We’ve constantly been sharing a list of our recent reads in our weekly emails for The Good Investors.
Do subscribe for our weekly updates through the orange box in the blog (it’s on the side if you’re using a computer, and all the way at the bottom if you’re using mobile) – it’s free!
But since our readership-audience for The Good Investors is wider than our subscriber base, we think sharing the reading list regularly on the blog itself can benefit even more people. The articles we share touch on a wide range of topics, including investing, business, and the world in general.
Here are the articles for the week ending 24 May 2026:
1. The end of the resource exponential – Brandon Carl
Financial analysts are currently engaged in a collective exercise in ruler-drawing. By mapping the trajectory of compute spending and GPU sales, they have constructed a future that is essentially a larger version of the present. In this view, the path to artificial intelligence is a matter of pure capital expenditure. If $100 billion buys a certain level of reasoning, then $1 trillion must buy ten times as much. It is an investment thesis built on extrapolation.
History, however, is not linear. In any technological cycle, the most dangerous moment is when the market begins to treat the status quo as a permanent law. Today’s AI logic—that hardware scale is the primary lever—is less a rule of physics and more a temporary workaround for inefficient architecture. As investors calculate the return on ever-larger clusters, they are ignoring a more fundamental lesson: what is built today is rarely what defines tomorrow.
The shift from brute force to elegance is not just likely; it is a mathematical necessity. Much of modern AI is built on transformer architectures that exhibit quadratic complexity. Double the input, and the requirements for compute and memory grow fourfold. Quadratic consumption of any limited resource will eventually consume everything. Efficiency is not an optional optimization; it is a condition for survival…
…This does not mean the end of the massive GPU cluster, but it does mean that algorithmic efficiency, rather than just raw silicon, will increasingly solve resource shortages. As architectural pivots reduce the dependence on brute-force scaling, the “mix” of hardware required by a data centre five years from now will look nothing like the procurement lists of today.
The risk for investors is to over-index on “selling the shortage” based on current constraints. Subsidizing the construction of yesterday’s architecture is a recipe for stranded assets. In the history of technology, the greatest returns have rarely accrued to those who simply bought the most hardware, but to those who understood how the math was changing. Ideas travel much faster than silicon.
2.AI Chip Mania Sows Seeds of Its Own Destruction – James Mackintosh
Memory chips are a perfect example of a highly cyclical industry. Heavy investment is required to build a fabrication plant, or fab. When demand rises, it takes several years for supply to catch up, during which prices and profits jump. Those high profits encourage CEOs to expand supply. And the high fixed costs encourage producers to run fabs at full capacity—even when supply overshoots demand. The cycle turns when excess supply pushes down prices and profits plunge, as they did in 2022-23.
Already the high profitability has encouraged heavy capital spending. Micron is spending $150 billion to build or expand fabs in New York, Idaho and Virginia, and new Korean fabs are opening…
…The risk of a downturn is embedded in Micron’s valuation. Two weeks ago it was the S&P’s third-cheapest stock measured by price to forward earnings, and it’s still at under 10 times, tame for a highflying stock. That doesn’t make it cheap, though. It just means investors recognize that the boom times in memory chips never last.
History shows how this works. In the last cycle Micron stock peaked at the start of 2022, with the forward P/E at just nine times, ahead of a halving in the shares that year. The stock bottomed out and subsequently doubled after the loss was baked into predictions…
…The biggest risk is impossible to quantify: AI technology could become far more efficient in its use of memory, meaning data centers need less of it. Memory stocks took a hit in March when Alphabet researchers published a paper showing dramatic improvements in memory efficiency, but have recovered. Large language models are an immature technology, and engineering improvements for specialized data centers should be expected—but how big they are and when they come is unknowable in advance.
Other risks apply to the whole AI supply chain: Data-center plans may be scaled back, AI uptake prove slower than hoped, or a political backlash may hinder expansion. All are plausible; none are considered that serious by the AI bulls driving stock prices.
A final risk is that supercharged profits attract new rivals to enter the market. For now, that seems unlikely in the superfast memory Micron makes, but it’s already happening with other highly profitable chips used in AI.
3. The toll booths of lending – Michael Fritzell
To manage risks, banks and companies gather information on their counterparties. And one way to do so is to buy data from so-called “credit bureaus”, also known as “credit reporting agencies”.
These credit bureaus gather information on borrowers’ creditworthiness. These include consumers, corporate borrowers, and trade counterparties. The data is then used to support lending decisions, ensuring that each lender is comfortable with their exposures…
…On the corporate side, credit bureaus collect all sorts of data on private businesses: business registration numbers, legal addresses, ownership data, the executive leadership, name changes, etc. And more importantly, they collect data on revenues, profitability, and leverage from public filings, interviews, payment data, etc. They also cooperate with debt collectors to understand whether each business has had payment issues in the past.
All this data then ends up in credit reports, which you can purchase for US$150 each. Historically, these credit bureaus made money by selling credit reports a la carte. But today, the entire industry has moved towards subscriptions that generate much higher-quality, recurring, and sustainable revenue. If you’re an ongoing subscriber, you’ll get alerts if there are any changes to the creditworthiness of any particular counterparty…
…Buyers of corporate credit data tend to be small- and medium-sized enterprises that want to know whether they extend favourable credit terms to their counterparties. Or banks that want to know how to extend credit to. The local Asian credit bureaus have almost impenetrable market positions, as they’ve gathered detailed information on millions of businesses. And the reports can be purchased for very little money, while costing almost nothing to produce. No serious lender would skip a US$50 credit check before extending a half-million loan…
…And because collecting consumer data is sensitive, it is highly regulated and therefore protected. The buyers of credit data tend to be financial institutions that want to know whether to extend a mortgage or consumer loans.
There are clear network effects: in many cases, credit bureaus get data on consumer borrowers from their bank customers, who willingly provide the information in exchange for data on other banks’ borrowers. So the bureaus almost become central exchanges that become difficult to displace.
On the other hand, the heavy regulation also means that pricing power tends to be limited. So it’s a scale business, with significant operating leverage if credit growth for whatever reason starts to accelerate.
And this is the exact bull case for Asia’s credit bureaus: the credit penetration in this part of the world remains low, especially in emerging Asian nations like Indonesia and the Philippines.
4. 18% IRR for 57 Years – Joe Raymond
George Batten founded the Batten Company in New York in 1891. At the time, advertising was mostly about placing ads in newspapers.
In 1919, Barton, Durstine & Osborn emerged, focused more on messaging, copywriting, and persuasion.
The two merged in 1928 to form Batten, Barton, Durstine & Osborn.
Over the next several decades, BBDO became a core player on Madison Avenue, helping large corporations build brands as radio and television expanded their reach.
BBDO International started trading over the counter in 1968…
…As Larry recalls:
“I came to realize advertising was a royalty business. If you had a consumer product, you needed to advertise. And you needed to use an ad agency like BBD&O. I viewed it as a royalty on consumer spending.”…
…He paid less than 8x earnings for a business generating 20% return on equity, growing in the low-double-digits, and yielding 7.5%…
…BBDO grew revenues from $49 million to $155 million from 1969 to 1979 (12% CAGR).
Net income tripled from $4 million to $12 million. Shares outstanding declined from 123 million to 106 million. As a result, EPS quadrupled from 3 cents to 12 cents (15% CAGR).
The P/E multiple ended the period at about the same 7.6x it started.
The stock went from 25 cents in 1969 to 85 cents in 1979 while also paying out 46 cents per share of dividends.
Including dividends, the IRR for his first decade of ownership was 20%…
…EPS over the 11 years from 1979 to 1990 grew from $0.12 to $0.25 (7% CAGR) while paying out a cumulative $0.99 per share of dividends. Not spectacular performance, but not terrible either.
The stock started the decade at $0.85 and finished at $2.73. Thus, Larry had a 10-bagger in his first 20 years of ownership, plus dividends worth nearly 6x his purchase price.
1979 to 1990 was a mediocre stretch for earnings growth. But dividends were consistently paid and the multiple expanded 45% from 7.6x to 11.0x. The result was a 17% IRR for the 11-year period…
…Like many other stocks (and the market averages), 2000 to 2010 represented a “lost decade” for Omnicom shareholders.
The business itself grew at a decent rate–EPS compounded at 8% and $5.48 of cumulative dividends per share were paid.
Counteracting these factors was a 50% reduction in the multiple. 32x in 2000 fell to 15x in 2010. The net result was a 1% IRR for the decade.
Operationally, the 2000s didn’t look that different than the 1970s (8% EPS growth in the former vs 7% in the latter). Yet the 1970s produced a 17% annualized return while the 2000s yielded only 1%.
Such is the power of valuation. The same quality business can deliver wildly different results depending on the price paid. In this case, paying 8x earnings resulted in an annual return of 17% for a decade while paying 32x delivered almost nothing for 10 years….
…BBDO was an ideal buy and hold investment in the 1960s and 1970s.
The economics were attractive (20%+ ROE) and growth prospects solid (decades of global advertising growth ahead). Capital allocation was sensible (small bolt-on acquisitions, share repurchases, and dividends), and the valuation was cheap (sub 10x earnings).
$10,000 invested in 1969 and held through today would be worth $3.2 million, with an additional $1.7 million of dividends received as well.
5. The American Rebellion Against AI Is Gaining Steam – Amrith Ramkumar, Katherine Blunt, and Lindsay Ellis
Delivering a commencement address at the University of Arizona, Schmidt told students the “technological transformation” wrought by artificial intelligence will be “larger, faster and more consequential than what came before.” Like some other graduation speakers mentioning AI, Schmidt was met with a chorus of boos.
In one poll after another in recent weeks, respondents have overwhelmingly voiced concerns about AI, a challenge to claims by industry executives that their technology would gain popularity by improving people’s lives…
…Pollsters and historians say the souring of public opinion is all but unprecedented in its speed. “I don’t think I’ve ever seen something intensify this quickly,” Gregory Ferenstein, who conducted a recent poll with researchers at Stanford University and the University of California, Berkeley, said of the backlash…
…Voters in Festus, Mo., ousted four city council members a week after they approved a $6 billion data center. Dozens of communities in states from Maine to Arizona are trying to ban new data centers. Some 360,000 Americans are in Facebook groups opposed to the facilities, roughly quadruple the number from December, figures from organizations fighting the AI build-out show…
…AI has risen in importance most quickly among 39 political issues studied by polling firm Blue Rose Research in the past year, though it still trails priorities including the economy, immigration and foreign policy…
…But all over the country, community-level organizations have been succeeding in blocking data-center projects. Local opposition blocked or delayed at least 48 projects valued at some $156 billion last year, according to Data Center Watch, an organization tracking the trend. A record of 20 were canceled in the first quarter of the year because of local backlash, figures from climate-media outlet and data provider Heatmap show. Dozens more are currently facing similar obstacles on top of obstructions because of permitting snafus and equipment shortages.
Disclaimer: The Good Investors is the personal investing blog of two simple guys who are passionate about educating Singaporeans about stock market investing. By using this Site, you specifically agree that none of the information provided constitutes financial, investment, or other professional advice. It is only intended to provide education. Speak with a professional before making important decisions about your money, your professional life, or even your personal life. We currently have a vested interest in Meta Platforms (parent of Facebook). Holdings are subject to change at any time.