Global Equities

Consumer Financial Protection Bureau (2025)

Disclaimer

This publication is for informational, educational, and entertainment purposes only and does not constitute financial, investment, legal, or tax advice. The content herein is a record of my personal research and investment process, and all analysis, forecasts, and opinions expressed are solely my own.

I make no representation or warranty, express or implied, as to the accuracy, completeness, or timeliness of the information provided. The stock market is highly volatile, and my forecasts, estimates, and assumptions may prove incorrect.

I am not acting as your financial advisor or fiduciary. You should not rely on any information in this publication to make investment decisions. Under no circumstances will I be held liable for any direct, indirect, or consequential losses or damages arising from your reliance on the content of this publication.

At the time of publication, I do not hold any positions in XYZ, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Pew Research Center (2022)

Hindenburg Research (2023)

Consumer Financial Protection Bureau (2025)

Payments Dive (2025)

XYZ (2026)

Bloomberg (2022)

iFAST: Singapore’s homegrown fintech, penetrating Hong Kong and beyond

Angsana Anderson — Tue, 23 Jun 2026 15:24:55 GMT

Summary

I passed on iFAST, Singapore’s homegrown fintech.

The business is what many investors would describe as a high quality compounder.

Indeed, iFAST has grown its revenue by 36% p.a. and EPS by 131% p.a. over the past 3 years.

However, I am not sure why customers are now taking more than 300 days to pay iFAST. I am also not sure about the durability of the profits from its UK banking subsidiary.

Finally, iFAST seems overvalued relative to global comparables. The consensus seems to be underestimating the industry’s cyclicality.

About (22 Jun 2026)

Share price: SGD 9.15

Market capitalisation: SGD 2,786 mn (USD 2,154 mn)

Enterprise value (EV): SGD 1,799 mn (USD 1,391 mn)

Average daily volume (ADV): SGD 11 mn (USD 9 mn)

NTM P/E: 23x

Time spent: ~2 days

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Pass

Background

Is iFAST Corporation Ltd. (AIY; IFAST SP) overvalued?

Over the past 10 years, shareholders have been rewarded with ~27% p.a. total return.

Today, IFAST trades at 23x NTM P/E. It targets SGD 100 bn assets under administration by 2030. Its UK banking division just turned profitable.

However, I have some questions.

Why did receivable days increase so much?

Why are global peers like Allfunds Group plc (ALLFG NA) trading at only 18x NTM P/E?

Business model

Estimated breakdown of 2025 net revenue (SGD 340 mn; +37% YoY):

39% iFAST ePension (+65% YoY)
38% Business-to-Business (B2B) Wealth Management (+18% YoY)
12% Business-to-Consumer (B2C) Wealth Management (+10% YoY)
11% iFAST Global Bank (+64% YoY)

iFAST is a global digital banking and wealth management platform with assets under administration (AUA) of SGD 33 bn as of 31 Mar 2026.

The iFAST ePension segment provides pension administration services and white-labelled solutions to enable digital access, management and processing of pension scheme transactions.

This segment is the prime subcontractor for the eMPF Platform project. This project, funded by the Hong Kong government, aims to create a digital experience for the city’s Mandatory Provident Fund (MPF) system.1

The Wealth Management segment provides access to investment products such as unit trusts, bonds, stocks, and ETFs. This segment connects product providers like fund houses with Business-to-Consumer (B2C) self-directed investors and Business-to-Business (B2B) partners, such as financial advisory firms and banks.

The most important source of revenue here is trailer fees. This is a recurring fee paid by fund houses to iFAST for distributing their investment products, typically calculated as a percentage of the AUA held on iFAST’s platform.

The iFAST Global Bank segment is a fully licensed UK bank acquired in 2022.

In 2025, iFAST earned 51% of its net revenue from Hong Kong, 33% from Singapore, and 11% from the UK. Malaysia contributed 5% while China’s revenue is immaterial.

The UK and Hong Kong were the fastest-growing markets, with +64% and +52% YoY net revenue growth respectively. China and Malaysia grew more slowly, at +24% and +14% YoY. Singapore grew +16%.

Unit trusts make up 56% of AUA. Stocks & ETFs contribute 24% while bonds contribute 11%. The remaining 9% is held in cash accounts and deposits.

The most relevant public comparables are Allfunds Group plc (ALLFG NA) for B2B and Futu Holdings Limited (FUTU US) for B2C.

My reasons

Why are customers taking so long to pay?

Revenue and profits are growing. The bull case involves high growth. In the past 3 years, iFAST has grown its revenue by 36% p.a. EPS grew even faster, at 131% p.a.

iFAST’s growth was mainly driven by strong net inflows in AUA and its involvement in Hong Kong’s eMPF Platform project.

But little cash. Despite growing profits, iFAST seems to need more cash. iFAST borrowed a combined SGD 220 mn in Jun 2024 and Mar 2026. Why?

The problem seems to be growing receivables. Over the past 3 years, receivables grew +88% p.a., far outpacing +35% p.a. revenue growth.

Are receivables increasing due to 3^rd party financial advisers? Unlikely. When a customer buys investment products on iFAST’s platform through a 3^rd party financial adviser, iFAST collects a fee from the customer. A portion of the fee is due to the 3^rd party financial adviser.

iFAST records both a receivable (from customer) and a payable (to financial adviser).

Could the receivables growth be due to increasing contribution from transactions like these? If this is true, both receivables and payables should grow in line. There should be no significant growth in excess receivables (receivables less payables).

But this is not what I observed.

Before 2022, excess receivables were stable. They started surging from 2022 onwards, suggesting the recent receivables growth is likely due to other factors.

Source: iFAST annual reports, Angsana Anderson estimates

In 2016, customers took around 91 days to pay iFAST. By 2025, this has increased 330%. Customers are now taking around 301 days to settle invoices.

Likely due to Hong Kong’s eMPF. The surge in excess receivables coincided with the start of revenue recognition from the eMPF Platform project.

In fact, the SGD 262 mn of excess receivables coincides with my estimate of all the revenue that iFAST has recognized from the project (~ SGD 244 mn).

Projects like this are usually based on milestone billing, where iFAST is paid when it completes certain predefined stages.

iFAST reported a credit term of 4 to 6 months for its project fee receivables.2 During the Q4’25 earnings call, management reported trustees onboarding was largely completed in 2026. If so, why did receivables still grow +3% YoY in Q1’26?

Problems with eMPF Platform. In Jun 2025, Bloomberg reported glitches on the eMPF platform, which it described as built by PCCW with operational support from iFAST. The glitches include log-in difficulty and missing information.3

As recently as Mar 2026, the platform is still receiving many complaints. The government has announced that affected members can claim compensation.4

iFAST stated that its role is limited to operational and user delivery services.5

Has iFAST’s UK bank finally turned around?

Finally profitable in 2025. The bull case also involves a turnaround at iFAST Global Bank, its fully licensed banking subsidiary in the UK. iFAST acquired this subsidiary in 2022 when it was suffering from persistent losses. By 2025, the bank finally turned a profit.

Steepening yield curve contributed to profits, but this is beyond the bank’s control. The UK yield curve steepened significantly during the first three quarters of 2025. This benefited banks like iFAST, which borrow at short-term rates and lend at long-term rates.

iFAST’s net interest margin (NIM) improved from 0.8% in the prior year to 1.1% in 2025. Net interest income increased by GBP 5.6 mn, contributing to the lion’s share of the GBP 1.9 mn profit before tax (PBT).6

Given that the yield curve is beyond the bank’s control, I don’t think the profits in 2025 should be seen as durable. In fact, the UK yield curve flattened during Q4’25. iFAST Global Bank reported PBT declined -31% YoY in Q1’26.

Why is margin so high? 2025’s profits were aided by GBP 3.0 mn increase in net fee and commission income. Fee and commission income increased while the corresponding transaction expense decreased. As a result, net fee margin jumped from 72% to 87%.

It is quite rare to see such drastic improvement in margins, and such high margins. For comparison, Remitly Global, Inc. ( US), another remittance provider, reports only margins around 68%.7

iFAST Global Bank explained that fee and commission income increased due to increased transaction volume at EZRemit, as a result of onboarding new originating counterparties in the Middle East.8

However, I could not find further details on these counterparties, and the end-customers.

Why is iFAST trading at a premium over comparables?

Industry is cyclical. In 2022, AUA declined -8%. The broader stock market decline caused valuation losses and also slowed down net inflows. As a result, operating profit fell -55%.

The bear market in 2018 caused a similar deterioration in performance.

Consensus seems to underestimate cyclicality. iFAST trades at 23x NTM P/E, much higher than its B2B comparable, ALLFG (18x) and its B2C comparable, FUTU (7x).

The consensus seems to justify the premium multiple because of iFAST’s growing earnings visibility and its AUA target of SGD 100 bn by 2030. Sell-side consensus forecasts double-digit growth rates in each of the next three years.

However, I believe such high multiple does not leave enough margin of safety for the industry’s cyclicality.

That said, short-term looks ok. Uncompleted contracts represent contract amounts for client trades that have been executed but not yet settled. It is a leading indicator because of the correlation between trades and AUA.

In Q4’25, uncompleted contracts increased +45% YoY and appears to have continued increasing in Q1’26.

Factors that could lead to a re-assessment of my decision

iFAST starts collecting its receivables (i.e. days receivable normalize below 6 months, the upper limit of the company’s credit terms).
iFAST Global Bank reaches sufficient scale. (e.g. Loan book scales up. Currently, investment grade corporate bonds still form the bulk of its interest-earning assets).
Share price falls >30% without a significant deterioration of business fundamentals.

Coming up next

Previously, I discussed PayPal: Undervalued at 7.8x P/E?

Like PayPal , Block, Inc. has sold off significantly over the past 5 years, dropping almost -70%.

Unlike PayPal, Block is expected to grow its EPS by +45% p.a. over the next two years. PayPal is expected to grow its EPS by only +4% p.a.

Could Block be more attractive?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in AIY, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

iFAST Global Bank (2025)

Remitly (2026)

iFAST Global Bank (2025)

PayPal: Undervalued at 7.8x P/E?

Angsana Anderson — Fri, 19 Jun 2026 11:24:21 GMT

Summary

I shortlisted PYPL.

At 7.8x NTM P/E, PYPL is priced like a terminally declining business.

However, investors seem to be underestimating its growth potential in buy now, pay later (BNPL), Venmo and advertising.

Despite tremendous growth in recent years, BNPL value per active user remains significantly below industry average.

BNPL offers higher take-rates and increases transaction value. Unlike traditional lenders, PYPL sells most of its BNPL receivables. This allows it to grow without being constrained by BNPL’s capital intensity.

PYPL has significant room to monetise Venmo. Compared to peers like Revolut, Venmo’s product offerings are still limited.

UBER’s advertising unit is already generating USD 2 bn annual revenue. PYPL has more active users and deeper data, but its advertising revenue appears to be immaterial.

This is changing. In 2024, PYPL hired a senior executive from UBER’s advertising unit. In 2025, PYPL’s advertising unit started launching major products.

Catalyst: potential takeover by Stripe.

About (18 Jun 2026)

Share price: USD 42.51

Market capitalisation: USD 37,498 mn

Enterprise value (EV): USD 39,826 mn

Average daily volume (ADV): USD 665 mn

NTM P/E: 8x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

PayPal Holdings, Inc. (PYPL US) has fallen -85% from its peak in Jul 2021.

It is now trading at only 7.8x NTM P/E, an all-time low. Historical average was ~ 27x.

Since Feb 2026, Stripe is said to be interested in acquiring PYPL.

Is there an opportunity?

Business model

Breakdown of 2025 revenue (USD 33,172 mn; +4% YoY):

90% Transaction Revenue (+3% YoY)
10% Other Value-Added Services (OVAS) Revenue (+14% YoY)

Transaction revenue consists mainly of net transaction fees charged to merchants and consumers for transactions completed on PYPL’s payments platform.

OVAS revenue consists of interest and fees from consumer and merchant loans, revenue-sharing partnerships for branded credit cards, subscription fees, and the interest earned on underlying customer account balances.

Main customers consist of everyday consumers who use its digital wallets to shop and manage money, and merchants who rely on its payment processing tools to drive sales and run their businesses.

In 2025, no customers accounted for more than 10% of revenue or net loans receivable.

Main suppliers include payment card networks, partner banks and payment processors.

In 2025, two payment processors accounted for 56% of transaction expense. These are likely Mastercard Incorporated ( US) and Visa Inc. ( US).

PYPL earns 57% of its revenue from US (+3% YoY) and 43% from the rest of the world (+6% YoY).

Block, Inc. ( US) is the most relevant public comparable.

It owns Cash App, which directly competes with PYPL’s Venmo for P2P transfers, direct deposits, etc. In 2022, XYZ acquired Afterpay to integrate BNPL into its ecosystem.

Some analysts compare PYPL to payment processors like Fiserv, Inc. ( US). However, the total return chart above suggests legacy payment processors are driven by other factors. XYZ seems to be the best comparable.

For a detailed explanation of PYPL’s business model, I recommend Bob’s Payment Stock Substack’s PayPal: Past, Present and Future.

My reasons

Under-recognised growth?

BNPL is under-penetrated. In 2025, PYPL generated ~ USD 40 bn in BNPL transaction volume1 across its 227 mn monthly active accounts (MAUs)2.

Blended across its entire active user base, this equates to ~ USD 176 in BNPL volume per MAU. For context, the US industry average sits significantly higher at USD 848 per active BNPL user.3

This comparison is not perfect because PYPL’s denominator includes customers for whom BNPL may never be suitable.

That said, the sheer magnitude of this gap strongly suggests under-penetration of BNPL among PYPL’s user base.

BNPL increases take-rate and transaction value. PYPL’s take-rate for an ordinary payment is only 2.6%, on average. If the customer completes the transaction using BNPL, PYPL’s take-rate increases to 4.5%.

BNPL also increases transaction value. When customers pay using BNPL, their average order value is on average more than 80% higher.4

BNPL is capital-light. Lending money is capital intensive. That’s why MoneyMax Financial Services Ltd. (5WJ; MMFS SP), Singapore’s leading pawnbroker, issued more shares recently.5 Without the infusion of new capital, growth will probably stall.

PYPL does not face the same problem because it has forward flow arrangements under which it sells most of its BNPL receivables to third‑party investors such as KKR & Co. Inc. (KKR US).6

Venmo is under-monetised. Venmo is PYPL’s US digital wallet and mobile app. It is popular among younger, mobile‑first consumers, especially Millennials and Gen Z. In 2025, Venmo accounted for ~ 19% of PYPL’s total payment volume (TPV).7

To show how under-monetised Venmo is, I’ve compared its product offerings against Revolut.

Revolut started in 2015 as a digital wallet. Today, it has evolved into a fully licensed digital bank and financial super-app.

Revolut monetises its users by charging subscription fees for premium perks like personalised cards, discounted airport lounge access, etc. It also distributes money market funds and offers stocks and commodities trading.

Source: Revolut (2026)

On the other hand, Venmo’s product offerings are sparse. It seems like Venmo does not even offer trading in stocks and ETFs.

Source: Venmo (2026)

Although Venmo is currently under-monetised, it also means growth opportunities.

Venmo already has more than 95 mn active accounts.8 It should not be too difficult to expand its product offerings and increase monetisation.

USD 2bn revenue from advertising? In 2022, Uber Technologies, Inc. ( US) launched its advertising business. By 2025, Uber Advertising brought in more than USD 2 bn of annual revenue.9

Despite being immaterial now, PYPL’s advertising revenue has the potential to match or even exceed USD 2 bn. PYPL’s 227 mn MAUs exceed UBER’s 202 mn. PYPL has deeper data on its users’ spending habits.

In 2024, PYPL hired Mark Grether, who formerly led UBER’s advertising business, to lead its newly created PayPal Ads division.10

PYPL’s advertising division seems to be gathering momentum. In 2025, PYPL launched PayPal Ads Manager, a zero-cost tool that enables merchants to monetize store traffic by displaying relevant advertisements on their websites.11

Capital returns?

Attractive capital return. In the last twelve months, PYPL returned USD 6,545 mn to shareholders, mainly through share buybacks. This represents ~17% of its market capitalisation today.

I estimate PYPL generates ~ USD 4,500 mn of free cash flow to firm (FCFF).

If capital returns normalise to FCFF, this will still represent ~12% yield on today’s market capitalisation. The 8% premium over United States 10y government bond looks attractive.

PYPL can likely sustain returning ~100% of its FCFF to shareholders because it is a capital-light business. Working capital requirements are not significant because PYPL sells most of its loan receivables upon origination.

Catalysts

Potential sale. In Feb 2026, Bloomberg reported that Stripe Inc., a payment processor, is considering an acquisition of all or parts of PYPL.12

The probability of a sale is high, in my view. PYPL is trading at an all-time low NTM PE of 7.8x, despite significant growth opportunities in BNPL, Venmo and advertising.

In Apr 2026, new CEO Lores simplified PYPL’s organisation into 3 segments: Checkout Solutions & PayPal, Consumer Financial Services & Venmo, and Payment Services & Crypto.13 This cleaner structure makes it easier for acquirers to buy part or all of PYPL.

Factors to focus on

Growth potential. My initial estimates suggest PYPL can generate an incremental annual revenue of at least USD 2 bn just from raising BNPL penetration.

I expect another USD 2 bn incremental annual revenue from its advertising division.

How much incremental revenue can Venmo generate?

How reasonable are my initial estimates? How successful will the new CEO be in realizing these growth opportunities?

Sustainability of transaction margin. Transaction margin refers to the portion of net revenues left after paying transaction expense and transaction and credit losses.

Despite the growth opportunities in BNPL, Venmo and advertising, traditional transaction fees remain the bulk contributor to revenue and profits.

PYPL’s sell-off on 3 Feb 2026 was partly driven by its guidance for a slight decline in transaction margin during 2026. This was unexpected.

Beyond 2026, will transaction margin continue declining?

Coming up next

In the past 10 years, shareholders of iFAST Corporation Ltd. (AIY; IFAST SP) enjoyed 27% return per year.

This is despite the saturation of the market by new players like Futu Holdings Limited ( US).

What’s the secret behind AIY’s success?

Is there still an opportunity for investors?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Bob’s Payment Stock Substack (2025)

Disclaimer

At the time of publication, I do not hold any positions in PYPL, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Consumer Financial Protection Bureau (2025)

The Edge (2026)

Bob’s Payment Stock Substack (2025)

Marketing Dive (2026)

Sido Muncul: 11% FCF yield. But 49% of revenue from related parties?

Angsana Anderson — Tue, 16 Jun 2026 18:15:39 GMT

Summary

I shortlisted SIDO, one of the top makers of herbal medicine in Indonesia.

Investors may be under-estimating its growth potential, particularly in international markets.

SIDO seems to be a good business. It is reinvesting capital at higher rates of return, as evidenced by its superior and growing return on assets (ROA).

Recent revenue decline seems to be a temporary headwind, driven by management’s decision to reduce channel inventory rather than a structural problem.

On that basis, SIDO’s 11% free cash flow yield seems sustainable and attractive.

That said, now may not be the best time to invest. Working capital trends suggest consensus remains too optimistic over near-term demand.

Although related parties contributed to only 29% of group revenue in Q1’26, it has averaged ~50% historically. I will need to investigate why, and whether this risk is too big.

I am not too concerned over macroeconomic risks. These can be managed through diversification or shorting of an Indonesia index ETF.

About (15 Jun 2026)

Share price: IDR 380

Market capitalisation: IDR 11,185 bn (USD 632 mn)

Enterprise value (EV): IDR 10,417 bn (USD 588 mn)

Average daily volume (ADV): IDR 8.7 bn (USD 0.5 mn)

NTM P/E: 9x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

Are investors wrong about Indonesia’s top herbal medicine?

PT Industri Jamu Dan Farmasi Sido Muncul Tbk (SIDO IJ) makes Tolak Angin, Indonesia’s most popular herbal medicine with ~72% market share.

At first glance, Sido Muncul is a lousy business. Revenue declined -19% in Q1’26. Its share price has fallen -30% in 2026.

However, investors are likely under-estimating the company’s growth potential.

Its headwinds are likely temporary.

Furthermore, the company also offers double digit free cash flow yield.

Business model

SIDO is an Indonesian company that manufactures and sells herbal supplements, food and beverages and pharmaceutical products.

Breakdown of 2025 revenue (IDR 4,080 bn; +4% YoY):

61% Herbal Medicine & Supplements (+0% YoY)
36% Food & Beverages (+12% YoY)
3% Pharmaceutical (+1% YoY)

The Herbal Medicine & Supplements segment manufactures health products such as Tolak Angin, Indonesia’s most popular jamu. Jamu is a traditional herbal medicine made from turmeric, ginger and other herbs used to boost immunity and relieve inflammation.

The Food & Beverages segment manufactures ginger coffee, ginger tea, energy drinks, etc. Brands include Kuku Bima Ener‑G and Alang Sari Cool.

The Pharmaceutical segment focuses on the production and distribution of various health products under the Berlico brand. These products include Anacetine, Berlosid, Anabion, Minyak Telon Cap Tiga Anak, and a variety of other products.

Most herbs and spices appear to be sourced from farmers across Indonesia, which implies a fragmented base of small and mid‑size agricultural suppliers.

Main customers are distributors across Indonesia. There are no individual customers that contributed > 10% of revenue in 2025.

SIDO earns 91% of its revenue from Indonesia. The remaining ~9% comes from international markets like Malaysia, Philippines and Nigeria.

According to management, Tolak Angin holds ~72% market share. The next closest competitors hold ~23% and 5%.

The most relevant public comparable is PT Kalbe Farma Tbk. (KLBF IJ).

Both SIDO and KLBF sell herbal medicine and supplements in Indonesia, but KLBF is a far larger, diversified pharma‑healthcare conglomerate while SIDO remains a focused herbal and FMCG player.

My reasons

Under-recognised growth?

Exports are growing fast. Since disclosure started in 2022, international revenue has grown 33% p.a. and reached 9% of overall revenue in 2025. During the same period, overall revenue increased by only 2% p.a.

Malaysia and Nigeria contributed 2% and 4% to overall revenue respectively.

There are precedents for Indonesian products to penetrate these markets. For example, Indonesia’s Indomie has captured ~74% of the market in Nigeria.1 Like Indonesia, both markets have significant Muslim population.

Good business, temporary headwinds?

Revenue declined in Q1’26. When SIDO released earnings on 8 May 2026, its share price started collapsing faster than the broader market.

SIDO’s revenue declined -19% YoY in Q1’26. This was unexpected, given management’s target of 5% to 8% revenue growth for 2026.

Headwind likely temporary. Management explained they deliberately restricted sales after discovering that channel inventory had accumulated to as much as two to three months’ supply.2

SIDO is targeting inventory levels of ~ 14 days for distributors in Java and ~ 21 days for areas outside Java.

Despite that, management reports consumer demand has not weakened. They believe sell-in remains healthy.

That’s why even though it no longer believes 5% to 8% revenue growth is achievable, SIDO is still targeting stable revenue in 2026.

Good business. Despite the headwinds, SIDO appears to be a good business.

Its flagship product, Tolak Angin, is priced at ~20% to 50% premium over competitors. Despite that, it still holds 72% market share in Indonesia.3

SIDO is reinvesting incremental capital to earn higher returns. Between 2016 and 2025, it improved its ROA from 13% to 25%. SIDO’s 25% ROA is significantly higher than the 10% earned by its domestic peer, KLBF.

Capital returns?

SIDO is generating cash. SIDO generated ~ IDR 1,177 bn of free cash flow to firm (FCFF) in the last twelve months (LTM).

Against today’s enterprise value (EV), SIDO is offering ~11% FCFF yield. This is a decent 4 percentage points (ppt) premium over Indonesia 10y government bond.

And returning cash to shareholders. Management demonstrated a good track record of returning excess capital to shareholders and not hoarding cash.

In 2016, cash was ~39% of revenue. By 2025, management had reduced it to a more reasonable 11% of revenue.

This was achieved mainly through a high dividend payout ratio. Management also started significant share buybacks in 2025, resulting in ~ IDR 1,600 bn capital return to shareholders in that year alone.

Catalysts

Potential sale. In Jan 2026, Bloomberg reported that the Hidayat family, SIDO’s controlling shareholders and management are open to bringing in a strategic investor to take the business “to a higher level”.

The family is in talks with banks. A transaction could value SIDO at ~ USD 1 bn, +70% higher than the current enterprise value.4

Factors to focus on

Timing of channel destocking. Even though SIDO looks like a great business, now may not be the best time to invest. The consensus still seems too optimistic.

According to Tikr, the consensus is expecting ~+2% revenue growth in 2026. This is more or less in line with management’s guidance of stable revenue.

If SIDO is to achieve this, it must grow revenue by +7% YoY in the remaining 3 quarters of 2026.

This seems too optimistic. Working capital trend suggests demand from distributors has not bottomed out.

In Q1’26, receivable and inventory days reached new highs and are not showing signs of peaking yet:

Furthermore, the % of current receivables declined from 64% in Q1’25 to 50% in Q1’26.

Despite revenue falling in Q1’26, inventory increased +17% YoY, with most of the increase concentrated on finished goods and work-in-progress.

With customers delaying payments and inventory piling up, near-term demand looks likely to be much weaker than consensus expects.

I will be monitoring the working capital trends over the next few quarters. When there are signs of inventory and receivable days peaking, that may be a better time to invest.

Significant related party transactions. In Q1’26, sales to related parties declined -43% YoY, significantly contributing to the -19% YoY decline in overall revenue.

During the quarter, sales to related parties contributed to 29% of revenue, down from 42% in Q1’25 and 49% in 2025.

The major related parties appear to be distributors owned by the Hidayat family, SIDO’s controlling shareholders and management.

Is such high concentration of related party transactions normal in Indonesia? What is the business rationale?

Are the transactions conducted on an arms-length basis?

Given such high % of sales from related parties, why didn’t SIDO foresee the -19% revenue decline in Q1’26?

Macroeconomic volatility. The Jakarta Composite Index (JCI), an index of Indonesian stocks, fell -28% in 2026. The IDR has depreciated -6% against the USD.

Such volatility is mainly driven by concerns over increasingly populist and interventionist government policies.5

It is difficult to predict the direction of government policies.

In any case, I don’t have to bet on macroeconomic factors. If SIDO is indeed attractive, I can manage the macroeconomic risks through diversification or shorting the JCI.

Coming up next

PayPal Holdings, Inc. (PYPL US) has fallen -85% from its peak in Jul 2021.

It is now trading at only 7.8x NTM P/E, an all-time low. Historical average was ~ 27x.

Since Feb 2026, Stripe is said to be interested in acquiring .

Is there an opportunity?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in SIDO, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

The Straits Times (2026)

MegaStudyEdu: Why South Korea's top private educator can be a 3x

Angsana Anderson — Fri, 12 Jun 2026 19:45:58 GMT

Summary

I bought shares in MegaStudyEdu Co. Ltd (215200 KS).

At 4x P/E and 8% dividend yield, South Korea’s top private educator is priced like a dying business. It’s not.

Despite declining student population, the industry is still growing. More parents are enrolling their children in private education, and they are also spending more per child.

South Korea’s expenditure on private education has grown steadily over the past 10 years. In 2025, it declined for the first time.

That -6% decline was concentrated among lower-income households, which supports my view that the weakness is cyclical rather than structural.

Since 2024, the government’s Neulbom School program has provided competing services. However, in Mar 2026, the program stopped expanding.

The trends in advance payments from customers, a leading indicator, suggests revenue has likely bottomed out.

MegaStudyEdu’s dividends and share buybacks in 2025 represented ~12% of its current market capitalisation. This is an attractive 7 percentage points (ppt) premium over South Korea’s 10-year government bond yield. It also appears sustainable.

Catalysts: activist investor and potential sale to private equity.

About (11 Jun 2026)

Share price: KRW 38,900

Market capitalisation: KRW 403 bn (USD 265 mn)

Enterprise value (EV): KRW 321 bn (USD 211 mn)

Average daily volume (ADV): KRW 2 bn (USD 1.3 mn)

NTM P/E: 4x

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

I bought my shares at KRW 37,700. Within 3 to 5 years, I expected to sell my shares for at least KRW 125,000.

Over 3 years, I expected to earn ~57% p.a. (49% capital gains; 8% dividend yield). Over 5 years, I expected to earn ~35% p.a. (27% capital gains; 8% dividend yield).

Background

A month ago, I shortlisted MegaStudyEdu Co. Ltd (215200 KS). Since then, its share price has fallen almost -17%.

Investors seem to believe the business will shrink indefinitely.

However, recent earnings showed MegaStudyEdu maintained revenue, while increasing net profit after tax by 8% YoY in Q1’26.

The indicators I monitor suggest MegaStudyEdu has reached a bottom.

Business model

Please see my prior post: MegaStudyEdu: Why is South Korea’s top online school trading at only 5x P/E?

My reasons

Under-recognised growth

Priced like a dying business. At 4x NTM P/E, the market seems to be pencilling in a significant decline in MegaStudyEdu’s profits.

“Don’t you know South Korea’s population is declining?”

Indeed, the number of elementary, middle and high school students has been declining in the past 10 years:

But the industry is growing. Despite declining student population, South Korea’s total private education expenditures increased 4% p.a. in the past 10 years.

More students are using private education. In 2016, 68% of all students used private education. By 2025, the participation rate has increased to 76%.

More students use private education, and spending per student has also been rising steadily.

Fewer kids, more investment per kid. At first glance, MegaStudyEdu seems to be operating in a dying industry. South Korea’s student population is declining.

However, I believe the market is underestimating MegaStudyEdu’s growth potential. The industry has been growing ~4% p.a. in the past 10 years.

South Korean couples are having fewer children, but more parents are enrolling their children in private education, and they are spending more per child.

Good business, temporary headwinds

Private education expenditure declined -6% in 2025. This was driven by -7% decline in the number of participating students, partially offset by +2% increase in spending per participating student.

Temporary economic weakness. I believe the -6% decline in 2025 reflects a temporary headwind, rather than a structural decline.

Maeil Business News Korea reported “[In 2025], real consumption expenditure for all households also declined by 0.7 percent, indicating that rising prices weakened purchasing power, leading families to cut education expenses.”1

Indeed, private education expenditure declined the most among lower-income households:

Government has halted the expansion of Neulbom Schools. MegaStudyEdu’s early childhood and elementary segment suffered the steepest revenue decline in 2025.

Peers like Digital Daesung Co., Ltd. (A068930) are reporting similar trend. I suspect Neulbom Schools.

In 2024, the Ministry of Education introduced Neulbom Schools. This program aims to provide after-school care for all elementary school students.

Instead of online lectures and studying at MegaStudyEdu’s academies (‘hagwon’), the children will participate in after-school activities and daycare for free or at low costs.

As of Mar 2026, 552,000 first and second graders in elementary school were using the service. The plan was to expand it to all elementary school grades starting 2027.2

However, I believe Neulbom Schools is no longer a significant headwind.

In Mar 2026, the new government led by Lee Jae Myung halted the expansion of his predecessor’s Neulbom School initiative.3

The program will be maintained only for first and second graders. Third graders will instead be provided with a voucher for after-school programs annually.

Effectively, the government has admitted that it is difficult to replace the services provided by MegaStudyEdu and its peers.

“There is a tendency for third graders and above to prefer after-school programs focused on subjects like English and Math over Neulbom School”.

Furthermore, the government is struggling to cope with the increased workload on school staff if more students participated in the program.

Favourable working capital trends. MegastudyEdu’s advance payment from customers (unearned revenue) has stabilised in Q1’26.

This suggests that these headwinds are indeed temporary, and that revenue has likely bottomed out.

Source: Tikr

Capital returns

Capital return represents ~12% of market capitalisation. In Apr 2025, MegaStudyEdu announced its corporate value-up program.4 It targets to return 60% of net income via dividends and share buybacks.

For 2025, MegaStudyEdu returned KRW 51 bn of capital to shareholders, representing 65% of net income. This is ~12% yield on market capitalisation, an attractive 7 ppt premium over South Korea’s 10-year government bond yield.

Attractive capital returns likely sustainable. For the reasons, I discussed above, I believe MegaStudyEdu’s 12% yield is sustainable.

MegaStudyEdu is also cash generative.

The company benefits from favorable working-capital dynamics. It collects fees in advance from students and pays its instructors later.

Besides some physical academies, there are no significant capital expenditure requirements. Most lectures are streamed online.

Valuation

On 11 Jan 2026, MegaStudyEdu was trading around KRW 37,700 per share.

At that price, I believed MegaStudyEdu was very attractive. 4x NTM P/E and 12% capital return yield on a business that is likely to maintain revenue and profits.

To take advantage of the share price decline, I bought my shares before completing my discounted cash flow (DCF) valuation.

My initial estimates suggest MegaStudyEdu is worth at least KRW 125,000 per share. At this price, it would trade around 15x NTM P/E and 3.6% capital return yield. This is more consistent with a stable business.

For context, the iShares MSCI South Korea ETF, which tracks an index of South Korean equities, trades at 17x LTM P/E.5

Historically, MegaStudyEdu has traded around 7x NTM P/E. The market misunderstood it as a business inside a structurally declining industry.

For such a quality business, it should trade much higher than 7x NTM P/E. Its corporate value-up program, which we just discussed, and interests from activist and private equity (PE) investors will help catalyse its value.

Catalysts

Activist investor. On 16 Apr 2024, Dalton Investments, an activist investor, disclosed it had been investing in MegaStudyEdu since 2020.

The investor commended the company’s responsiveness to their recommendations for higher capital returns. “We are now truly delighted to see management making such a full commitment [to capital returns].”

Although I could not find Dalton Investments in MegaStudyEdu’s shareholder registry, both current and historical, I believe the investor is probably still a shareholder. Dalton Investments describe their investment horizon as between 5 to 10 years.6

Having a prominent activist investor on its shareholder registry will likely incentivise MegaStudyEdu to create more value for minority shareholders.

Potential sale. On 4 Feb 2026, Maeil Business reported that the management (and major shareholders) are looking to sell 32% of the company to private equity. The sale price would value the entire company at around KRW 925 bn.7

The company denied this report on 5 Feb 2026.8

Despite that, I believe there is a good chance of an eventual sale.

The major shareholders (and management) are near retirement age. The company has not publicly identified a clear next-generation successor.

In 2022, the major shareholders were negotiating a sale to MBK Partners, one of North Asia’s largest private equity (PE) firms.9 The negotiation eventually broke down. Apparently, they disagreed on price.

Risks

Substitution risks. The Ministry of Education (MOE) provides free online CSAT lectures via Educational Broadcasting System internet (EBSi). The purpose is to help curb private education costs.

The MOE explicitly targets a ~50% linkage rate between EBSi materials and the actual CSAT exams. As a result, MegaStudyEdu faces substitution risks.

However, EBSi acts as a price and share ceiling at the low end of the market but has not displaced premium private CSAT prep providers like MegaStudyEdu. The company continues to differentiate on content quality, teacher reputation, and data-driven services.

EBSi has been free since it launched in 2004. In contrast, a MegaPass costs around KRW 600,000 (~ USD 400) per season. A MegaPass is the company’s all-access online subscription service providing unlimited access to lectures.

Despite the price difference, I estimate MegaStudyEdu grew its revenue by 4.4% p.a. since 2004. A cursory browse through DC Insider, the Reddit-equivalent in South Korea, reveals students still perceive the company’s star lecturers as superior.

Regulatory risks. On Jul 2021, the government in mainland China banned for‑profit K‑12 after‑school tutoring.

The share price of tutoring companies collapsed. New Oriental Education & Technology Group Inc. (EDU US) fell almost -92% from its peak.

There are two reasons why I believe EDU is not a good precedent for MegaStudyEdu.

First, mainland China is governed by one party. South Korea is governed through a democratic process. Drastic laws like an overnight ban of for-profit tutoring will be much more difficult.

Second, before the ban, EDU was trading at ~20x NTM P/E. Such high valuation meant there was little margin of safety for investors.

Whereas today MegaStudyEdu is trading at only 4x NTM P/E. If a ban wipes out 50% of its earnings, the company will still trade at an undemanding 8x NTM P/E.

AI disruption. I believe this risk is low.

The lecturers on MegaStudyEdu are not just teachers. They are celebrities.

According to Korea JoongAng Daily, “The relationship between students and star lecturers is similar to the one between fans and K-pop stars. Students even buy merchandise that includes bags, mugs and stickers.”10

The teachers not only teach, but they motivate, entertain, console and even admonish.

Factors that could lead me to increase investment

MegaStudyEdu grows its revenue more than mid-single digit per year (e.g. strong inflection in the growth of advance payment from customers).
Indicators of a potential sale of the business (e.g. higher legal and consulting fees).
Share price falls > 30% without significant deterioration in business fundamentals.

Factors that could lead me to decrease investment

Unfavourable regulatory developments (e.g. lawmakers proposing new laws to ban private education)
Substitutes gaining traction (e.g. government reviving Neulbom Schools and expanding its coverage)

Upcoming events

Tue 11 Aug 2026: Q2’26 earnings release (expected)

Valuation model

Subscribers may request a free copy of my valuation model by emailing me at angsana.anderson@gmail.com.

I will send out my model by the end of Sun 14 Jun 2026.

Coming up next

The IDX Composite, an index of all stocks listed on the Indonesia Stock Exchange, has fallen -31% year-to-date. This is in response to seemingly erratic policymaking in the country.

Despite the chaos, could there be opportunities in Indonesia?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Maeil Business News Korea (2026)

Disclaimer

At the time of publication, I hold shares in 215200. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

The Chosun Daily (2026)

MegaStudyEdu (2025)

iShares (2026)

Dalton Investments (2026)

Maeil Business (2026)

MegaStudyEdu (2026)

KED Global (2022)

Korea JoongAng Daily (2023)

GQG: Good business, temporary headwinds?

Angsana Anderson — Tue, 09 Jun 2026 15:59:41 GMT

Summary

I shortlisted GQG.

Consensus believes funds under management (FUM) will decline.

However, GQG seems to be a good business facing temporary headwinds.

Since Q3’25, investors have been redeeming their funds because GQG is underperforming its benchmarks. GQG’s relative underperformance is driven by its underweighting of AI and semiconductor stocks.

I believe there is a low risk of a sudden and large collapse in FUM.

Net outflows are decelerating. This suggests short-term investors have largely redeemed.

With the S&P 500 concentrated in AI and semiconductor stocks carrying stretched valuations, there is demand for non-AI and non-US funds, like those offered by GQG.

Finally, GQG offers an attractive 15% dividend yield. This seems sustainable. The business does not require significant investment in working capital or capital expenditures.

Key factors to look into: (a) probability and timing of inflection in net flows; (b) customer concentration; (c) key man risks.

About (05 Jun 2026)

Share price: AUD 1.39

Market capitalisation: AUD 4,112 mn (USD 2,897 mn)

Enterprise value (EV): AUD 3,949 mn (USD 2,783 mn)

Average daily volume (ADV): AUD 6.0 mn (USD 4.2 mn)

NTM P/E: 7x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

Is there an opportunity in 15% dividend yield?

An investment manager is now back at its IPO price.

The company is wrestling with fund outflows.

Its relative performance has been deteriorating because of its decision to avoid AI stocks.

However, these headwinds seem to be temporary.

Business model

GQG is an investment boutique investing in listed equities globally.

In 2025, 98% of its revenue comes from management fees (+8% YoY). Performance fees (-44% YoY) are immaterial, making up only 2% of total revenue.

GQG earns 81% of its revenue from the United States. No individual foreign country contributed more than 10% of total revenue.

GQG manages four strategies: Global Equity, International (non-US) Equity, Emerging Markets Equity, and US Equity.

As of 30 Apr 2026, the International Equity strategy accounts for 44% of funds under management (FUM). The Emerging Markets strategy accounts for 24%.

Source: GQG (2026)

41% of FUM is managed through the sub-advisory channel. A sub-advised fund is an investment fund that is formed and managed by a third-party firm that retains GQG to manage part or all of the fund.

32% of FUM is managed through the institutional channel. These funds come from investors with large pools of investable assets such as insurance funds, pension/superannuation funds, sovereign wealth funds and ultra-high net worth investors.

The remaining 27% is managed through the wholesale channel. This channel is typically financial intermediaries such as financial advisers, wealth management administration platforms or private banks.

Source: GQG (2026)

The investment management industry is fragmented but increasingly consolidating.

Public comparables of similar size include Magellan Financial Group Limited (MFG AU), Perpetual Limited (PPT AU) and Jupiter Fund Management Plc (JUP LN).

My reasons

Good business, temporary headwinds?

Consensus expects decline. According to Tikr, consensus sell-side forecasts -3% and -4% revenue decline in 2026 and 2027.

Net outflows. Consensus expects revenue to decline because investors have been redeeming their capital, resulting in net outflows since Q3’25:

As of 30 Apr 2026. Source: GQG (2026)

The net outflows are largely the result of relative underperformance.

GQG has outperformed its benchmark across all 4 strategies since inception. However, its performance has been lagging benchmarks in recent years:

Source: GQG (2026)

Defensive positioning. GQG’s relative underperformance has been mainly driven by GQG’s decision to avoid AI-related stocks.

Goldman Sachs GQG Partners International Opportunities Fund is a sub-advisory fund managed by GQG. It accounts for ~33% of GQG’s FUM.

The top 10 holdings consist of companies in consumer staples and energy like Philip Morris International Inc. (PM US) and TotalEnergies SE (TTE FP).1 There are no AI-related companies.

Net outflows are decelerating. The net outflows since Q3’25 likely consist of clients who cannot tolerate short-term underperformance stemming from a defensive positioning.

There are signs that such clients have largely exited.

Monthly net outflows are starting to decelerate:

Low risk of FUM collapse. With the S&P 500 becoming more concentrated in AI-related companies, and valuation reaching historical highs, there will be higher demand for non-US and non-AI funds to diversify risks.

GQG’s products offer that diversification, especially with its largest strategy, International Equity, which focuses on non-US stocks. This is likely what the existing clients are staying for.

I also believe the risk is low of sudden redemptions large enough to disrupt the business. GQG’s FUM is well diversified across institutional, sub-advisory and wholesale channels.

According to my discussion with an equity product specialist, institutional clients can take months before making final decisions on redemptions. The process is long.

They will usually engage consultants, who then conduct extensive performance attribution and studies before they come up with recommendations, etc.

GQG’s relatively low fees are another incentive for clients to stay. On average, GQG charges management fee that is ~0.5% of FUM. The industry average is around 0.7% to 1.0%.

Good business. The focus on the recent net outflows overshadows the quality of the business and management.

CIO Rajiv Jain and CEO Tim Carver founded GQG in 2016. In less than 10 years, they grew the startup into a USD 167 bn FUM business with ~200 professionals.

They built a good long-term track record. Since inception, all strategies have beaten their benchmarks, despite the recent underperformance.

CIO Jain and CEO Carver achieved these despite the challenges confronting active equity managers.

Peers like Aberdeen Group Plc (ABDN LN), MFG and JUP have been experiencing stagnant or even declining FUM, driven by market share losses to low-cost passive equity ETF offered by the likes of BlackRock, Inc. (BLK US).

Capital returns?

15% dividend yield seems sustainable. GQG targets a payout ratio between 50% to 95% of distributable earnings. In the past few quarters, the payout ratio averages ~90%. This means an attractive dividend yield of ~15%.

It seems like GQG can maintain the current level of dividend. The business does not require significant investment in working capital or capital expenditure. As discussed earlier, I believe the risk is low of a sudden and large collapse in FUM.

Catalysts

Inflection in net flows. Net flows is probably the single most important metric for investment managers.

In Apr 2025, JUP’s share price started recovering after it reported significant institutional net inflows in Q1’25, after multiple quarters of net outflows.

Factors to focus on

Probability and timing of inflection in net flows. When the AI and semiconductor rally fizzle, GQG’s relative performance will likely improve, given its defensive positioning. GQG should return to net inflows at that time.

As discussed in Where are we in the semiconductor cycle?, we are likely in the later stages of a semiconductor shortage.

The shortage could end around Q1’27. Because of forward guidance, investors will start reacting one or maybe two quarters before Q1’27. This means the market rally may end around Q4’26.

This is what I expect.

Could there be anything that can break my expectations? Are there any precedents?

Customer concentration. In its 2025 financial statement, GQG disclosed “One of our Mutual Fund clients accounted for 24.6% and 24.9% of the total revenue for the years ended December 31, 2025 and 2024, respectively.”

GQG added, “One of our SMA clients accounted for 22.7% and 20.6% of total revenue for the years ended December 31, 2025 and 2024, respectively.”

I suspect the Mutual Fund client refers to GQG’s sub-advisory relationship with Goldman Sachs, most notably the Goldman Sachs GQG Partners International Opportunities Fund. I believe the concentration risk is low because the end clients are likely numerous.

Who is the separately managed account (SMA) client? Is the concentration risk acceptable? If GQG loses this client, is current valuation still attractive?

Key man risk. Talent is the biggest asset in the investment management business.

The sudden departure of its CEO and CFO contributed to the downfall of a peer, MFG AU.2

In Feb 2025, JUP said outflows in 2024 were driven by more than GBP 6 bn pulled from strategies managed by Ben Whitmore, one of its star portfolio managers.3

For GQG, the key talents are its founders, CIO Jain and CEO Carver. CIO Jain has primary and veto authority over all portfolio management decisions and is chiefly responsible for portfolio parameters.

I do not worry about them leaving GQG. They own ~70% and 4% of GQG, respectively. However, if anything untoward happens to them, it could cause significant disruption to the business.

If the founders were to retire, how would GQG perform? Does the business have potential successors that can take over the CIO and CEO’s duties?

Coming up next

A month ago, I shortlisted MegaStudyEdu Co. Ltd (215200 KS).

Investors seem to be pricing in an indefinite revenue decline and underestimating its growth.

The company is returning ~12% of its market capitalisation to shareholders via dividends and buybacks.

Since then, its share price has fallen almost -17%.

This is despite Q1’26 earnings showing the business maintained revenue, while increasing net profit after tax by 8% YoY during the quarter.

Is there an opportunity?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in GQG, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Goldman Sachs (2026)

InvestorDaily (2023)

Financial News (2026)

Grab: Not as cash-generative as you think

Angsana Anderson — Thu, 04 Jun 2026 16:45:39 GMT

Summary

I decided to pass on GRAB US.

GRAB’s adjusted free cash flow overstates its true cash-generating ability because it includes stock-based compensation (SBC). SBC is actually financing from employees.

Despite being the dominant platform across South-east Asia, GRAB’s pricing power is limited.

Several factors contribute to this: more price-sensitive consumers, stronger competition. The factor least appreciated: regulatory pressure.

Although GRAB’s shares are not obviously attractive to me now, I believe the business will likely still do well.

CEO Anthony Tan and his team have demonstrated strong ability to raise funds.

About (4 Jun 2026)

Share price: USD 3.41

Market capitalisation: USD 13,947 mn

Enterprise value (EV): USD 9,437 mn

Average daily volume (ADV): USD 171 mn

NTM P/E: 33x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Pass

Background

“Buy GRAB US! A cash-generative monopoly with high growth in fintech”

That is the consensus bull case.

However, Grab is not as cash-generative as many investors think.

Its pricing power still looks weak, despite being the dominant platform in South-east Asia.

Its expansion into fintech reinforces the challenges in its core segments.

Although its shares are not obviously attractive to me, I believe Grab, as a business, will still do well. The secret weapon? Anthony Tan.

Business model

Breakdown of 2025 revenue (USD 3.4 bn; +20% YoY):

53% Deliveries (+21% YoY)
36% Mobility (+16% YoY)
11% Financial services (+37% YoY)

GRAB operates an app providing deliveries, mobility, and financial services across South-east Asia. As of Q1’26, the company serves 52 mn monthly transacting users.

The Deliveries segment connects drivers and merchants with consumers looking for deliveries of meals, groceries and parcels.

The Mobility segment connects drivers with consumers looking for transport. Consumers can hire private cars, taxis and motorcycles through the Grab app.

The Financial Services segment provides banking services such as digital payments, deposits, loans and receivables factoring.

This segment is carried out by GXS Bank Pte. Ltd (GXS), a joint venture with Singapore Telecommunications Limited (Z74; ST SP). Because GRAB retains control, GXS is consolidated into GRAB’s financial statements.

Customers are typically urban residents across South-east Asia who rely on the Grab app for everyday conveniences like ride-hailing, food delivery, and digital financial services.

No significant portion of GRAB’s revenue can be attributed to a particular customer or group of customers.

GRAB earns 31% of its revenue from Malaysia, 22% from Singapore and 21% from Indonesia. The remaining 26% is derived from the Philippines, Thailand, Vietnam and other countries in South-east Asia.

In 2025, Malaysia and Singapore were the fastest growing markets, with +27% and +26% YoY growth respectively. Indonesia and Vietnam grew the slowest, at +11% and +12% YoY growth in 2025.

The closest public comparable is PT GoTo Gojek Tokopedia Tbk (GOTO IJ).

GOTO is GRAB’s most direct regional competitor in South-east Asia. It competes fiercely with GRAB in the key Indonesian market for mobility, food delivery, and digital payments.

Other comparables include Uber Technologies, Inc. (UBER US) and DiDi Global Inc. (DIDIY US).

UBER is the most direct global comparable due to its dual dominance in both ride-hailing (Uber) and food delivery (Uber Eats). GRAB actually acquired UBER’s South-east Asian operations in 2018.1

DIDIY is the dominant ride-hailing platform in mainland China.

My reasons

Not as cash-generative as you think

Adjusted free cash flow includes SBC. GRAB uses adjusted free cash flow to monitor business performance and assess cash flow activity.

GRAB defines adjusted free cash flow as net cash flows from operating activities less capex plus proceeds from disposal of property, plant and equipment. It excludes changes in working capital related to its banking business.

In my opinion, this definition overstates the true cash flow generated by the business because it includes stock-based compensation (SBC).

SBC is financing from employees. To simplify, GRAB recognizes SBC when it pays salaries with newly issued shares instead of cash. It is the equivalent of raising cash from employees.

This may sound too technical, but it’s actually easy to understand. Let’s use an example.

Company A pays employees with SBC. Company B issues shares and uses the cash to pay salaries.

The economic outcome is the same.

But if we follow GRAB’s definition, then Company A will report higher free cash flow.

Same economic outcome, different numbers reported.

That’s why, to show the business’ true cash-generation ability, SBC should be deducted from operating cash flow.

True free cash flow is much lower. In 2025, GRAB reported USD 290 mn of adjusted free cash flow.

If GRAB had paid its employees with 100% cash, it would have reported USD 49 mn of free cash flow. This is -83% lower than the adjusted free cash flow reported by GRAB.

Source: Angsana Anderson’s estimates based on GRAB’s 10-K

Pricing power is weaker than expected

Dominant player. Grab is dominant in almost all the countries it operates in. In food deliveries, GRAB is estimated to have ~55% gross merchandise value (GMV) market share in South-east Asia.2

But pricing power is weak. Despite its dominance, GRAB seems to struggle with raising its commission. Revenue as % of GMV represents GRAB’s commission or take-rate.

Post COVID-19 boom, GRAB’s take-rate has largely plateaued:

Source: Angsana Anderson’s estimates based on GRAB’s 10-K

UBER, GRAB’s global peer, demonstrates stronger pricing power.

It increased its Mobility take-rate from ~20% in 2021 to ~30% by 2025. During the same period, UBER raised its Deliveries take-rate from 18% to 19%.3 4

Why does UBER enjoy higher take-rate? Why can it raise its take-rate faster than GRAB?

I believe there are multiple factors.

Consumers in South-east Asia generally have lower income and are more price-sensitive. Competition is more intense. In Singapore alone, GRAB faces off against Gojek, TADA, Ryde, etc.

However, I believe the most underappreciated factor is regulatory pressure.

Regulatory pressure. In 2024, Singapore introduced the Platform Workers Act. The law became effective from 1 Jan 2025 and mandated better benefits for platform workers like Grab drivers.

GRAB mitigated the increased costs by raising its platform fees.5

On 1 May 2026, President Subianto announced Indonesia will reduce the maximum commission of ride-hailing companies from 20% to only 8%.6

GRAB later clarified this new rule will only impact 2-wheelers, which is less than 6% of their GMV.

However, this goes to show that, despite being the dominant platform, GRAB faces significant regulatory pressure. Its pricing power is constrained.

Its drivers are numerous, and are generally among the most economically vulnerable. Between “filthy rich” shareholders and “honest hardworking” drivers (and voters), which group will governments choose?

I suspect the challenge in raising take-rate is a main driver behind GRAB’s hard push to transform itself into a super-app.

GRAB acquired Jaya Grocer, a Malaysian supermarket chain, in 2022. During the same year, GRAB entered into banking services through GXS.

Other matters

Although the shares do not fit into my investment strategy, I believe Grab, as a business, will likely do well.

Anthony Tan and his team are very good at selling their vision to investors.

From nabbing Softbank as an investor to going public in Dec 2021, right before the 2022 bear market, they have demonstrated strong fund-raising abilities.

In Jun 2025, GRAB issued convertible notes. The offer size was USD 1.5 bn, which will support the business’ cash requirements for some time.

The terms were favourable. The notes carried zero interest, and the high conversion premium (40%) minimises dilution.

Factors that could lead to a re-assessment of my decision

GRAB continues to scale and generate significantly more cash than I expect
GRAB can increase take-rate significantly above current levels

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in GRAB, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Grab (2018)

The Straits Times (2026)

UBER (2025)

UBER (2022)

Channel News Asia (2026)

Reuters (2026)

Springer Nature: Not an AI-loser

Angsana Anderson — Tue, 02 Jun 2026 06:03:12 GMT

Summary

In early 2026, Springer Nature (SPG GR) was caught in the software sell-off, driven by AI disruption fears.

However, SPG faces low AI-disruption risk. SPG is in the business of supporting the creation, verification and dissemination of new knowledge. In fact, AI helps accelerate the research and publication process, improves the dissemination of research and maintains research integrity.

By and large, the transition to open access is beneficial not only to publishers but also to researchers and readers. The transition will likely remain gradual, thereby mitigating risk of sudden disruptions.

While dividends and organic investments will remain core uses of free cash flow, SPG is also open to mergers and acquisitions. Share buybacks are also possible, but they are constrained by the relatively low free float.

If BC Partners, the private equity firm and also the second largest shareholder, continues to exit, share buybacks can become a more practical and significant way to return capital to shareholders.

About (22 May 2026)

Share price: EUR 19.94

Market capitalisation: EUR 3,966 mn (USD 4,603 mn)

Enterprise value (EV): EUR 5,208 mn (USD 6,045 mn)

Average daily volume (ADV): EUR 1.4 mn (USD 1.6 mn)

NTM P/E: 12x

Background

How is generative artificial intelligence (AI) impacting the scientific publishing industry? Why are more and more researchers publishing their papers for free? What do investors usually misunderstand about Springer Nature?

On 5 May 2026, Springer Nature AG & Co. KGaA (SPG GR) released its Q1’26 earnings. The results were good.

I met Tom Waldron, the new Head of Investor Relations. He joined SPG in Sep 2025. Tom brings a wealth of experience from the buy-side and sell-side, and most recently served as Head of Investor Relations at WPP plc, one of the world’s largest advertising agencies.

In this post, I summarise the main takeaways from our conversation.

I would like to thank Tom for his time. He has reviewed this post for factual accuracy. I retained full editorial independence.

Our conversation

How will generative AI impact you?

SPG caught in the software sell-off. In early 2026, investors sold off software stocks because of AI disruption fears. SPG was caught in the sell-off.

Low disruption risk. Although there is some software element in SPG’s business, the company’s core value proposition is creating trust.

Large language models (LLM) are useful in mining old knowledge. However, SPG’s business is about the creation and verification of new knowledge.

Accelerate the research and publication process. Historically, the volume of research increases ~5% p.a. Generative AI has the potential to increase this growth by accelerating the research and publication process.

The most obvious example is translation. Researchers who are non-native English users now find it easier to publish in English and reach a wider audience.

With higher volume of research, the scientific community will need help. AI can help.

SPG has incorporated AI into their proprietary publishing platform (SNAPP). For example, SNAPP uses AI to scan the research paper’s content to suggest suitable peer reviewers.

SNAPP currently covers ~50% of their journals. Processes in academia evolve gradually. But the advantages of SNAPP and its AI tools are real, so the coverage and application of its AI tools will continue increasing. This is the first area to apply AI.

Dissemination of information. The second application of AI is around the dissemination of new knowledge.

Nature Research Assistant is Springer Nature’s AI‑powered tool that helps researchers discover, read, and write papers more efficiently.

It can generate structured summaries of articles and lets the users “chat with” a paper. The tool also surfaces related and foundational literature around a topic to speed up building a mental map of the field.

Another opportunity is licensing its huge catalogue of proprietary data and knowledge to corporate clients like R&D departments and LLM developers.

On 7 May 2026, SPG announced it is integrating directly into Claude to give institutional subscribers access to its proprietary drug development and clinical trial data. This integration is powered by a secure Model Context Protocol (MCP) server.1

The MCP server allows Claude to retrieve and ground its responses in real-time, licensed content from SPG. At the same time, it prevents the LLM from ‘ingesting’ the data. SPG maintains control over its proprietary knowledge.

There is potential for licensing deals like this to form a part of SPG’s revenue streams. People have become much more sensitive to the sources used by LLM because of their tendency to hallucinate.

However, we have not published how significant this revenue potential will be. We are taking a careful approach to licensing revenue because we want to maintain control over our proprietary data.

Maintain research integrity. The third AI application is maintaining research integrity.

After a researcher submits a paper, SNAPP uses tools and checks to help editors spot genuine errors, deliberate misconduct (including the illegitimate use of AI) and fraudulent papers created by “paper-mills”.

SNAPP runs AI-assisted integrity checks to detect manipulated images and hallucinated citations. These allow SPG to flag potential paper-mill or fabricated content early before the papers even reach human editors and peer reviewers.

There is an element of an arms race. Just as SPG is adopting advanced AI tools, the paper mills are also becoming more sophisticated.

Peer review is the final firewall that protects the scientific record, the communities of scientists involved and the reputation of SPG’s brands. In the vast majority of cases, human peer reviewers will spot fabricated papers from paper mills. Their feedback improves SPG’s AI screening process.

Because of all these mitigating factors, the level of retraction across SPG’s portfolio of journals is very low.

In 2025, SPG received over 3.1 mn submissions, of which only 539,000 articles were published. During the same year, 1,462 retractions took place. 57% (833 articles) of retractions were for papers published before Jan 2024.

Some level of retraction is natural because of genuine mistakes. Scientific methodology improves continuously. What used to be acceptable may no longer be reliable. Other times, articles can get retracted due to technicalities, such as authors making unintentional copyright errors.

How will open access impact you?

From pay-to-read to pay-to-publish. Historically, publishers were paid through journal subscriptions. As distribution became digital, access widened, but the model remained fundamentally pay-to-read.

Around 25 years ago, the industry began shifting toward pay-to-publish. In this model, articles are free for readers, while authors or their institutions pay an article processing charge (APC). This is also known as open access (OA).

The APC covers services such as peer-review coordination, editorial support and dissemination. That shift is the foundation of the open-access transition.

Open access works. Open access benefits researchers and funders because it allows research to reach the widest possible audience. It also works well for publishers because payment is tied more directly to the services they provide, such as managing peer review and dissemination rather than to limiting access behind a paywall.

Furthermore, funding increasingly comes from the broader research budget rather than narrower library budget. Research budgets have generally grown by 3% to 4% p.a., versus 1% to 2% p.a. for library budgets.

Transformative agreements are accelerating the shift. Pure open access is transactional, but transformative agreements (TAs) have become a key mechanism for accelerating adoption. SPG signed its first transformative agreement in 2015 with the explicit goal of driving open access.

Under these agreements, university consortia pay in advance for APCs, making the process more predictable and scalable. SPG now has more than 85 transformative agreements covering around 4,000 institutions.

SPG leads the open access transition partly because its subscription portfolio has long offered strong value for money. On a per article basis, SPG’s subscription fees are relatively low. That made the shift easier while still allowing SPG to be paid appropriately for the services it provides.

Revenue visibility remains strong. SPG still retains good revenue visibility during the transition. Around 60% of research revenue is contracted in advance, while roughly 40% comes from transactional sources such as APCs.

Even for transactional revenue, the roughly 200-day period between article submission and publication provides good visibility.

The transition will likely be gradual. Last year, SPG published around 53% of its articles under open access. SPG also owns two of the most cited open-access journals, Nature Communications and Scientific Reports.

In 2025, the company launched 48 new open-access journals and signed 19 new TAs. The transition is continuing, but it remains gradual rather than disruptive.

Europe is already heavily oriented towards TAs, while the US has moved more slowly.

Some large scholarly societies still earn relatively high subscription revenue per article and therefore have less incentive to change quickly to OA.

Incentives also differ across universities: research-intensive institutions may still prefer subscriptions, while less research-intensive universities may favour open access.

What are your shareholders’ plans?

Free float likely to continue increasing. The Holtzbrinck family owns ~50% of SPG. The family will likely remain a long-term shareholder.

BC Partners, a private equity (PE) firm, owned just under 50% before SPG’s IPO. The PE firm partially exited at the IPO and created the current free float.

The market expects BC Partners to continue exiting over time. If that happens, the free float should increase, which could improve liquidity and broaden the shareholder base.

How do you plan to use your free cash flow?

Free cash flow likely to continue growing in 2026. For the full year 2026, SPG expects free cash flow (FCF) to grow broadly in line with, or slightly faster than, adjusted operating profit (AOP).

In Q1’26, FCF was better than expected. This was mainly because of timing benefits from promissory notes and investments.

Deleveraging still matters. The EUR 165 mn dividend payment in Q2’26 will likely increase financial leverage slightly.

Management still sees value in reducing leverage. Lower debt supports lower interest costs and should help ahead of upcoming refinancing.

Organic investments. As discussed earlier, SPG continues to see attractive opportunities to invest in AI tools and launch new journals.

SPG also considers mergers and acquisitions (M&A). Management is also open to acquisitions, especially assets that strengthen SPG’s technological capabilities or further the company’s strategies.

Dividends remain a core use of cash. The company intends to maintain a strong balance sheet and a progressive dividend. Management has committed to paying out 50% of its adjusted net income, which should allow the dividend to grow over time. This makes the current 4% dividend yield even more attractive.

Buybacks are possible, but not yet ideal. At the upcoming annual general meeting (AGM), SPG is requesting approval to repurchase up to 10% of its shares. That said, the current free float limits what it can realistically buy back.

In the medium term, if BC Partners continues to exit and the free float rises, buybacks could become a more practical and significant way to return excess cash.

Why is CFO departing?

Orderly handover. On 1 Apr 2026, SPG announced Alexandra Dambeck has decided to step down from her role as the Chief Financial Officer (CFO) of SPG.2

The departure is amicable. The CFO is leaving to pursue a new external opportunity in a larger role. Alexandra played an important role through SPG’s IPO and has built a strong finance team to support a smooth transition.

Search is underway. The company is already looking for a successor. SPG is an attractive business, and management is confident of getting strong candidates.

The CFO plans to leave in Q4’26 and remains involved in key projects, which should give the company time to manage a stable handover.

What do investors typically misunderstand?

SPG is more AI-resistant than many assume. SPG’s core value proposition is in assisting the creation, verification and dissemination of new knowledge. Many analysts and shareholders already know this.

However, investors who are new to SPG may view the business like any other software or media businesses. Indeed, some peers may be more exposed, especially when their products rely on publicly available content.

Unless AI reaches the same level of intellect as Nobel Prize winners, it’s hard to imagine the business being disrupted by AI. If disruption happens, it will mean computers have become more sentient and intelligent than humans. There won’t be any investors left at that point.

Anyone who looks closely at SPG’s business will struggle to make a strong AI-disruption thesis. (Angsana Anderson’s note: I agree. That’s essentially why I became a shareholder.)

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I am a shareholder of SPG. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

SPG (2026)

What many investors still don't understand about P/E ratios

Angsana Anderson — Thu, 28 May 2026 16:18:42 GMT

Photo of Aswath Damodaran via Wikimedia Commons

Summary

When we apply a P/E multiple to a stock, we are essentially performing a DCF valuation with specific assumptions on growth and required rate of return.

Aswath Damodaran differentiates between valuation and pricing. Valuation is absolute while pricing is relative.

However, new students often wrongly assume that multiples meant pricing, and that it is inferior.

The NTM P/E ratio is a simplified discounted cash flow (DCF) valuation.

Background

Is Damodaran wrong?

New students of Aswath Damodaran, a finance professor at NYU, like to differentiate valuation and pricing.

They say valuation means discounted cash flow (DCF) calculations. Using P/E multiples is pricing, not valuation.

In this post, I will show that the difference is artificial.

P/E is actually a simplified DCF.

What do you mean by P/E is a simplified DCF valuation?

Let’s start from the basics.

What is P/E?

We talked about this before. P/E is calculated by dividing share price by its earnings per share (EPS). P/E of 15 times means shareholders are paying $15 for every $1 of profits earned by the company.

That’s great. You still remember.

How about DCF?

A share is worth the cash the shareholder is expected to receive in the future, discounted back to today’s value. Future cash should be discounted because $1 ten years from now is worth less than $1 today.

That’s the essence of DCF valuation.

Very good.

What is a share worth if its cash payouts grow at a steady rate forever?

You’re talking about the Gordon Growth Model. Here’s the equation:

For simplicity, let’s assume earnings convert completely into free cash flow (FCF = earnings).

Rewrite your equation using these assumptions. You can use abbreviations now.

Sure. I will use k_e to represent required rate of return for equity holders, P for value of a share today, E₁ for next twelve months (NTM) EPS and g for growth.

Great. Divide both sides by E₁.

What are you trying to get at? Just tell me the point already.

This is the final step.

After this, you’ll see the complete picture.

Ok. Here’s the final equation.

Let’s call this equation, “NTM P/E is simplified DCF”.

On the left-hand side, we have the NTM P/E ratio. Many investors like to use this. Essentially, this ratio is saying, “what is the value of every $1 of future earnings, today?”

What’s on the right-hand side?

That’s a variation of Gordon Growth Model. It is saying, “what is the value of every $1 of future earnings, today?”

Now we have it.

NTM P/E is actually a simplified DCF!

Hmmmm… I am still trying to process this. Do you have an example?

Sure.

I created an example in the Excel file below.

2026 05 28 Ntm Pe Is A Simplified Dcf

39.2KB ∙ XLSX file

Download

First, I calculated the value of a share using DCF valuation. I assumed 3% growth rate and 10% required rate of return.

Based on the DCF valuation, the value of a share today is $14.71. NTM EPS is $1.03.

The NTM P/E implied by the DCF valuation is 14x.

Ok.

Next, I used the “NTM P/E is simplified DCF” equation. Given 3% growth rate and 10% required rate of return, the shares should trade at 14x NTM P/E.

This is the same P/E implied by the DCF valuation.

The upshot is that whenever we slap a 14x P/E on a share, we are actually performing a simplified DCF valuation with 3% growth rate and 10% required rate of return.

NTM P/E is simplified DCF.

This makes more sense now. So, Damodaran is wrong?

Aswath Damodaran is not wrong. He already knows this.

For example, in pages 758 and 760 of his paper from 2005, he derived similar equations.1

Then why does Damodaran say that multiples and DCF are different?

Damodaran never said that.

He only said that valuation is different from pricing.

Valuation is determining an asset’s intrinsic value based on its underlying fundamentals. This is different from pricing, which is based on what other market participants are willing to pay for similar assets.

Since most assets derive their intrinsic value from their cash flows, DCF is the most common way to value assets.

Multiples are the most common way to price assets. Comparing P/E ratios is a quick way to gauge how expensive or cheap a company is, relative to other companies.

Essentially, valuation is absolute. Pricing is relative.

However, new students often misunderstand Damodaran. They wrongly assume that multiples meant pricing, and that it is inferior. In fact, P/E ratio is a simplified DCF.

I see. But there’s still one thing bothering me. Your key assumption is not realistic. Earnings usually do not convert 100% into FCF.

You’re right.

Because of capex and investments in working capital like inventories, a company can report $100 mn of earnings but only $50 mn of FCF.

Long story short, all else equal, a company with higher capital intensity will naturally trade at a lower NTM P/E.

I alluded to this in one of my earlier posts: The lower the P/E, the better… really?

Can you explain in more detail how capital intensity influences a company’s NTM P/E?

Let me do that in a future post.

I will also try to explain why capital-light businesses like software may be a bargain at 15x NTM P/E, but capital-intensive businesses like telecommunication may be overpriced at the same 15x NTM P/E.

Subscribe to receive my analysis immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

References

Aswath Damodaran (2005)

Australian Clinical Labs: Why I see an opportunity in Australia's pathology service provider

Angsana Anderson — Tue, 26 May 2026 00:28:19 GMT

Summary

I shortlisted Australian Clinical Labs Limited (ACL AU).

Consensus seems to be pricing little to no growth. However, the business can likely grow revenue at mid-single digit p.a.

The market may be underestimating ACL’s medium-term growth from pathology indexation, demographic tailwinds, and a recovery in GP attendance.

ACL appears to be a good business going through temporary pressure.

Despite being smaller than its closest competitor, ACL earns a superior ROA. Margin improvement suggests the recent revenue decline is driven more by portfolio clean-up than by structural deterioration.

I estimate sustainable capital returns at ~10% of market capitalisation. This is an attractive ~5% premium over Australia 10y government bond yield.

On 18 May 2026, KKR, a private equity (PE) firm, quietly disclosed it increased its stake to 8.2%, up from 7.2%.1

About (25 May 2026)

Share price: AUD 1.99

Market capitalisation: AUD 370 mn (USD 265 mn)

Enterprise value (EV): AUD 663 mn (USD 475 mn)

Average daily volume (ADV): AUD 1.5 mn (USD 1.1 mn)

NTM P/E: 11x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

ACL’s share price fell -30% in a year. KKR responds by lifting its stake to 8%.

Why are they buying more when everyone else is selling?

I believe the market is underestimating ACL’s growth potential and overreacting to short‑term headwinds.

It runs a high‑quality business with a clear path to recovery, yet it’s still priced like the problems are permanent.

In the last twelve months, capital returns were about 13% of its market cap. I believe such attractive capital return is sustainable.

Maybe that’s what KKR sees.

Business model

ACL’s financial year ends in June. Unless stated otherwise, all time references will follow the company’s financial year. For example, FY2025 refers to the year ending on 30 Jun 2025.

ACL provides pathology diagnostic services in Australia. This involves the collection, examination and testing of body tissues, fluids, and cells to analyze diseases and medical conditions.

In FY2025, the company recognised revenue of AUD 741 mn (+6% YoY). ACL earns all its revenue in Australia.

~63% of revenue comes from community pathology. ACL operates 50 National Association of Testing Authorities (NATA) accredited laboratories and 1,288 approved collection centres (ACCs) from which it collects samples for community pathology.

ACL’s ACCs are mostly co-located at medical centres, where ACL leases an area within the medical centre and provides an ACL employee to collect samples.

~15% of revenue comes from specialist outpatient and 11% from inpatient hospitals. The remaining 11% comes from commercial contracts. For example, contracts with the Australian Defence Force (ADF).

Medicare bulk billing accounts for ~70% of ACL’s revenue.

This is a payment arrangement under Australia’s national health insurance scheme where a healthcare provider directly bills the government (Medicare) for a medical service, rather than charging the patient. The patient does not have to pay anything out-of-pocket.

Because pathology is a highly regulated and consolidated industry, ACL likely sources its large-scale testing equipment and proprietary chemical reagents from major global medical device manufacturers like Roche Diagnostics, Siemens Healthineers, Abbott, or Beckman Coulter.

The closest public comparables are Healius Limited (HLS AU) and Sonic Healthcare Limited (SHL AU).

SHL is the largest player in Australia with ~51% revenue market share. I estimate HLS and ACL hold ~30% and ~18% respectively.

SHL earns ~36% of its revenue from Australia, with the remaining mainly from United States and Germany. Both HLS and ACL operate purely in Australia.

My reasons

Under-recognised growth?

Consensus is pricing no growth. The iShares Core S&P/ASX 200 ETF (IOZ AU), which tracks the 200 largest listed companies in Australia, trades on a P/E ratio ~21x.2

At 11x NTM P/E, the consensus seems to be pricing ACL for no or low growth indefinitely. Tikr shows that the consensus expects -1% revenue decline in FY2026.

Investors dismiss ACL due to an apparent lack of pricing power.

The company charges for its pathology services based on the Medicare Benefit Schedule (MBS), which in turn, is determined by the Australian government.

For more than 20 years, MBS rates for most pathology services have not been increased to keep pace with inflation (indexation).

Likely to grow at least MSD. However, ACL is likely to grow revenue at mid-single digit (MSD) p.a. over the long-term.

The government has announced that, starting from Jul 2025, certain pathology services will return to annual indexation.3 The MBS indexation factor applied most recently was ~2.4%.4

I estimate that this indexation factor benefitted ~33% of ACL’s pathology services, contributing to ~0.80 percentage points (ppt) of revenue growth in H1’26.

Investors’ focus on pricing misses the growth potential from volume and mix shift.

Since FY2017, overall benefits paid for pathology services under the Medicare system increased ~5% p.a.5 This was mainly driven by 3% volume growth. Despite no annual indexation, fees per service grew ~2% p.a. because of favourable shifts in service mix.

Even without annual indexation, the industry has grown and will likely continue growing revenue at ~MSD p.a., supported by Australia’s growing and ageing population.

People aged over 65 years are the largest users of pathology services, consuming 38% of all services in FY2020 and this population cohort is estimated to grow faster than the general population at a CAGR of 2.9% between FY2020 and FY2030.6

Good business, temporary headwinds?

Headline revenue decline is mainly driven by ACL exiting unprofitable business. ACL’s revenue declined -1.0% YoY (-AUD 3.8mn) in H1’26. At first glance, this is bad.

However, ACL explained that the revenue decline was mainly driven by the company exiting unprofitable ACCs (-AUD 8.2 mn) and unprofitable public hospital/commercial contracts (-AUD 1.4 mn).7

As a result, underlying EBIT margin improved from 7.4% in H1’25 to 7.7% in H1’26.

Near-term revenue is likely flat, driven by lower GP attendance. General practitioners (GP) are the largest referrers of pathology tests, ordering approximately 70% of Medicare funded tests.8

In Feb 2026, ACL lowered its FY2026 mid-point revenue guidance by -4%. This was largely driven by weaker-than-expected GP attendance.

The chart below shows that GP attendance in Australia remains below trend and was essentially flat in H1’26.

Source: ACL (2026)

Survey data from the Australian Bureau of Statistics (ABS) suggests this is partly because of the increasing cost of GP services.

The survey shows the percentage of people who said they avoided going to the GP at least once in a year due to the cost is now at the highest level in the 11-year period the survey covers.9

But GP attendance expected to return to growth. From 1 Nov 2025, the Australian government expanded the bulk billing incentives (BBI) to all Australians and created an additional new incentive payment for GPs that bulk bill every patient.10

As a result, the overall bulk billing rate has increased from 77% to 81% between Nov 2025 and Jan 2026. The number of registered Medicare Bulk Billing Practices increased almost +60%.11

GP attendance is expected to rise over the next six months. Because of the lag between a GP’s referral and the patient going for the pathology service, I expect the benefits to flow to ACL only after FY2026.

More efficient. FY2026 guidance downgrade overshadows the quality of the business.

At 18% revenue market share, ACL is almost half the size of HLS (~30%). Despite that, ACL earns 8% return on assets (ROA) while HLS is still suffering losses.

ACL attributes this mainly to the fact that they are the only national pathology provider with a single national Laboratory Information System (LIS) (operating system).

This reduces labour costs. Pathologists from different states can analyse samples across the nation, increasing labour utilization rate. ACL’s labour cost is only ~43% of revenue, compared to the 50%+ reported by HLS.

Capital returns?

Attractive capital returns. In the last twelve months, share buybacks amounted to ~AUD 25 mn and dividends were ~AUD 24 mn. In total, I estimate this is ~150% of free cash flow to firm (FCFF). Capital returns were ~13% of its market capitalisation.

ACL authorised its current share buyback program on 23 Oct 2025, likely in response to the share price falling below the IPO price. The company can still repurchase up to 10.4 mn shares before the program expires in Nov 2026.12

When the share buyback program ends and capital returns normalise to 100% of FCFF, I estimate the sustainable capital return will be ~10% of market capitalisation. This still offers an attractive 5% premium over the Australia 10y government bond.

Catalysts

Private equity takeover? Over the past year, ACL’s share price has fallen around -30%. On 18 May 2026, KKR, a private equity (PE) firm, disclosed it increased its stake to 8.2%, up from 7.2%.

KKR first disclosed its substantial stake (>5%) in ACL around Jul 2024.13 At that time, ACL’s share price was higher at around AUD 2.60.

KKR has shown interest in Australia’s healthcare industry. In Apr 2022, KKR led a consortium to buy Ramsay Health Care Limited (RHC AU), one of Australia’s largest private-hospital operators.14

This offers some downside protection. If ACL’s share price continues to deteriorate despite improving business performance, the probability of a privatisation should increase.

Factors to focus on

Probability and timing of recovery. The expansion of the BBI to all Australians and the introduction of additional incentive payments has increased the GP bulk billing rate.

This should help GP attendance recover. Since GPs are the single largest source of referrals to pathology services, higher GP attendance should accelerate volume growth at ACL.

Are there any factors that can break this relationship?

CEO not renewing contract. On 16 Feb 2026, CEO Melinda McGrath announced that she will not renew her contract when it expires on 30 Aug 2026. ACL is currently conducting an executive recruitment process and will announce an update when available.15

At first glance, CEO McGrath’s departure does not raise immediate red flags. The six-month notice period suggests a planned transition. It should also provide ACL sufficient time to identify and integrate a suitable successor.

That said, her resignation still creates some uncertainty. CEO McGrath has been a key part of ACL’s management over the past 10 years.

Management is important, more so in a regulated industry like healthcare. It will be important to keep an eye on who the board chooses as her successor.

Coming up next

I will be diving deeper into ACL over the next few weeks, alongside other companies on my shortlist: MegaStudyEdu Co. Ltd (215200 KS) and Xinyi Solar Holdings Limited (968 HK).

If any of them turn out to be really attractive opportunities, I will publish my thesis on Substack.

Subscribe to receive the updates immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Australian Government, Department of Health, Disability and Ageing

Disclaimer

At the time of publication, I do not hold any positions in ACL, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

ACL (2026)

Blackrock (2026)

Australian Government, Department of Health, Disability and Ageing

Topsports: 10% dividend yield. But I decided to pass. Why?

Angsana Anderson — Wed, 20 May 2026 23:31:02 GMT

Summary

I decided to pass because Topsports’ competitive advantage is deteriorating. Consumer preference is shifting towards competitors’ brands, and it is difficult to tell how far this trend can run.

Corporate governance can be improved.

Most importantly, I believe the high dividend yield is not sustainable.

About (19 May 2026)

Share price: HKD 2.93

Market capitalisation: HKD 18,170 mn (USD 2,320 mn)

Enterprise value (EV): HKD 21,650 mn (USD 2,764 mn)

Average daily volume (ADV): HKD 30 mn (USD 4 mn)

NTM P/E: 12x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Pass

Background

Topsports International Holdings Limited (6110 HK) is one of the largest retailers of sportswear in mainland China. It is the main retailer of Nike and Adidas sportswear.

Competitors like ANTA and Li Ning are trading around 14x NTM P/E and 3% dividend yield.

Meanwhile, Topsports is trading at only 12x NTM P/E and offers an attractive 10% dividend yield.

What’s the catch?

Business model

Topsports’ financial year ends in February. Unless stated otherwise, all time references will follow the company’s financial year. For example, FY2025 refers to the year ending on 28 Feb 2025.

Breakdown of FY2025 revenue (CNY 27 bn; -7% YoY):

85% Retail (-7% YoY)
15% Wholesale (-6% YoY)

Topsports is the largest distributor and retailer of NIKE, Inc. (NKE US) and adidas AG (ADS GR) products in mainland China.

As of FY2025, Topsports operates 5,020 stores, down from 6,144 in the prior year. I estimate it holds ~17% revenue market share.

Customers are mainly mass-market consumers in mainland China. No revenue derived from transactions with any single customer represent 10% or more of total revenue.

Purchases from its five largest suppliers accounted for ~95% of Topsports’ total purchases. Purchases from its largest supplier included therein accounted for ~67% of the company’s purchases. This is likely .

The closest public comparable is Pou Sheng International (Holdings) Limited (3813 HK), which also distributes Western sportswear like Nike and Adidas under the brand of ‘YYsports’. Pou Sheng is smaller. I estimate it holds ~ 12% revenue market share.

Other competitors include ANTA Sports Products Limited (2020 HK), Li Ning Company Limited (2331 HK) and Xtep International Holdings Limited (1368 HK).

These are domestic sportswear brands. I estimate they hold 45%, 18% and 8% revenue market share respectively.

My reasons

Deteriorating competitive advantage

Declining return on assets (ROA). Over the past 5 years, Topsports’ ROA averaged ~13%. This is similar to ANTA.

However, in recent years, Topsports’ ROA started diverging and is on a downward trend. In FY2025, Topsports’ ROA is 9%, much lower than ANTA’s ~12%.

Source: Angsana Anderson’s estimates based on Tikr

Topsports’ declining ROA and ANTA’s improving ROA suggest Topsports’ competitive advantage is deteriorating.

Consumer preference shifting towards domestic brands. Since FY2017, Topsports’ revenue market share has halved from around 30% to 15%.

Source: Angsana Anderson’s estimates based on ANTA, Topsports, Pou Sheng, Li Ning and Xtep data

In FY2022, Topsports’ market share suffered ~7 percentage points (ppt) drop. This was because of the Mar 2021 Xinjiang cotton incident.1

Consumers in mainland China boycotted Western brands like Nike and turned towards domestic brands. ANTA gained ~6 ppt market share. Li Ning gained ~4 ppt.

This is part of a broader trend. Consumers in mainland China increasingly prefer domestic brands.

Starbucks is losing market share to Chinese rivals like Luckin and Cotti. According to Euromonitor, Starbucks’ market share in mainland China has declined from 34% in 2019 to 14% in 2024.2

In calendar 2025, Huawei edged out Apple to reclaim the No. 1 spot in mainland China’s smartphone market.3

It is difficult to tell how far this shift in consumer preference will go on.

Corporate governance

Concentration of power. Yu Wu is the CEO, board chairman, member of the nomination & remuneration committee and is also a substantial shareholder.

Key management pay rising as profits fall. In FY2022, compensation to key management personnel is ~4.4 bps of revenue. By FY2025, it increased to 5.3 bps.

Over the same period, Topsports’ operating profits halved.

Limited disclosures. Despite being significantly impacted by the 2021 Xinjiang cotton incident, I could not find any management discussion over this matter in their annual reports.

Furthermore, I am surprised there has been no impairment of goodwill. Market share and operating profits have declined significantly, and there seems to be little visibility over recoverability.

High dividend yield likely not sustainable

Dividends funded by borrowings. In the past 5 years, operating cash flow totalled CNY 21.6 bn.

After capex, debt repayment and other business needs, there was actually a deficit of – CNY 3.5 bn. Topsports borrowed CNY 15.5 bn and used that to fund CNY 12.2 bn of dividends.

In the last twelve months (LTM), net debt has more than doubled to reach CNY 3.2 bn (2x of operating profits).

Topsports has already begun to reduce dividends. In FY2025, dividends per share declined -22% from CNY 0.36 to CNY 0.28.

Factors that could lead to a re-assessment of my decision

Strong evidence that the unfavourable shift in consumer preference will reverse.
Improvement in corporate governance (e.g. more transparent disclosures, separation of CEO and Chairman roles).

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in 6110, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

BBC (2021)

Reuters (2025)

SCMP (2026)

lululemon: A quality compounder on sale?

Angsana Anderson — Tue, 19 May 2026 07:17:41 GMT

Summary

I decided to pass because international growth looks more fragile than expected.

Near-term demand still looks weak.

Corporate governance is unstable, making any turnaround look unlikely.

About (18 May 2026)

Share price: USD 120.26

Market capitalisation: USD 13,710 mn

Enterprise value (EV): USD 13,701 mn

Average daily volume (ADV): USD 346 mn

NTM P/E: 10x

Time spent: ~2 days

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Pass

Background

lululemon: A quality compounder on sale?

lululemon was a ‘quality compounder’.

Since 2013, it has always traded around 20x NTM P/E. At its peak during COVID-19 lockdowns, its NTM P/E breached 80x.

Today, lululemon athletica inc. (LULU US) is selling for only 9x NTM P/E. This is despite its +29% revenue growth and 40% operating profit margin in mainland China.

Could this be the bargain of our lifetime?

Green/red: LULU; Purple: SPY (SPDR S&P 500 ETF Trust)

Business model

LULU’s financial year ends on the Sunday closest to January 31 of the following year. Unless stated otherwise, all time references will follow the company’s financial year. For example, FY2025 refers to the year ending on 1 Feb 2026.

Breakdown of FY2025 revenue (USD 11 bn; +5% YoY):

63% Women’s apparel (+5% YoY)
24% Men’s apparel (+4% YoY)
13% Accessories and others (+8% YoY)

LULU sells pants, shorts, tops, and jackets designed for a healthy lifestyle including athletic activities such as yoga, running, and training.

In FY2025, 71% of revenue comes from the Americas, 16% from Mainland China and 14% from the rest of the world.

Mainland China is the fastest growing region, with +29% YoY growth, followed by the rest of the world (+16% YoY). Revenue from the Americas shrank -1% YoY.

There is some supplier concentration. In FY2025, 48% of fabrics were produced by LULU’s top ﬁve fabric suppliers, with the largest manufacturer producing 20%. During 2025, 34% of the company’s fabrics originated from Taiwan.

Public comparables include Nike, Inc. (NKE) and adidas AG (ADS).

My reasons

International growth at risk

The Americas shrinking, mainland China growing. In FY2025, revenue from the Americas declined 1% YoY, driven by -3% decline in comparable sales. Operating profit margin (OPM) declined from 38% in FY2023 to 33% in FY2025.

Sales in mainland China grew +29% YoY, driven by +20% growth in comparable sales. OPM improved from 35% in FY2023 to 40% in FY2025.

The bulls argue that even though the Americas is shrinking, LULU is still undervalued because its international segment, especially mainland China, shows tremendous growth potential.

Profits attracting competition. However, I am less optimistic. It is no secret that mainland China is an extremely competitive market. You have probably heard of 内卷 (involution), a self-defeating cycle of excessive competition where companies invest more effort and resources for diminishing returns.1

40% OPM is very high in the industry. It is bound to attract competition.

Vuori, LULU’s fierce competitor in the US, entered China in 2024 and is expanding its store network to compete with LULU. Another competitor, Alo, is planning flagship stores in Beijing and Shanghai, slated for the second half of calendar 2026.2 ANTA Sports Products Limited (2020 HK) is repositioning Maia Active in 2026 as a more direct domestic challenger to LULU.3

Limited patents. It is not clear to me that LULU can fend off these challengers.

In its latest 10-K, LULU warns that “The intellectual property rights in the technology, fabrics, and processes used to manufacture our products generally are owned or controlled by our suppliers and are generally not unique to us.”

LULU adds, “We hold limited patents and exclusive intellectual property rights in the technology, fabrics or processes underlying our products. As a result, our current and future competitors are able to manufacture and sell products with performance characteristics, fabrics and styling similar to our products.”

Eclat Textile Co., Ltd. (1476 TT) is a top supplier. It helped LULU develop its signature Luon fabric, as well as newer materials. In 2017, Bloomberg reported Eclat is helping Amazon.com, Inc. (AMZN US) to make a new line of private-label athletic wear.4 Chip Wilson, LULU’s founder, reportedly regretted not buying Eclat back in 2011.5

These competitive pressures will likely only show up in the long-term. In the near-term, however, demand still looks weaker than expected.

Near-term demand still looks weaker than expected

Inventory days stabilized. At first glance, near-term demand seems to have steadied. Average inventory days appear to have stabilized at around 119 days.

But driven by heavier discounting and advertising. LULU increased its advertising expense from 4.5% of revenue in FY2023 to 5.6% in FY2025.

LULU reported credit card affiliate programs, a form of discount for credit card users, helped drive higher e-commerce traffic in the Americas. FY2025 seems to be the first year LULU used such programs.

The heavier discounts and advertisements contributed to overall OPM declining from 23% in FY2023 to 20% in FY2025.

It seems to me the stabilization of inventory days was driven by heavier discounts and advertisements rather than significant improvement in demand.

Absent a demand recovery, LULU faces steeper-than-expected margin compression to complete normalising its inventory levels.

No strong inflection in contract liabilities. LULU’s contract liabilities mainly consist of unredeemed gift cards. Historically, the YoY % change has been a leading indicator of revenue growth:

Source: Angsana Anderson’s estimates using Tikr

In FY2025, unredeemed gift cards grew +3%, up from +1% during the prior year.

However, the cash inflow from unredeemed gift cards declined -3% in FY2025, suggesting the underlying demand remains weak. The +3% increase in unredeemed gift cards was likely due to timing effects (e.g. customers delaying the redemption of their gift cards).

Corporate governance looks unstable

Board needs more suitable experience. Out of the 11 board members, only 4 have direct operational experience in apparel or fashion.

In FY2020, LULU acquired Mirror, a maker of in-home fitness equipment.6 LULU paid USD 500 mn, about 71% of the prior year’s operating income. By FY2022, LULU had written down the value of its investment by USD 443 mn.7

Out of the 10 board members that oversaw LULU at the time of the acquisition, 5 remained on the board of directors today.

I suspect this is one of the reasons why Chip Wilson, LULU’s founder, is fighting a public proxy battle to install new board members.

New CEO is unproven. After CEO Calvin McDonald stepped down in Jan 2026, the board appointed Heidi O’Neill as the new CEO.

According to Piper Sandler, an investment bank, “Investors likely remember O’Neill most as the key part of John Donahoe’s tenure (2020-2024) that was marked by a push towards DTC as well as declining innovation, strained retail partnerships and a falling share price”8

I am surprised the board did not appoint a CEO with proven track record in turnarounds.

Activist investor Elliott Investment Management suggested Jane Nielsen, the former CFO and COO of Ralph Lauren. Nielsen built her reputation when she helped turn around Ralph Lauren Corporation (RL US).9

Difficult turnaround. Retail is a tough business, especially fashion retail. It’s hard to gain any sustainable structural advantage. If you are doing well, competitors can walk into your stores and copy your innovation. That’s an oversimplification, but you get my idea: management is very important.

With LULU’s founder fighting a high-profile proxy battle and an aggressive activist investor at the gates, management may struggle to find enough time and resources to turnaround the business.

Factors that could lead to a re-assessment of my decision

Strong evidence of durable competitive advantage (e.g. patented fabric)
Early indicators of recovery in demand
Improved corporate governance (e.g. management with proven track record in turnarounds)
Favourable resolution of the proxy battles

Coming up next

Topsports International Holdings Limited (6110 HK) looks like a bargain.

9.6% dividend yield. 9x NTM P/E.

It is one of the largest distributors and retailers of Adidas and Nike sportswear products in mainland China, with revenue market share of ~17%.

Why is a business like this selling so cheaply?

In my next post, I will share my analysis.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

CommonWealth Magazine (2024)

Disclaimer

At the time of publication, I do not hold any positions in LULU, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

CNA (2025)

Jing Daily (2026)

Bloomberg (2017)

Bloomberg (2020)

Bloomberg (2023)

Bloomberg (2026)

Xinyi Solar: Poised for another upcycle?

Angsana Anderson — Fri, 15 May 2026 14:31:32 GMT

Summary

I decided to shortlist.

Xinyi Solar seems to be approaching a cyclical inflection.

The industry has been cutting capex. Demand likely to grow above expectations. Gross margins are bottoming out. Working capital shows favourable trends.

About (14 May 2026)

Share price: HKD 3.09

Market capitalisation: HKD 28,264 mn (USD 3,608 mn)

Enterprise value (EV): HKD 42,668 mn (USD 5,447 mn)

Average daily volume (ADV): HKD 185 mn (USD 23 mn)

NTM P/E: 14x; LTM P/B: 0.9x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

Xinyi Solar is the world’s largest maker of solar glass.

In the most recent shortage, its share price jumped almost 5x and peaked around HKD 20.

Today, it trades at HKD 3. The collapse reflects the severe glut in the solar industry.

Most investors have written off another upcycle.

But there are some signs that the company has reached the cycle trough and may be approaching an upcycle.

In this post, I will explain these in detail.

Business model

Xinyi Solar makes the glass that covers solar panels.

Xinyi Solar holds ~30% worldwide market share by production capacity. Its closest competitor, Flat Glass, holds ~20% market share.1

Breakdown of FY2025 revenue (CNY 20,861 mn; -5% YoY):

85% Solar Glass (-5% YoY)
14% Renewable Energy (-1% YoY)
1% Others (-57% YoY)

The Solar Glass segment produces the glass that covers solar panels. Solar glass is designed specifically to protect solar cells from harsh weather while maximizing the amount of sunlight that passes through.

The Renewable Energy segment develops and operates solar and wind power farms. It generates revenue by selling the electricity to local power grids and industrial users, supplemented by government tariff subsidies.

In 2025, Xinyi Solar earned 71% of its revenue from Mainland China, 20% from other parts of Asia, 7% from North America and Europe and 2% from others.

North America and Europe is the fastest growing region (+123% YoY), followed by other parts of Asia (+25% YoY). Mainland China declined -15% YoY.

Xinyi Solar’s main customers for its solar glass are major global solar panel manufacturers such as LONGi Green Energy Technology Co., Ltd. (601012 CH) and JinkoSolar Holding Co., Ltd. (JKS US).

Sales to the top 1 customer contributed 12% to total revenue while sales to the top 5 customers contributed 44%.

The closest public comparable is Flat Glass Group Co., Ltd. (6865 HK), the second largest solar glass maker in the world.

My reasons

Cyclical inflection?

Capex cuts. Cash capex in the industry peaked in 2022 and has been trending down ever since.

In 2025, Flat Glass’s cash capex (green) declined to CNY 3,000 mn from CNY 5,000 mn. The chart could not display this. Source: Tikr (2026)

Capex in 2026 will likely be lower too. As of 31 Dec 2025, Xinyi Solar’s contracted capex for 2026 declined by -36% YoY.

Consensus often underestimates demand. It is no secret that the solar industry is in a severe oversupply.

Not too long ago, I casually mentioned the solar industry might be interesting, only to be curtly cut off by a portfolio manager, “Do you know how many people lost their job because of solar companies?”

The media likes to quote the International Energy Agency (IEA)’s report that “In 2024, there was sufficient manufacturing capacity to have produced more than twice as many solar PV modules as were actually deployed”2

However, the consensus often underestimates demand growth. The orange hairy lines below illustrate this.

Source: The Economist (2024)

In 2025, the growth in annual installation decelerated to +12% YoY.3

However, there are two reasons why demand growth will likely re-accelerate.

Higher-than-expected oil prices. The Economist argues that investors are underestimating the crisis in the oil markets. “Even in the most optimistic scenario, the world will continue to suffer a severe shortfall in supply for several more months.”4

If the Economist is right, solar energy will become more attractive. In fact, even before this crisis, solar was already cheaper. In 2025, solar energy was 41% cheaper on average than the lowest-cost fossil fuel alternatives.5

Already, we are seeing early signs of acceleration in exports:

Source: Ember (2026)

To be fair, the expiration of export tax rebate in Apr 2026 also contributed to the export surge in Mar 2026. Data for Apr 2026 will give a clearer picture.

Higher for longer oil prices will likely re-accelerate the demand for solar panels.

Declining cost of battery energy storage systems (ESS). The penetration of solar energy has historically been curtailed by its lack of dispatchability. Grid operators cannot easily adjust solar output to match demand because electricity is only generated when the sun is shining.

As a result, solar cannot act as a reliable baseload power source. This is a huge disadvantage. Natural gas or diesel power plants can quickly ramp their electricity production up or down to match fluctuating market needs.

The obvious solution is to store the solar energy and release it according to electricity demand. In the past, this was not very commercially feasible. Battery ESS is expensive.

However, the declining cost of battery ESS is starting to change this.

Source: Energy Storage News (2025)

In May 2026, the International Renewable Energy Agency (IRENA) reported that, in certain regions, solar energy with battery storage is already delivering cost-competitive electricity compared with coal and gas.6 IRENA expects the gap to widen further.

All these factors suggest demand for solar panels will likely re-accelerate. Coupled with capex cuts, the current oversupply in the solar industry will likely end sooner than expected.

Gross margins bottomed out. Already, we are seeing some early signs of cyclical inflection. Gross margins for both Xinyi Solar and Flat Glass reached all-time low in 2024 before recovering slightly in 2025.

Gross margins appeared to have bottomed out. In Q1’26, Flat Glass’ gross margin improved to 14%, up from 12% a year ago. Xinyi Solar has not reported Q1’26, but will likely show a similar improvement.

Flat Glass’ gross margin (green) also recovered in 2025, reaching 16%, up from 15% in 2024. The chart could not display this. Source: Tikr (2026)

Favourable working capital trends. Xinyi Solar’s working capital trends suggest an inflection point may be approaching.

After declining -30% in 2023, contract liabilities (advance payment from customers) grew +3% in 2024. The growth accelerated to +14% in 2025. Although the absolute amount is immaterial, the trend in contract liabilities has historically been a good leading indicator of revenue growth.

Receivable collection improved in Q4’25, with average receivable days declining from a near all-time high in the prior quarter.

When customers start settling their bills and even pay for some orders in advance, an inflection is likely approaching.

Finally, inventories decreased -40% in 2025, much faster than the -5% decline in revenue. The inventory drawdown was concentrated in finished goods.

Factors to focus on

Where are we in the solar cycle? Xinyi Solar’s share price trades in line with LONGi, one of the world’s largest solar wafer and panel makers.

Source: Tikr (2026)

To invest in Xinyi Solar over the long-term, it is important to know where we are in the broader solar industry cycle.

The solar cycle seems to share similarities with the semiconductor cycle. Using the same framework, back in 2024, I created this chart for Jinko Solar.

Source: Angsana Anderson’s estimates using Jinko Solar

The average selling price (ASP) and unit cost of JKS’ solar panels seem to offer enough explanatory power. It will be worth exploring this further in a future post.

Long overdue receivables. Tariff adjustment receivables make up ~60% of Xinyi Solar’s total trade receivables.

These receivables represent government subsidies on renewable energy generated by the company’s solar farms.

Xinyi Solar disclosed that 52% of its tariff adjustment receivables pertain to electricity generated over 2 years ago (!).

The government is apparently delaying subsidies payments because of a mounting deficit in China’s renewable energy development fund (REDF). The REDF is funded by renewable surcharge fees levied on consumers. However, considerable political pressure to keep electricity cheap meant difficulty in raising the surcharge fees.7

Such long overdue receivables, even from the government, create a high risk that Xinyi Solar will not receive the cash. A write-down may be necessary. A complete write-down would cost around CNY 4,900 mn, ~135% of 2025’s operating profits.

Until the company starts receiving payments, it may also be prudent for investors to disregard the revenue recognized from tariff adjustments. Such revenue accounted for ~5% of total revenue in 2025.

Perhaps Flat Glass Group Co., Ltd. (6865 HK) offers a better way to invest in the solar industry. Its trade receivables appear healthier because it focuses on making solar glass rather than solar farms.

Coming up next

Lululemon: Lulu or lemon?

Lululemon was a quality compounder. Since 2013, it has always traded around 20x NTM P/E. At its peak during COVID-19 lockdowns, its NTM P/E breached 80x.

Today, Lululemon is selling for only 9x NTM P/E.

Could this be the bargain of our lifetime?

Let’s find out in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in 968, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Energy Iceberg (2019)

ThaiBev: After falling -50%, is it time for me to buy?

Angsana Anderson — Tue, 12 May 2026 15:14:05 GMT

Summary

I decided to pass because near-term demand looks weaker than expected, but capex continues to increase.

While management has a great track record in Thailand, their capital allocation outside Thailand has delivered disappointing returns so far.

About (11 May 2026)

Share price: SGD 0.43

Market capitalisation: SGD 10,681 mn (USD 8,405 mn)

Enterprise value (EV): SGD 21,207 mn (USD 16,688 mn)

Average daily volume (ADV): SGD 8 mn (USD 6 mn)

NTM P/E: 10x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Pass

Background

It’s 2016. You’re lying on the beach in Phuket. Hot sweltering heat. Cold refreshing beer. Paradise on earth.

Back in your office, you discovered ThaiBev brews your favourite beer in Thailand. Chang and its rival Singha form a cosy duopoly.

Your heart skipped a beat when you realised ThaiBev is listed in Singapore. Despite its seemingly expensive NTM P/E of 20x, you bought a significant block.

After all, you know the product well. The company has steadily grown its revenue by ~6% p.a. over the past 10 years.

Surely, 20x P/E is not that expensive for such a quality stable stock?

Fast forward 10 years, you’ve suffered a total loss of -20%. What happened?

More importantly, with its NTM P/E at an all-time low of 9x, is ThaiBev finally cheap enough to buy?

Business model

ThaiBev’s financial year ends in Sep. Unless stated otherwise, all time references will follow the company’s financial year. For example, Q2’26 refers to Jan 2026 – Mar 2026.

ThaiBev is the largest beverage company in ASEAN by sales.

The company holds ~90% volume market share in Thailand’s spirits market. Through its beer brands like Chang and Bia Saigon, ThaiBev holds a strong market share of 35% to 40% in Thailand and Vietnam.1

Breakdown of FY2025 revenue (THB 333 bn; -2% YoY):

37% Beer (-2% YoY)
36% Spirits (-2% YoY)
19% Non-alcoholic beverage (-1% YoY)
7% Food (-2% YoY)
1% Others (-11% YoY)

The Beer segment produces and sells branded beer such as Chang beer, one of Thailand’s most popular beers.

The Spirits segment produces and sells branded spirits such as Hong Thong and Ruang Khao. These are the best-selling brown and white spirits in Thailand, respectively.

The Non-alcoholic beverages segment produces and sells branded water, ready-to-drink coffee, energy drinks, green tea, fruit flavored drinks and dairies product. Brands include Fraser & Neave (F&N) and Oishi.

The Food segment operates restaurants such as KFC Thailand and Oishi Sushi.

In 2025, 65% of revenue comes from Thailand, 15% from Vietnam, 7% each from Malaysia and Myanmar.

Main customers include distributors and retailers. ThaiBev did not disclose any major customers.

Public comparables include Emperador Inc. (EMI PM) and Carabao Group Public Company Limited (CBG TB).

EMI is a Philippines‑based global spirits company that produces brandy and Scotch whisky brands such as Emperador, Fundador, The Dalmore and Jura.

CBG is a Thai beverage company that produces Carabao‑branded energy drinks and related beverages.

My reasons

Demand looks weaker than expected

Declining spirits. In FY2025, the Spirits segment recorded a -2% YoY decline in revenue. Sales volume declined by -3%. This was partially offset by +1% increase in average selling price (ASP).

FY2025 is the third consecutive year of sales volume decline. In fact, Spirits sales volume has never recovered to its peak in FY2020.

Management denied this was a structural issue. They attributed the weak volume to a weak post-COVID economy and high household debt levels resulting in weak consumer purchasing power.2

I am not fully convinced. Since 2020, Thailand’s real annual GDP growth rate has recovered and hovered around low-single-digit.

Near-term demand looks weaker than expected. Consensus forecasts revenue growth will recover from -2% in FY2025 to +3% in FY2026. However, working capital trends suggest near-term demand remains weak.

Average inventory days increased for the second consecutive year and is now above 10-year average.

The increase was concentrated on finished goods. Coupled with the revenue decline, this suggests demand was weaker than expected in FY2025.

At the same time, raw materials were flat. Management likely does not see demand recovering in FY2026.

But capex continues to increase

Capex reached all-time high. In the past 10 years, capex averaged 3% of revenue. It started increasing from FY2023 and reached THB 18 bn in FY2025 (~6% of revenue). ThaiBev allocated 40% of the elevated capex to the non-alcoholic beverage segment and 20% to the beer segment.

I estimate the bulk of the capex for non-alcoholic beverage went into F&N AgriValley, a dairy farm in Malaysia. This is a ~THB 25 bn project that will see up to 20,000 dairy cattle when completed.

But demand is not obvious. The dairy industry is quite mature in Malaysia. Roland Berger, a consultancy, forecasts low growth over the next 10 years.3 Milk consumption per capita is already among the highest in Southeast Asia.4

Management explained, “It’s more that Malaysian actually consume a lot of recombined milk, right, not the fresh milk. So our goal is to provide good access to quality fresh milk for Malaysian people.”

Competition is strong. The market is dominated by incumbents such as Dutch Lady Milk Industries Berhad (DLADY MK) and Farm Fresh Berhad (FFB MK).

DLADY is the market leader with ~28% overall market value share. It has been operating in Malaysia since 1963. FFB started around 2009. It grew quickly by focusing on chilled fresh milk, eventually dominating the segment with ~60% market share.

Mixed record in capital allocation

Great track record in Thailand. The Sirivadhanabhakdi family showed extraordinary management skills. They first launched Chang beer in 1995. Through low price, good timing, and aggressive business practices, Chang has overtaken Singha, the incumbent beer brand.5

Not so great outside Thailand. However, outside of Thailand, their track record is mixed.

In 2007, ThaiBev acquired 54% of Saigon Beer - Alcohol - Beverage Corporation (SAB VN) for USD 6.5 bn.6 SAB is Vietnam’s largest beer brewery.

Today, ThaiBev’s stake is worth only USD 1.2 bn. I estimated ThaiBev collected around USD 0.6 bn of dividends from SAB.

It is difficult to calculate the exact cost per share that ThaiBev paid for Vietnam Dairy Products Joint Stock Company (VNM VN).

However, the fair value of its stake in VNM is only 40% of its carrying value. This suggests ThaiBev likely is incurring a loss on this acquisition too.

Risk of write-down. So far, ThaiBev has not written down the carrying value of its stake in VNM.

I suspect this is because they believe the value in use of VNM remains above the carrying value.

However, if the fair value of VNM remains significantly below the carrying value, I would view this assumption as increasingly optimistic.

Factors that could lead to a re-assessment of my decision

Spirits volume recovers better than expected
Inflection in working capital trends that suggest near-term demand will be better than expected
Better-than-expected returns on capex in Malaysia
Turnaround in VNM and SAB
Free cash flow to firm (FCFF) yield on enterprise value increases to ~7.2% (5% premium over Thailand 10y government bond yield). Current FCFF yield is ~6.0%.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Singapore Business Review (2017)

Disclaimer

At the time of publication, I do not hold any positions in THBEV, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

All About Beer (2003)

Duolingo: Is it AI-resistant?

Angsana Anderson — Fri, 08 May 2026 15:25:53 GMT

About (07 May 2026)

Share price: USD 113.61

Market capitalisation: USD 5,293 mn

Enterprise value (EV): USD 4,134 mn

Average daily volume (ADV): USD 293 mn

NTM P/E: 16x

Time spent: 1 day

My decision

Pass

Background

In Apr 2015, Tim Cook launched the Apple Watch.

This looked like a huge threat to Pebble, a smartwatch pioneer.

Eric Migicovsky, the CEO of Pebble, brushed these concerns aside. Migicovsky revealed that Pebble’s sales continued to increase, experiencing a double-digit percentage growth. “Apple Watch had no material impact.”

By Dec 2016, Pebble had shut down.

In Nov 2022, OpenAI released ChatGPT.

The bulls argue generative AI cannot replace the gamified and social experience Duolingo offers. “Look at how DUOL grew its revenue by +41% p.a. since 2022!”

Is Duolingo really AI-resistant?

I am not sure.

In my latest post, I explain why I believe generative AI threatens Duolingo’s business model, but not in the way that most investors think. I also highlight some early warning signs.

Business model

Breakdown of 2025 revenue (USD 1,038 mn; +39% YoY):

84% Subscription (+44% YoY)
8% Advertising (+45% YoY)
4% Duolingo English Test (-8% YoY)
4% In-App Purchases (+5% YoY)

DUOL earns subscription revenue mainly through two products: (a) Super Duolingo and (b) Duolingo Max. These products unlock additional features such as Video Call, where subscribers can practice speaking with Lily, an AI persona.

Subscribers pay in advance. DUOL offers monthly or annual subscriptions. DUOL also offers an annual family plan, which includes up to six users on one subscription.

DUOL also earns revenue from in-app purchases (IAP). Users can purchase consumable in-app virtual goods such as ‘energy refills’, which let users continue lessons once they run low on ‘energy’.

DUOL’s main customers are individual language learners. It is especially popular among Gen Z and young millennials. As of 2025, DUOL reported 133 mn monthly active users (MAU), 53 mn daily active users (DAU) and 12 mn paid subscribers.

In 2025, 38% of revenue comes from the United States. This is down from 45% in 2023. The remaining 62% comes from the rest of the world. No other country accounted for more than 10% of revenue.

Public comparables include Chegg, Inc. (CHGG US), Coursera, Inc. (COUR US) and Udemy, Inc. (UDMY US).

DUOL delivers gamified, bite-sized language lessons. COUR and UDMY both focus on professional upskilling. COUR offers formal university credentials while UDMY hosts a crowdsourced marketplace of independent courses.

Finally, CHGG is a student homework-help platform. To survive AI disruption, it is currently pivoting toward career training.

My reasons

Not 100% AI-resistant

Prima facie, DUOL appears AI-resistant. In Nov 2022, OpenAI released ChatGPT. Many EdTech companies suffered.

CHGG’s revenue has declined -21% p.a. since 2022. Its main products were paid homework help and study support. CHGG charged students a subscription fee to see worked solutions, ask questions to human “experts,” and access study materials.

These were suddenly matched or replaced by AI tools that are free, instant, and always available. Students could paste the same questions into ChatGPT and get reasonably good explanations without paying. Even search engines like Google began answering questions directly with AI instead of sending traffic to CHGG.

In contrast, over the same period, DUOL grew its revenue by +41% p.a. It hasn’t suffered the same pain mainly because it offers a gamified and social experience to learning new languages.

One of the most important features is streak. It tracks the number of consecutive days a user has completed at least one lesson. As the streak grows, it transforms into a digital trophy that users are emotionally invested in and afraid to lose.

Streak creates a strong anchor that drives consistent daily engagement. DUOL even offers “Friend Streaks” with up to five people, which increase only when all participating users complete their daily lessons.

Prima facie, there seems to be strong evidence that AI does not impact DUOL.

ChatGPT’s chatbot interface is great for answering homework questions. That’s why CHGG was significantly affected. However, the chatbot interface is not great at replicating DUOL’s gamified and social experience of learning languages.

However, I believe DUOL’s gamified and social experience may only delay the impact of AI. It does not completely remove the AI-disruption threat.

In other words, it seems to me that DUOL is not 100% AI-resistant.

But AI has lowered the barriers to entry. Before generative AI, the language learning app market was dominated by a small set of incumbents that had been around for years. The major players were all well-established:

Source: Perplexity (2026)

There were not many new entrants. Designing lessons was labour intensive and expensive.

It all changed when OpenAI launched ChatGPT in Nov 2022. Large language models (LLMs) suddenly made it possible for startups to build convincing AI tutors at low cost, triggering a wave of new entrants:

Source: Perplexity (2026)

Early signs of weakness

Bookings growth decelerated. On 4 May 2026, DUOL released their Q1’26 earnings. Their share price dropped as much as -10% mainly because management guided a weak bookings growth in Q2’26.

Are we starting to see the negative impact from increased competition enabled by generative AI?

“Nothing to worry about!”, the bulls argue. Management expects bookings growth to recover and accelerate during the second half of 2026.

I am not so sure of recovery. Humans tend to extrapolate the status quo and fail to recognise paradigm shifts.

In Apr 2015, Tim Cook launched the Apple Watch. This looked like a huge threat to Pebble, a smartwatch pioneer.

Eric Migicovsky, then CEO of Pebble, publicly downplayed these concerns, noting that Pebble’s sales were still growing at a double‑digit rate.

According to Betanews, he said the Apple Watch had ‘no material impact’ on the company and attributed this to differences in their target markets.1

“Pebble is dealing with the threat of Apple’s upcoming Apple Watch by making jokes about the enthusiasm Apple employees have shown for the device.” Source: MacRumors (2014)

In just over a year after Apple Watch launched, Pebble shut down.2

Corporate history is filled with victims of technological change. Pebble is just one of them. Others include Sound Blaster developed by Singapore’s Creative Technologies3 and WordPerfect4, the predecessor to Microsoft Word.

DUOL does not sell watches, sound cards or word processors. I highlighted these examples to show that in most cases, people failed to recognize the paradigm shift until it was too late.

User engagement may have declined. In DUOL’s case, I see a few early warning signs.

Revenue from in-app purchases (IAP) declined -11% YoY in Q1’26. IAP revenue comes from users buying consumable virtual goods such as ‘energy refills’.

This decline is inconsistent with the supposedly higher user engagement suggested by the improvement in DAU/MAU ratio.

High short interest. There is increasingly more short selling of DUOL. By 15 Apr 2026, the short interest in DUOL has increased from 3.7% to 20.0%. Such extraordinarily high short interest suggests high conviction on the part of short sellers.

Factors that could lead to a re-assessment of my decision

New entrants fail to gain meaningful market share in the next 3 to 5 years.
Bookings growth accelerates more than expected.
Sustained growth in IAP revenue.

Coming up next

ThaiBev: When will the hangover stop?

It’s 2016. You’re lying on the beach in Phuket. Hot sweltering heat. Cold refreshing beer. Paradise on earth.

Back in your office, you discovered Thai Beverage Public Company Limited (Y92; THBEV SP) is the brewer of your favourite beer in Thailand. Chang and its rival Singha form a duopoly that dominates the beer market.

Your heart skipped a beat when you realised ThaiBev is listed on the Singapore Stock Exchange. Despite its seemingly expensive NTM P/E of 20x, you bought a significant block.

After all, you know the product well. The company has steadily grown its revenue by ~6% p.a. over the past 10 years. Surely, 20x P/E is not that expensive for such a quality stable stock?

Fast forward 10 years, you’ve suffered a total loss of -20%.

What happened?

More importantly, with its NTM P/E at all-time low of 9x, is ThaiBev finally cheap enough to buy?

I’ll discuss this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I do not hold any positions in DUOL, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

Betanews (2015)

Business Insider (2016)

Asianometry (2026)

W E Pete Peterson (1998)

MegaStudyEdu: Why is South Korea’s top online school trading at only 5x P/E?

Angsana Anderson — Tue, 05 May 2026 18:38:32 GMT

About (05 May 2026)

Share price: KRW 46,800

Market capitalisation: KRW 485 bn (USD 328 mn)

Enterprise value (EV): KRW 450 bn (USD 305 mn)

Average daily volume (ADV): KRW 2 bn (USD 1.3 mn)

LTM P/E: 5x

Time spent: ~1 day

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

Shortlist

Background

“Every year in November, Suneung brings the whole country to a standstill.

Silence descends across the capital Seoul as shops are shut, banks close, even the stock market opens late. Most construction work halts, planes are grounded and military training ceases.

Occasionally the stillness is broken by distant sirens - police motorbikes racing to deliver students running late to their exam.”

That’s how the BBC described CSAT (Suneung).

CSAT is South Korea’s most important exam. A one-day exam will decide your university, your career, your social standing and even who you date.

With such high stakes, many turn to private educators. MegaStudyEdu is one of the top 3 CSAT prep providers in the country.

It boasts a deep bench of star tutors like Stanford-educated Hyun Woo-jin. He reportedly earns up to USD 20 mn a year.1

Yet, MegaStudyEdu’s shares present a puzzling disconnect. Its dividends and share buybacks amount to ~12% of its market capitalisation. It looks likely to sustain its revenue in 2026.

But it trades at only 5x P/E.

Could MegaStudyEdu be an opportunity?

In my latest Substack post, I’ll explain why I shortlisted the company for further investigation.

Business model

MegaStudyEdu is South Korea’s educational ‘Netflix’.

It is one of the country’s largest “online cram school”. Students can watch lectures from famous star teachers, do practice questions, and get data-based advice on their chances of getting into certain universities.

The company was one of the first to focus mainly on online lessons.

It grew by hiring very popular teachers, recording their lectures once, and then selling those videos to many students at a lower price than traditional offline classes.

Today, 67% of its revenue comes from lectures, mostly online. Students pay their fees in advance. MegaStudyEdu recognises revenue over time as it provides access to the lectures.

The remaining 33% comes from selling textbooks, mock exams, study rooms, etc.

Breakdown of 2025 revenue (KRW 885 bn, -6.1% YoY):

63% from High school (-4.4% YoY)
24% from Early childhood, elementary and middle school (-5.3% YoY)
9% from University (-1.2% YoY)
4% from Employment (-34.2% YoY)

High school revenue mainly comes from selling lectures to high school students, especially those sitting for the College Scholastic Ability Test (CSAT).

According to MegaStudyEdu, it holds ~ 74% market share, up to 3x higher than the next closest competitor.2

Early childhood, elementary and middle school revenue comes from selling lectures to toddlers, elementary and middle school students.

These students are relatively younger, making it difficult for them to sustain self-directed online learning. Consequently, they tend to prefer offline academy classes where there is relatively more interaction between teachers and students.

The University segment earns revenue from providing lectures to students seeking university transfers.

MegaStudyEdu earns almost 100% of its revenue in South Korea.

The largest expense is instructor fees (21% of revenue), followed by employee salaries (16%).

The closest public comparable is Digital Daesung Co., Ltd. (068930 KS). They operate the Daesung MyMac online learning platform and physical cram schools, competing directly for high school students preparing for the CSAT.

My reasons

Under-recognised growth?

The market appears to be pricing in an indefinite revenue decline. I plotted the 10-year total return of MegaStudyEdu against Digital Daesung. As a baseline, I included EWY, an ETF tracking the MSCI Korea 25/50 Index.

Source: Tikr

MegaStudyEdu and Digital Daesung are generally uncorrelated with EWY.

Both companies rallied in 2021. COVID-19 lockdowns drove higher enrolment in online cram schools. In just one year, MegaStudyEdu more than tripled its earnings per share (EPS) while Digital Daesung grew its EPS by +52%.

The consensus extrapolated the tremendous growth and paid up for MegaStudyEdu and Digital Daesung. In 2021, LTM P/E peaked at 34x and 19x respectively.

The tremendous growth could not be sustained after the lockdowns ended. Revenue growth at MegaStudyEdu decelerated and eventually turned -6.1% in 2025. LTM P/E compressed.

Today, at 5x LTM P/E, MegaStudyEdu is priced like a terminally declining business.

However, there are reasons to be skeptical. First, the headline is overstating the decline. MegaStudyEdu disposed a loss-making subsidiary (Next Study Co., Ltd.) in 2024. If not for this disposal, the revenue decline in 2025 would be -3% instead of the headline -6%.

Second, there are signs that MegaStudyEdu is near the trough in revenue.

In 2025, prepayments from customers (contract liabilities) remained essentially flat, an improvement from -15% in 2024. This suggests revenue will likely remain flat or even grow slightly in 2026.

Capital returns?

Capital return represents ~12% of market capitalisation. In Apr 2025, MegaStudyEdu announced its corporate value-up program. It targets to return 60% of net income via dividends and share buybacks.

For 2025, MegaStudyEdu returned KRW 51 bn of capital to shareholders, representing 65% of net income. This is ~12% yield on market capitalisation.

If earnings remain stable and the company continues this policy, this will imply a recurring 11% yield on market capitalisation. This is an attractive 7% premium over South Korea’s 10-year government bond yield.

Attractive capital returns likely sustainable. I believe there is a very good chance that MegaStudyEdu can maintain such attractive capital returns.

For the reasons I discussed above, revenue looks near a trough.

MegaStudyEdu is cash generative.

The company operates with favourable working capital. It collects fees in advance from students and pays its instructors later.

Besides some physical academies, there are no significant capital expenditure requirements. Most of the lectures are streamed online.

Catalysts

On 4 Feb 2026, Maeil Business reported that the management (and major shareholders) are looking to sell 32% of the company to private equity. The sale price would value the entire company at around KRW 925 bn.3

The company denied this report on 5 Feb 2026.4

On a side note, Perplexity did not catch this company announcement. I guess generative AI still can’t completely replace me rolling up my sleeves and combing through the regulatory filings. Or maybe I should buy a more powerful model?

That said, I believe there is a good chance of an eventual sale.

In 2022, the major shareholders were negotiating a sale to MBK Partners, one of North Asia’s largest private equity (PE) firms. The negotiation eventually broke down. Apparently, they disagreed on price.

The major shareholders (and management) are near retirement age. The company has not publicly identified a clear next-generation successor.

Factors to focus on

Revenue deceleration? Is post-lockdown normalisation the complete explanation behind the recent years of revenue deceleration?

In 2025, management also attributed the -4% decline in High School revenue to ‘the trend of top-tier students preferring self-study for retaking the college entrance exam persists’.

Digital Daesung, its direct competitor, also experienced similar deceleration. However, in recent quarters, it managed to maintain low positive revenue growth while MegaStudyEdu experienced low single-digit decline. Why?

Substitution risks? The Ministry of Education (MOE) provides free online CSAT lectures via Educational Broadcasting System internet (EBSi). The purpose is to help curb private education costs.

The MOE explicitly targets a ~50% linkage rate between EBSi materials and the actual CSAT exams. As a result, MegaStudyEdu faces substitution risks.

Despite the price difference, I estimate MegaStudyEdu grew its revenue by 4.4% p.a. since 2004.

A cursory browse through DC Insider, the Reddit-equivalent in South Korea, reveals students still perceive the company’s star lecturers as superior.

AI disruption? I believe this risk is low.

The lecturers on MegaStudyEdu are not just lecturers. They are celebrities.

The lecturers not only teach, but they motivate, entertain, console and even admonish.

Upcoming events

Fri 15 May 2026: Q1’26 earnings release (expected)

Coming up next

Is Duolingo AI-resistant?

OpenAI released ChatGPT in Nov 2022.

Chegg, Inc. (CHGG US), a popular edutech company, has been on the decline ever since. Revenue first declined -1% in 2022, then -7% in 2023 before collapsing -39% in 2025. The share price is down almost -95%.

In contrast, Duolingo, Inc. (DUOL US) is going from strength to strength. Revenue continued growing. +47% in 2022, +44% in 2023 and +39% in 2025.

Why?

I will explore this in my next post.

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Korea JoongAng Daily (2023)

Disclaimer

At the time of publication, I do not hold any positions in 215200, either long or short. I may change my views, predictions, or personal portfolio positioning at any time without notice.

References

MegaStudyEdu (2025)

Maeil Business (2026)

MegaStudyEdu (2026)

Korea JoongAng Daily (2023)

Micro-Mechanics: Thesis has materialised

Angsana Anderson — Sun, 03 May 2026 17:44:54 GMT

About (30 April 2026)

Share price: SGD 3.34

Market capitalisation: SGD 464 mn (USD 365 mn)

Enterprise value (EV): SGD 442 mn (USD 347 mn)

Average daily volume (ADV): SGD 1.5 mn (USD 1.2 mn)

LTM P/E: 34x; LTM P/B: 9x

My decision

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

I decided to continue holding my shares (previous: buy).

At SGD 3.34, 5DD shares now look fairly valued.

My investment has grown to a large portion of my portfolio. As a matter of prudent portfolio management, I plan to hold my shares while looking to redeploy the capital into new opportunities.

Background

What are my key takeaways from Micro-Mechanics’ earnings call on Wed 29 Apr 2026?

The cycle has turned. Revenue growth continued to accelerate. Operating profit margins continued to expand, finally recovering to the levels reached during the previous upcycle.

Near-term demand remains healthy. Management reported their lead time continued to increase. The risks of irrational capacity expansion remain low.

That said, there are indicators that we are no longer in the early stages of the upcycle.

Inventory growth finally increased faster than revenue. My framework suggests the current semiconductor shortage may end around calendar Q1’27.

I discuss these in detail in this post.

Business model

5DD’s financial year ends in Jun. Unless stated otherwise, all time references will follow 5DD’s financial year. For example, Q3’26 refers to Jan 2026 – Mar 2026.

If Micro-Mechanics (Holdings) Ltd. (5DD; MMH SP) is new to you, check out their business model here: Micro-Mechanics: After a +97% rally, is it too late to buy?

My reasons

Cyclical inflection

No longer in the early stage of upcycle. In Q3’26, inventory increased +23% YoY. This is the second consecutive quarter of YoY increase.

Management explained the increase was mainly due to higher material costs and 5DD building up inventory to support the fulfillment of confirmed backlog orders.

That the inventory (+23% YoY) increased faster than revenue (+16%) suggests we are no longer in the early stage of the current upcycle.

Risk looks low in the near-term. Management reported their lead time continued to increase. Demand still looks strong over the next year. The risks of irrational capacity expansion appear low.

Coupled with the positive industry read-throughs and healthy channel dynamics discussed in my thesis, near-term risks appear low.

Potential downcycle from ~ Q3’27 (calendar Q1’27). For the reasons I discussed in my previous post, Where are we in the semiconductor cycle?, I believe the next semiconductor downcycle could start ~ Q3’27 (calendar Q1’27).

I will not repeat the explanation here. Instead, I will apply the framework to 5DD and show you the resulting chart.

Source: Angsana Anderson’s estimates; Tikr; Investing.com

5DD typically experiences revenue growth during memory chip shortage (when ASP flatlines or increases). These are also the periods when its share price rallies.

At the early stages of the current shortage, 5DD’s revenue growth was weaker than expected.

I suspect significant downstream inventories delayed customer orders. Inventory days at major OSATs like Amkor Technology, Inc. remained elevated during 2023 and early 2024.

When the memory chip industry flips to oversupply, ASP collapses towards cost. During these periods, 5DD’s revenue growth and share price typically decline.

In my previous post, I estimate that the current shortage will end around calendar Q1’27. If this is accurate, this means 5DD’s revenue growth could enter a downcycle from ~ Q3’27 (calendar Q1’27).

I am not betting on the exact timing. Of course, I expect my prediction to be off by one or two quarters, maybe even three.

Why, then, do I try to understand where we are in the cycle and forecast the next downturn?

It’s about positioning rather than perfect prediction.

After a huge rally, there is a natural bias to hold onto cyclical winners regardless of valuation. Recognizing that we are in the late stages of a shortage helps me overcome this bias.

It reminds me that current earnings are not sustainable, preventing me from extrapolating peak earnings.

Furthermore, this upcycle has caused semiconductor companies to grow into a disproportionately large part of my portfolio. Recognizing that a downturn is approaching pushes me to begin diversifying.

Prudent portfolio management. This is why I am looking to redeploy a portion of my capital from 5DD.

It is not because I have lost confidence in the company. On the contrary, 5DD’s management has navigated past downturns and capitalised on the current upcycle so well that the position has grown too large.

I believe 5DD has set a benchmark that many other public companies in Singapore would do well to follow.

Management continues to demonstrate strong alignment with minority shareholders. They maintain a high standard of disclosure, consistently providing important and candid insights into their business.

“A crucial test of management is whether they explain things clearly, honestly and transparently. Companies that succeed in this tend to be better run, in my experience.”

- Charlie Huggins, a former fund manager (20 lessons from 20 years of investing)

Other matters

Capex likely to increase. Since FY2022, capex has been running below depreciation. Management guided SGD 2.8 mn capex for FY2026. This will likely fall below depreciation (~ SGD 6 mn). Property, plant and equipment is now ~ 77% depreciated.

I asked whether they foresee the need to raise capex significantly above depreciation in the future.

Management explained they had been focusing on preventive maintenance. Going forward, they expect higher capex.

No loss of major customer. During the same period last year, 5DD reported 1 major customer that contributed more than 10% of revenue. 5DD did not report any major customer in this period.

Management explained there was no loss of business with this customer. The contribution from this customer fell below 10% because revenue from other customers increased.

Advance from customers not a meaningful leading indicator. Even though the amount is immaterial (~0.2% of revenue), I asked management to explain the nature of advances from customers. Does the company usually require customers to pay in advance?

5DD typically sells on credit. Sometimes, to manage credit risks in certain geographies, the company may require certain customers to pay in advance.

I believe this is why advances from customers does not show any meaningful correlation to revenue growth.

Factors that could lead me to increase investment in 5DD

Evidence that the current upcycle will be much stronger than historical cycles (e.g. current capex has not reached peak levels, the spread between ASP and unit cost of memory chip can continue to widen).
Share price falls > 30% without significant deterioration in business fundamentals.

Factors that could lead me to decrease investment in 5DD

Strong evidence that we are approaching the peak of the semiconductor cycle, and other investors are extrapolating peak earnings and overpaying for 5DD shares.

Upcoming events

Thu 27 Aug 2026: Q4’26 and FY2026 earnings release (expected)

Fri 28 Aug 2026: Q4’26 and FY2026 earnings call (expected)

Coming up next

Is there an opportunity in South Korea’s largest private education platform?

In Apr 2025, the company announced its corporate value-up program.

It planned to grow revenue by +9% p.a. and operating income by 19% p.a. through 2027.

The company also planned to increase capital returns to shareholders, targeting to return 60% of net income through dividends and share buybacks.

This implies 11% yield on enterprise value, an attractive 7.4% premium over South Korea’s 10-year government bond yield.

Finally, a catalyst is on the horizon. In Feb 2026, the media reported that management is looking to sell 32% of the company to private equity. The sale price would value the entire company at around KRW 925 bn.

Yet, the company is trading at only KRW 456 bn enterprise value and 5x NTM P/E.

What’s the catch?

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

At the time of publication, I hold shares in 5DD. I may change my views, predictions, or personal portfolio positioning at any time without notice.

Where are we in the semiconductor cycle?

Angsana Anderson — Wed, 29 Apr 2026 11:36:26 GMT

In your Micro-Mechanics thesis, you promised to explain why the semiconductor industry is so notoriously cyclical, and why, despite the recent growth, it will likely remain cyclical.

You also promised to explain your framework for forecasting the cycle.

Let’s hear it.

Before I begin, I must make two confessions.

First, I did not discover this framework myself.

I learned the key foundations from Jim Handy of Objective Analysis, and then built on them through my work and experience.

I absolve you. What’s your second confession?

Although I said I will explain the semiconductor cycle, in reality, I will explain the memory chip cycle.

With few exceptions, overall semiconductor growth follows memory growth.

Memory chips make up 25% to 30% of the overall semiconductor market by revenue.1 That’s why the memory cycle is a useful proxy for the broader semiconductor cycle.

If we know the memory cycle, we will have a good understanding of the overall semiconductor cycle.

Sounds good. Let’s get into the memory cycle.

First, the most important memory chips are DRAM and NAND.

They store data. DRAM stores short-term data while NAND stores long-term data. You need both in a computer, smartphone, etc.

That’s all the technical detail you need to know.

What most people don’t fully appreciate is that DRAM and NAND are commodities.

How can such high-tech chips be commodities?

The key characteristic of a commodity is that it is fungible. When something is fungible, it means it is identical and interchangeable.

DRAM and NAND fit this definition. It doesn’t matter whether you use DRAM from Samsung or Micron . They all work the same in your laptop.

Ok. Identical and interchangeable = commodity.

That’s right.

Because memory chips are commodities, it doesn’t matter which supplier you buy from. You just want the cheapest price.

Therefore, suppliers compete mainly on price.

In the chart below, I plot the average selling price (ASP) and cost per GB (unit cost) of DRAM and NAND.

Source: Angsana Anderson’s estimates based on Samsung Electronics, Micron and SK hynix

Explain your chart.

Unit cost declines over time because of technological advancements (Moore’s Law).

As unit cost declines, price competition among suppliers pushes ASP to follow cost. ASP follows cost until we reach a shortage.

During a shortage, ASP flattens or even increases while unit cost continues to decline.

Memory chip makers start raking in outsized profits. They invest these profits into new fabs.

Generally, it takes around 2 years to build new fabs.

When these fabs are completed, they flood the market with memory chips.

The market flips from shortage to oversupply. ASP collapses towards cost. Profits disappear.

Because of lower profits or even losses, memory chip makers start cutting capital expenditures (capex). Supply stops growing.

At the same time, demand continues to grow. ASP recovers and follows cost.

The cycle repeats.

Can you summarise this?

The memory cycle has 3 stages:

ASP follows cost: recovery from oversupply
ASP flattens or increases: shortage
ASP collapses towards cost: oversupply

Which stage are we in now?

I estimate we are now already 9 quarters into the current shortage.

Historically, shortages last only 7 to 9 quarters. If history repeats, we could be near the end of the current shortage.

This method is not quite scientific, isn’t it? It feels borderline astrology?

Shortages end when new fabs are completed.

In the chart below, I plotted historical capex growth against memory ASP and unit cost.

It shows ASP usually collapsed 2 to 4 quarters after capex growth peaked.

The peak signals the heaviest investments have been made, meaning new supply is just quarters away from hitting the market.

Source: Angsana Anderson’s estimates based on Samsung Electronics, Micron and SK hynix

Capex growth seems to be forming a peak in Q4’25. This suggests the memory market may flip into oversupply within the next 2 to 4 quarters.

What are the implications for investors?

Let me plot the QoQ % change in SOXX against memory ASP and unit cost. refers to the iShares Semiconductor ETF, one of the most popular semiconductor ETFs.

Source: Angsana Anderson’s estimates based on Samsung Electronics, Micron, SK hynix, SOXX

The last shortage ran from 2020 to 2021. Memory chip makers raked in massive profits, catching investors’ attention and fuelling a +160% rally in SOXX.

By early 2022, after two years of extraordinary gains, investors grew euphoric. “This time is different!” They extrapolated the recent growth indefinitely.

Imagine their shock when ASP collapsed around early 2022. The new fabs opened and flipped the market into oversupply. Profits disappeared.

SOXX crashed as investors raced to dump anything related to semiconductors. “No more lousy cyclicals!” Just like in the late stages of the shortage, they extrapolated the oversupply indefinitely.

Chip makers start cutting capex to stem supply growth. When demand, boosted by generative AI, eventually caught up around early 2024, the current shortage began.

The cycle repeats.

Can you be more straightforward?

I don’t provide financial advice or recommendations. But like any other private investor, I can always share my forecasts and how I am positioning my personal portfolio.

I estimate the shortage will end around Q1’27. Capex growth peaked around Q4’25. Historically, ASP starts to collapse 2 to 4 quarters after the peak in capex growth.

Because of forward guidance, investors will start reacting one or maybe two quarters before Q1’27. This means the rally in SOXX will probably end around Q4’26.

Could you be wrong?

There is a chance the shortage ends in Q1’28, a year later than I expect.

Here’s the same chart, but with consensus forecasts of capex growth.

Source: Angsana Anderson’s estimates based on Samsung Electronics, Micron, SK hynix; Tikr

The consensus is forecasting that after a brief pause in Q1’26, capex growth will accelerate and peak only around Q1’27.

If this happens, the shortage will probably end around Q1’28, a year later than I expect. Accordingly, SOXX will likely only start declining in Q4’27.

However, I am skeptical.

For reasons I will discuss later, during shortages, consensus tends to overestimate future capex growth. Time and again, these estimates are revised downward when the shortage begins to ease.

Furthermore, the target production date of new fabs cluster around early 2027. I asked Perplexity to compile data on major memory fabs under construction and their expected production date:

Finally, if the shortage will end only in Q1’28, this would make the current shortage the longest since 2011. At 17 quarters long, this shortage would be twice the duration of the past 3 shortages.

I am confident the current shortage will eventually end. I expect this to happen around Q1’27. If the shortage persists beyond 2027 and into 2028, I would be very surprised.

Why can’t semiconductor companies forecast their demand accurately and avoid oversupply?

I believe there are 2 big reasons.

These 2 reasons also explain why I believe the semiconductor industry will remain cyclical, despite the tremendous growth recently.

First reason: true demand unknown

When a shortage begins, customers start worrying over inventory. So, like any prudent people, they start hoarding inventories.

When distributors order laptops from HP , they will order extra to create buffer inventory.

HP receives the order. It is bigger than expected. At the same time, HP also sees the shortage starting.

What does HP do? It plays safe and orders its own buffer inventory!

The same thing happens when laptop manufacturers like Quanta order chips from SK hynix.

By the time SK hynix receives the order, everyone’s buffer inventory will have significantly inflated the true demand.

Isn’t this called the bullwhip effect?

Yes.

A small flick of the wrist creates a massive movement at the end of a whip.

A small change in end-demand causes large fluctuations as orders move up the supply chain.

What’s the second reason?

Inelastic supply.

Suppose SK hynix’s order book consists of 50% true demand and 50% panic-driven buffer inventory. Perceived demand has effectively doubled.

Reacting to this massive signal, SK hynix aggressively scales up its capex, doubling the size of its planned fab and ordering twice as much equipment.

In the meantime, the shortage begins easing. Customers start reducing their buffer inventory and true demand becomes clearer.

By then, it is too late. The shell is built, the equipment is installed and calibrated. Wafers are rolling off the line.

Because semiconductor equipment represents such a massive sunk cost, SK hynix has every incentive to keep pumping out volume to recoup its investments, producing as long as marginal revenue exceeds marginal cost.

What is completely rational for an individual chipmaker is disastrous for the industry.

As every player maxes out their utilization rates to cover fixed costs, the market plunges into a full-blown oversupply.

There is no quick fix. Chipmakers just have to live with it until demand catches up eventually.

Some don’t even live through the oversupply. See Elpida Memory, once the world’s third-largest DRAM maker. It went bankrupt in 2012.2

Coming up next

Micro-Mechanics (Holdings) Ltd. (5DD; MMH SP) released their earnings on Tue 28 Apr 2026 and held their earnings call on Wed 29 Apr 2026.

Is the earnings release good or bad? Is my thesis still on track?

Let me update you in my next post.

After that, I want to highlight a potential opportunity: “Why is South Korea’s largest private education platform trading at only 5x NTM P/E?”

Subscribe for free to be notified immediately when I publish.

Subscribe for 2 to 3 analyses of global SMID equities every week.

Discover overlooked ideas and rethink familiar names.

Published by Andrew Wong, ACA, CFA

Disclaimer

I make no representation or warranty, express or implied, as to the accuracy, completeness, or timeliness of the information provided. The semiconductor industry and financial markets are highly volatile, and my forecasts, estimates, and assumptions may prove incorrect. I may change my views, predictions, or personal portfolio positioning at any time without notice.

I am not acting as your financial advisor or fiduciary. You should not rely on any information in this publication to make investment decisions. Always conduct your own independent research and consult with a qualified financial professional before making any investment. Under no circumstances will I be held liable for any direct, indirect, or consequential losses or damages arising from your reliance on the content of this publication.

References

WSTS (2025)

TECRO (2012)

Micro-Mechanics: After a +97% rally, is it too late to buy?

Angsana Anderson — Sat, 25 Apr 2026 06:02:53 GMT

On 28 Jan 2026, I bought 5DD shares and wrote the first version of this thesis.

I have reproduced the thesis here, with corrections and edits for brevity and clarity. The core thesis is unchanged.

Under the ‘Subsequent events’ section, I added updates from the most recent earnings, as well as my latest assessment and decision.

About (28 Jan 2026)

Share price: SGD 1.61

Market capitalisation: SGD 224 mn (USD 177 mn)

Enterprise value (EV): SGD 198 mn (USD 156 mn)

Average daily volume (ADV): SGD 0.2 mn (USD 0.2 mn)

LTM P/E: 17x; LTM P/B: 4.2x

My decision (28 Jan 2026)

I bought my shares at SGD 1.62. Within 3 to 5 years, I expected to sell my shares for at least SGD 3.20.

Over 3 years, I expected to earn ~29% p.a. (26% capital gains; 3% dividend yield). Over 5 years, I expected to earn ~18% p.a. (15% capital gains; 3% dividend yield).

Background

Micro-Mechanics: After a +97% rally, is it too late to buy?

At the start of the year, Micro-Mechanics was an unloved orphan. The consensus? “A lousy cyclical business with no liquidity.”

Today, the narrative has flipped. After surging +97% year-to-date, there is palpable excitement surrounding the shares.

In this post, I’ll break down why I bought shares in 5DD on 28 Jan 2026. I will also try to make sense of the sudden rally that even surprised me.

Finally, I will tackle the most important question: Is it too late to buy?

Business model

5DD’s financial year ends in Jun every year. Unless stated otherwise, all time references will follow 5DD’s financial year. For example, Q2’26 refers to Oct 2025 – Dec 2025.

Breakdown of FY2025 revenue (SGD 65 mn, +13% YoY):

77% consumable tools (+6% YoY)
23% wafer fabrication equipment (WFE) parts (+46% YoY)

The consumable tools segment focuses on the design and manufacturing of miniature consumable tools used in the assembly and testing of semiconductors.

The main consumable tools are die attach tools, wire-bonding tools and encapsulation tools. These tools are used to package bare (‘naked’) chips into finished products. Elastomer (rubber) products represent ~30% of group revenue.1 For details, please see Micro-Mechanics’ product catalog.

The WFE parts segment focuses on making metal components and precision parts such as wafer‑handling parts and precision mounts for semiconductor wafer fabrication equipment. It is largely carried out by its US subsidiary (MMUS).

5DD has > 600 customers worldwide. Historically, major customers include leading outsourced assembly and testing (OSAT) companies like Amkor. In FY2025, 1 customer of the wafer fabrication equipment parts segment represented 10% of group revenue.

31% of revenue comes from China, 22% from USA, 18% from Malaysia and the remaining 29% from the rest of the world.

The largest expense is cost of sales (51% of revenue). This mainly consists of cost of inventories sold.

5DD operates in a niche, fragmented industry. 5DD is much larger than most consumable tooling competitors, but smaller than its WFE parts competitors.

In its consumable tooling segment, its direct competitors are mostly private. These include SPT Roth Ltd (SPT) and Micro Point Pro Ltd (MPP).

SPT seems like the largest competitor. It has 6 locations around the world. LinkedIn shows 156 employees. 5DD’s LinkedIn shows 144 employees and 5 locations.

5DD’s competitors in its WFE parts segment include Frencken Group Limited (E28; FRKN SP) and UMS Integration Limited (558; UMSH1 SP). That said, these competitors are not directly comparable because they have other segments like contract assembly and contract manufacturing.

Founder Christopher Borch and his family control over 43% of 5DD. Former COO Low Ming Wah owns another 5%. Yeo Seng Chong, the founder of Yeoman Capital, owns 0.5% in his personal capacity.

Institutions remain a small presence. No private equity or activists on the shareholder registry.

For a more in-depth introduction to 5DD’s business, check out:

· MONEY FM 89.3’s Chua Tian Tian interviews Micro-Mechanics CEO Kyle Borch

· NUS Professor Mak Yuen Teen’s case study on Micro-Mechanics (page 54). It is one of the few companies that has made the mark on Prof. Mak’s books.

· The Fifth Person interviews CEO Borch

My reasons

Cyclical inflection

Customers report higher demand. The semiconductor industry is notoriously cyclical.

A quick scan of the earnings call of major semiconductor companies suggests an inflection point is near. In their latest earnings call, direct customers such as Amkor Technology, Inc. and ASE Technology Holding Co., Ltd. (3711 TT) provided positive outlooks.

On 28 Jan 2026, Texas Instruments , a leading analog chipmaker and probably an important customer of 5DD, reported better-than-expected outlook. This prompted a +10% share price jump in just one day.

Customers’ inventory is low. The higher demand reported by customers will likely flow quickly to 5DD because customers’ inventory is low.

Although inventory days remain above average across analog names like TXN, analog integrated device manufacturers (IDMs) like TXN are only a part of 5DD’s customer base.

Inventory days at major OSATs (5DD’s other major customer group) like Amkor, ASE, JCET have largely trended below average and are near all-time low.

5DD reaching an inflection point. After 2 consecutive years of revenue decline, revenue finally grew +13% in FY2025.

Inventory days are near all-time low ~38 days in Q2’26. The 10-year average is ~53 days.

Good business, temporary headwinds

5DD is a high-quality business. Since 2006, 5DD’s return on asset (ROA) averaged ~14%. Such high ROA is rare in the industry and signals a high-quality business.

I believe 5DD achieves high ROA because its products are low-ticket, yet mission critical, thereby creating high switching costs for its customers.

5DD’s largest product category is probably its rubber tips. After cutting (‘dicing’) completed silicon wafers into individual chips (‘dies’), semiconductor manufacturers use these rubber tips to pick up the dies and package them.

These rubber tips are mission critical. If the tips are poor quality or worn out, they will contaminate the dies. The dies may need to be discarded.

Despite their importance, these rubber tips are low-ticket. A rubber tip can retail for less than USD 1.00 per unit.2 Wholesale prices should be lower.

In contrast, a completed silicon wafer costs from USD 3,000 for a mature node to USD 30,000 for the most advanced nodes.3

If 5DD’s rubber tips work, why switch to an unproven new supplier and risk spoiling expensive wafers just to save a few cents?

But the market has not fully recognised 5DD’s quality. Between Oct 2025 and Dec 2025, 5DD’s share price sold off after it faced headwinds. 5DD reported +3% YoY quarterly revenue growth, down from +12% in the previous quarter.

The headwinds are likely temporary. The headline +3% is masking the business quality and its true growth potential.

Revenue at consumable tools segment actually grew +8% YoY. WFE parts revenue fell -15% YoY because of delays in materials and shortages.

However, order book increased +20% QoQ. Revenue growth will likely accelerate in future quarters as 5DD resolves the materials shortage.

Indeed, by mid-Nov 2025, the CEO said the material shortage in the WFE parts segment had been resolved.4

Reading between the lines, this implies pushing of revenue from Q1’26 (Jul 2025 to Sep 2025) to Q2’26.

Valuation

At the share price of SGD 1.61 on 28 Jan 2026, 5DD offered 9% free cash flow yield on EV. This yield is likely sustainable because I believed 5DD is at the bottom of the semiconductor cycle.

9% is an attractive 6.8 percentage points (ppt) premium over Singapore’s 10-year government bond yield of 2.2%.

I estimated intrinsic value at SGD 3.20 per share based on a discounted cash flow valuation.

This implies ~50% margin of safety and ~98% upside. This looked very attractive to me.

Catalysts

Capital return when the cycle turns. The semiconductor cycle will eventually turn. When it inflects, 5DD has always returned the excess profits to shareholders. Dividend payout ratio averaged 85% in the past 10 years.

Singapore EQDP. Singapore has injected SGD 4 bn into the local stock market, with plans to inject another SGD 1 bn in 2026.

The total SGD 5 bn will be placed with fund managers to invest in Singapore equities, especially small and mid-caps.

This has driven price discovery among Singapore mid-caps, but not yet for small-caps like 5DD.

As the fund managers build up their research capabilities, small-caps like 5DD will likely appear on their radar.

Risks

What if the semiconductor cycle fails to turn?

In an unexpected downturn, 5DD’s downside is protected by 2 factors: (a) recurring revenue from low-ticket, mission critical consumable tools and (b) no debt (cash > 10% of market cap.).

Previously, we discussed how 5DD’s high-quality business stems from selling low-ticket, mission critical tools.

The other advantage that we have not talked about is that these tools create recurring revenue because they are consumables. These tools will wear out, some need to be replaced within hours.

Replacement cannot be delayed. Otherwise, the production line will stop. In any case, it does not make sense for customers to delay buying replacements from 5DD. These are low-ticket items.

That’s why, during downturns, 5DD is more resilient.

In the most recent downturn during FY2023, 5DD’s revenue declined only -19% compared to -27% for SK hynix Inc. (000660 KS) and -50% for Micron Technology, Inc. .

Factors that could lead me to increase investment in 5DD

Evidence that the current upcycle will be much stronger than historical cycles.
Share price falls > 30% without significant deterioration in business fundamentals.

Factors that could lead me to decrease investment in 5DD

Strong evidence that we are approaching the peak of the semiconductor cycle, and other investors are extrapolating the peak earnings and overpaying for 5DD shares.

Subsequent events

Q2’26 earnings

On 29 Jan 2026, 5DD released their earnings, followed by an earnings call the next day.

My thesis seems to be playing out.

Good business, temporary headwinds

Headwinds temporary. As I expected, revenue growth accelerated to +15% YoY in Q2’26, up from +3% YoY in Q1’26.

WFE parts segment recovered from material delays and shortage. Revenue increased +6% YoY in Q2’26, up from -15% YoY in Q1’26.

Revenue growth from consumable tools accelerated to +17% YoY in Q2’26, up from +8% YoY in Q1’26.

Cyclical inflection

Customers placing more rush orders. Demand is strong and improving.

CEO Kyle Borch reported WFE parts customers placing more advance orders.

Consumable tools customers are placing more rush orders and are increasingly concerned whether they can get their consumable tools fast enough to support their own customers.

Working capital trends signal cyclical inflection. Although 5DD does not provide guidance, working capital trends suggest the cycle has likely bottomed out and outlook is better than expected.

After multiple quarters of destocking, inventory finally increased slightly by +2% YoY in Q2’26. Coupled with the acceleration of revenue growth to +15% YoY in Q2’26, this suggests 5DD likely sees strong demand and is stocking up to meet that demand.

VP Finance Wendy Tan explained that the inventory increase in Q2’26 was concentrated on raw materials.

This supports my hypothesis. Demand likely grew stronger than expected because 5DD is finally restocking inventory after multiple quarters of destocking.

Low risk of irrational capacity expansion over the next year. I asked, “What are the risks of your competitors irrationally expanding capacity?”

CEO Kyle Borch believes this risk is low over the next year or so. “There is certainly a lot of demand… real demand…”

Catalysts

Singapore EQDP. Besides engaging the funds that received the SGD 5 bn EQDP allocation, CEO Kyle Borch did not offer more details on how they plan to benefit from Singapore’s EQDP. Reading between the lines, it seems like this is not their priority.

On the positive side, they noted multiple new faces during this earnings call, including myself.

Besides boutique fund managers and sell-side, I noticed analysts from at least 2 funds that received SGD 5 bn EQDP allocation.

In my thesis, when commenting on catalysts, I said ‘As the fund managers build up their research capabilities, small-caps like 5DD will likely appear on their radar.’ The participation of these analysts supports this.

5DD will likely increase capital returns only after multiple quarters of good performance. The negative point is that 5DD kept interim dividend unchanged at 3 cents.

In FY2024, 5DD cut their interim dividend in half. I believe management wants to see multiple quarters of good performance before increasing dividends again.

My decision (24 Apr 2026)

This is a record of my investment decisions, not financial advice. Read the full disclaimer at the end.

In less than 3 months, 5DD’s share price has climbed to SGD 3.24. I do not expect every stock I buy to perform like this. Furthermore, performance should be assessed across the entire portfolio, not based on a single winning call.

I have made my share of mistakes, including Haw Par Corporation Limited (H02; HPAR SP).

I will certainly make more. What’s important is that, over a full market cycle, my portfolio outperforms.

I have been in this business long enough to know that skill matters, but luck and emotional discipline matter just as much, if not more.

Peter Lim, a Singaporean billionaire investor reportedly said: “When you are holding stocks, if it goes up, don’t be too happy; when it goes down, don’t be too sad. Otherwise, how? Your life will also be fluctuating, and you’ll die of a heart attack.”5

I decided to hold. I believe 5DD is not overvalued, despite the near +100% rally this year. At the current share price of SGD 3.24, I believe it is only trading around its intrinsic value.

Looking to redeploy. That said, I am leaning towards redeploying a portion of the capital to other more attractive opportunities when they arise.

Upside not obviously greater than downside. My leading indicators signal that we could be approaching the peak of the memory cycle.

Even though I estimate 5DD has more exposure to analog chips than memory chips, the memory cycle is important because it leads the analog cycle.

Margins on memory chips at unprecedented highs. The gap between the average selling price of memory chips (ASP) and unit cost is the widest it has ever been since 2011.

Is it possible that the margins continue to exceed expectations and widen further? Yes.

Is it probable? I believe not.

But capex growth peaking. Capex growth at major memory chip makers like SK Hynix has peaked around Jun 2025.

Historical precedents show that ASP typically collapses within a few quarters of capex growth peak. The opening of new fabs floods the market with more memory chips. This pushes down ASP.

In summary, I am maintaining my position in 5DD for now. However, with the risk/reward profile becoming less attractive, I am actively looking to rotate this capital once a more compelling opportunity presents itself.

Risk looks low in the near-term. My decision to hold through the near term is driven by a low probability of a major disappointment in the upcoming earnings call on Tue 28 Apr 2026.

Positive industry read-throughs. On 22 Apr 2026, TXN reported a better-than-expected outlook, prompting a nearly +20% share price jump in just one day.

Healthy channel dynamics. Average inventory days for major OSATs ticked up slightly in calendar Q4’25 but remain comfortably below historical averages.

Upcoming events

Tue 28 Apr 2026: Q3’26 earnings release

Wed 29 Apr 2026: Q3’26 earnings call

Valuation model

Subscribers may request a free copy of my valuation model by emailing me at angsana.anderson@gmail.com.

I will send out my model by the end of Sun 26 Apr 2026.

Coming up next

Where are we in the semiconductor cycle?

When I was a junior semiconductor analyst, I hung onto every word of the CEOs of major semiconductor companies.

After multiple quarters of “seeing the bottom pushed out”, I realised that even industry leaders have limited visibility into the cycle’s timing, or their public guidance is inherently constrained by corporate policies.

I was lost in the wilderness. That is, until I found this framework:

Source: Samsung Electronics Co., Ltd. (005930 KS); Angsana Anderson

In my next post, I will explain why the semiconductor industry is so notoriously cyclical, and why, despite the recent tremendous growth, it will likely remain cyclical.

I will also demonstrate this framework.

How do I use it to understand where we are in the cycle, and what likely lies ahead of us now?

Subscribe for free to get my next analysis when I publish it.

Subscribe for 2 to 3 analyses of global SMID equities every week. Discover overlooked ideas and rethink familiar names. Published by Andrew Wong, ACA, CFA

The Business Times (2025)

Disclaimer

This is a record of my investment decisions, not financial advice. I may change my decisions without notice. Use this only for education and entertainment. All analysis and opinions expressed are solely my own, and 5DD has not reviewed or endorsed this post. Do not rely on this for your investment decisions. I hold shares in 5DD.

References

5DD (2025)

SGC Equipment (2026)

Silicon Analysts (2026)

Vulcan Post (2020)

Chagee: Profits attract competition, competition destroys profits

Angsana Anderson — Wed, 22 Apr 2026 12:20:28 GMT

About (22 Apr 2026)

Share price: USD 10.60

Market capitalisation: USD 1,969 mn

Enterprise value (EV): USD 1,046 mn

Average daily volume (ADV): USD 5 mn

NTM P/E: 7x

My Decision

Pass (previous: pass)

Background

Chagee: Profits attract competition, competition destroys profits

At only 7x NTM P/E, Chagee looks attractive.

However, I decided to pass.

Chagee’s latest earnings show an industry that remains too competitive. Company-owned stores nearly doubled from a year ago. However, this is likely the result of Chagee taking over unprofitable stores from franchisees.

Does Chagee’s opportunity lie overseas? Unlikely.

In Q4’25, overseas same-store sales growth continued to deteriorate. I found out why after visiting Chagee’s flagship store in Singapore.

Consumer tastes are rapidly changing, yet Chagee became slow to respond.

In this Substack post, I will discuss these concerns in detail.

Business model

Breakdown of 2025 revenue (RMB 12.9 bn, +4% YoY):

88% from franchised stores (−2% YoY)
12% from company-owned stores (+93% YoY)

CHA earns revenue from franchised stores by selling products and services to franchisees.

Out of the 88% revenue from franchised stores, 75 ppt is earned from sale of raw materials, 7 ppt from store equipment and other supplies and the remaining 6 ppt is earned from franchising and other services.

Revenue from company-owned stores is generated from the sale of freshly made tea drinks directly to customers.

As of 2025, CHA has 7,453 stores. 7,108 stores are in mainland China and HK. The remaining 345 stores are located overseas, mainly in Malaysia (217 stores).

6,838 stores are franchised while the remaining 615 stores are directly owned by CHA.

Relevant public comparables include Nayuki Holdings Limited (2150 HK), Auntea Jenny (Shanghai) Industrial Co., Ltd. (2589 HK), and Sichuan Baicha Baidao Industrial Co., Ltd. (2555 HK).

My reasons

Industry is oversaturated and too competitive

Franchisees quitting. At first glance, CHA seems to be doing well. Overall same-store gross merchandise value (GMV) growth improved from -28% in Q3’25 to -26% in Q4’25. CHA even increased its store count by 115 in Q4’25.

However, a closer examination reveals the market remains oversaturated, business is getting worse and likely to continue deteriorating.

The increase in store count is mainly driven by company-owned stores in China. In Q4’25, company-owned stores increased by 168. At the same time, CHA closed 136 franchises stores in China.

“So what? Maybe Chagee wants to operate more stores themselves for quality control.”

I interpret this differently. I believe the oversaturation in the market is pushing many franchisees into losses.

According to 曹阳 (Cao Yang), a franchisee from the Beijing-Tianjin-Hebei region, CHA currently offers 2 solutions for underperforming franchised stores: optimisation or closure.1

Optimisation allows franchisees to find stores with lower rents and relocate, while closure means CHA directly acquires the stores and converts the franchised store into a company-owned store.

The fact that company-owned stores increased significantly just when franchised stores started closing en masse suggests that many franchisees are exiting because of significant losses. This could also explain why CHA slipped into operating losses in Q4’25.

It also suggests a bleak outlook. Put yourself in the shoes of a franchisee. If there are any signs that business will improve very soon, you would likely have held on and continued operating.

Working capital trends support this view. Contract liabilities, which primarily comprise upfront fees and advance payments from franchisees, declined 8% in 2025. This is the first drop since the company began disclosing this metric.

Second-hand equipment piling up. According to Mr. Liu, a dealer of second-hand equipment in China, “At this time last year, you had to fight to get a set of Chagee equipment.

Now? I have more than a dozen sets piled up in my warehouse.

Many franchisees have been listing their equipment for half a year but haven’t been able to sell it. In the end, they can only sell it as second-hand equipment at scrap metal prices.”2

He added that many franchisees who have listed their equipment still have very new store decorations.

“Some have been open for less than a year.”

Low barriers to entry

The grass is greener on the other side? The bulls argue that international markets will come to CHA’s rescue. After all, overseas store count increased +32% QoQ in Q4’25. Chagee will be the Starbucks of tea!

I am less optimistic.

Overseas same-store GMV growth deteriorated from -23% in Q3’25 to -26% in Q4’25. I believe it will deteriorate further, and the consensus is likely too optimistic.

During a Saturday in Apr 2026, I visited Orchard, Singapore’s busiest shopping district.

Orchard is where Chagee opened its Singapore flagship store in Aug 2024. It is also where Molly Tea, a rival, opened its flagship store in Mar 2026.

Let’s look at the newcomer first.

Source: Angsana Anderson

I was immediately struck by how busy it was. This was the scene at around 9pm on a Saturday. Look at how packed the counter is. Everyone is busy.

Look at the far-right corner: the mountain of cups stacked together. Each cup represents an unfulfilled order.

Here’s a close-up:

Source: Angsana Anderson

Besides the mountain of teacups, I want to highlight the automated tea-making machine.

Part of the bull case for CHA is its automated tea-making machines. CHA wrote in its prospectus: “we have developed automated tea-making machines that significantly enhance the efficiency of store staff, reduce consumer wait times, and elevate the overall consumer experience.”

CHA continues: “By incorporating our self-developed tea drinks recipe parameters into these machines, we enable our store staff to craft each cup of tea drinks [sic] in consistent high quality at approximately 8 seconds, eliminating the need for memorizing intricate operational protocols.”3

Sadly, a competitive advantage is not much of an advantage if competitors can easily copy it. Automated tea-making machines are now even available through third-party suppliers.4

At Molly Tea, I watched a staff prepare tea using an automated tea-making machine. The machine functions like the one in CHA’s stores. Based on the crowd, I reckon the taste is just as good or better than CHA’s.

Let’s move on to CHA’s flagship store in Orchard. It’s less than 5 minutes’ walk.

Source: Angsana Anderson

This is the best photo I could take without attracting attention, so I will have to lean more on my written observations. I have two observations.

First, CHA’s store looks much less busy than Molly Tea.

Second, I saw at least two staff idling. Excess staff suggests demand is weaker than expected. Store managers prepare staffing roster based on their expectation of demand.

Rapidly shifting consumer tastes.

Active customers likely declined. One of my favourite techniques is to compare the latest earnings release with the prior one.

Management usually drafts the new release by rolling forward the previous quarter’s document. If I notice a subtle shift from positive to more neutral language, it often points to a weaker outlook. Most importantly, I pay close attention to any KPIs that are redefined, deprioritised, or quietly removed.

In Q4’25, CHA removed ‘active members’ from its key operating data disclosure table.

CHA explained that it had expanded its definition of active members to include “traceable member activity across other third-party online delivery and social networking platforms.”5 As a result, its active members finally broke its declining trend:

Source: CHA

CHA did not provide a like‑for‑like active member number under the previous definition. If it had not expanded its definition, I believe active members would likely have shown a significant decline.

What is driving the declining trend?

“You’re hot and you’re cold” I believe the answer lies in the rapid change in consumer tastes, and CHA’s failure to respond quickly.

According to Cao Yang, “Boya Juexian used to account for more than 50% of the store’s sales, making it a solid pillar of sales. But in the past two years, Boya Juexian has become increasingly difficult to sell…”6 Boya Juexian is Chagee’s bestselling drink, a jasmine milk tea.

The franchisee added: “New tea drinks lack the addictive qualities of coffee products, making it easy for consumers to experience taste fatigue. On social media platforms, many consumers are saying they’re tired of Boya Juexian and don’t want to drink it again after a few tries…”

CHA is struggling to keep up with the rapid shift in consumer tastes.

From Jan to early Dec 2025, CHA launched just 8 new products, and only 4 after Jun, 2 of which were iterations of Boya Juexian. Between 2022 and 2024, rivals launched many more new products than CHA.

Source: 蓝鲸新闻 (2026)

CHA seems to have recognized this problem. During the Q4’25 earnings call, CEO Zhang admitted: “Honestly, we took some detours in new product launch rhythm and marketing execution, and we did not fully keep pace with how fast the market was moving.”

Factors that could lead to a re-assessment of my decision

Strong evidence that the industry is not as saturated and competitive as I believed (e.g. sustained recovery in same-store GMV growth).
New entrants stop entering, existing competitors start exiting.
CHA demonstrates consistent ability to follow consumer tastes.
1. According to CEO Zhang, CHA’s new “Signature Pu-Erh Tea” launched in Dec 2025 drove a 16% week-over-week GMV uplift, significantly exceeding the historical average for all new products.
2. If CHA’s new “Signature Pu-Erh Tea” and future new products continue to perform well, this will warrant a re-assessment.

Coming up next

Micro-Mechanics: After a +90% rally, is it still cheap?

At the start of the year, Micro-Mechanics was an unloved orphan. The consensus? "A lousy cyclical business with no liquidity."

Today, the narrative has flipped. After surging +90% year-to-date, there is palpable excitement surrounding Micro-Mechanics (Holdings) Ltd. (5DD; MMH SP).

In my next post, I’ll break down why I bought shares in 5DD on 28 Jan 2026. I will also try to make sense of the sudden rally that even surprised me.

Finally, I will tackle the most important question: Is it still cheap?

Subscribers get my analysis first. Free to join.

Subscribe for 2 to 3 analyses of global SMID equities every week. Discover overlooked ideas and rethink familiar names. Published by Andrew Wong, ACA, CFA