Why China isn't chasing the smartest AI, growth but not at any price, & Howard Marks's test for AI investors — 3-Bullet Sunday


Hey Reader - Thomas here.

Good day to my fellow compounders!

In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies.

Enjoy!


Reputation Takes Decades to Build and Moments to Unravel

As the saying goes, there are decades where nothing happens, and there are weeks where decades happen. This was one of those weeks.

Like many of you, I was shocked when the CEO of Nvidia, Jensen Huang, resigned. I did not expect 2026 to end this way, or for it to unravel this quickly.

It is a sad story, not just for Nvidia and the AI industry, but for everyone involved, especially his family.

When I think about what to take away from it, the first thing is how important reputation is. It takes decades to earn, and it can be lost within minutes.

The second is for us as investors. Obviously he was the most important person at Nvidia, and I know some investors are extremely concentrated in the stock. I don't think any analyst had a black swan like this in their model, but it happened.

The only way to deal with a risk like this is to diversify adequately (not excessively either, but that's a topic for another day). That way no single company, and no single person, can sink the whole portfolio.

Compound Steadily,

Thomas


I. Why China Isn't Chasing the Smartest AI

I've always enjoyed Rob Vinall's postcards, and I learn a lot from every one of them. This one is from his latest trip to Hong Kong and Shenzhen.

I thought it was interesting because of his observations on the differences between China and the US when it comes to AI investments. In the US, the top AI labs are racing to be first to the most powerful AI, and everything else comes second. This sort of explains why they are spending without any regard for cost. In China, AI is seen as a business opportunity, and the biggest focus is driving down the cost of inference, which is what it costs to run a model every time someone uses it.

Rob puts Tencent's capex at around $30bn this year, one sixth of the largest US hyperscalers (the tech giants building these huge data centres), and thinks it would spend more if it could get the best Nvidia chips.

I agree that, for most business operations or for most people, we don't need the most frontier model or the smartest model, but we need something good enough. Most importantly, the cost must make sense when measured against the returns.

​Click here to read​

II. The Dolphin and the Anchor: Growth, But Not at Any Price

This essay from Investment Masters Class is about growth, and the insightful part for me is how important growth is in the first place. As Warren Buffett put it, "The real money is going to be made by being in growing businesses."

But growing for its own sake isn't enough. The quote I like the most is from Howard Schultz. After Starbucks expanded too fast and closed about 600 stores, he wrote, "Growth, we now know all too well, is not a strategy. It is a tactic."

The title comes from an old Roman coin with a dolphin wrapped around an anchor, the dolphin for speed and the anchor for what holds it back. The essay has plenty of companies that hold back on purpose. By its count, Discount Tire could open 80 to 100 stores a year and opens 40 to 50.

The summary at the end on what investors should look out for is helpful as well:

Is the business growing faster than its culture and people can keep up with?
Where is the money for the growth coming from?
Is management chasing size, or is size the result?

Growth is important, but not growth without a filter or growth at any price.

​Click here to read​

III. Howard Marks's Simple Test for AI Investors

Howard Marks has a new memo out on AI. He asks eleven questions about investing in it and says upfront that he can't answer them definitively.

One of them comes with a simple test. Pick an AI company and guess its earnings per share five years from now. Would you bet the real number lands within 50% of your guess? If you say $10, that's anywhere from $5 to $15.

It reminded me of what I wrote above. I don't think any analyst had this week in their model either.

He ends with a few questions that apply to any investment, like how many things need to go right before you make a decent return, and how much you stand to lose if you're wrong.

​Click here to read​

That's all I have for you today!

Till next week.

Thomas

Author of the #1 Straits Times bestseller The Lunch Break Investor​

​

Steady Compounding

I write about investment concepts, business breakdowns and timeless lessons from super investors. Featured on Business Times, Channel News Asia (CNA) and more. Read by over 10,000 investors.

Read more from Steady Compounding

Hey Reader - Thomas here. Good day to my fellow compounders! In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies. Enjoy! SQ 656 to Fukuoka in 2020. I couldn’t have been more excited about seeing cherry blossoms for the first time in my life... But I never took that flight. COVID-19 shut travel down in March 2020, about a week before departure. That same month, the US stock market’s...

Hey Reader - Thomas here. Good day to my fellow compounders! In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies. Enjoy! AI Stocks Are Having a Rout. This week I spent half a day setting up Grok Bot as my AI assistant. The universe is funny this way, because the very next morning Meta announced Muse is here in Singapore, and of course I cleared my calendar again to set that up too....

Hey Reader - Thomas here. Good day to my fellow compounders! In 3-Bullet Sunday, I share with you the three most interesting piece of article, video or podcast I came across for the week on investing and growth philosophies. Enjoy! Chasing the Northern Lights This December In early December, I'm going to the Nordics for the first time in my life, to see the Northern Lights. Apparently the sun runs on a cycle of about 11 years. It peaked in late 2024, but the years right after a peak can still...