The Law of Large Numbers Comes for Nvidia

Nvidia’s New Normal

One way to think about being a $4 trillion company is that even your spectacular successes can start to look like a slowdown. This, in essence, was the story of Nvidia's latest earnings report. The company posted another quarter of record revenue, with sales surging 56% from a year ago. And yet, the stock dropped.

The basic situation is that Nvidia has become a victim of its own success. The company is now so large, and its growth has been so meteoric, that it is running into the fundamental limits of mathematics and geopolitics. The problem, of course, is that the stock market is not always fond of limits.

The first speed limit is the law of large numbers. A 56% growth rate is, for any normal company, a triumph. For Nvidia, it was the slowest in more than two years. At this scale, it is simply mathematically harder to maintain the triple-digit growth that investors had come to expect. The market is priced for a rocket ship that can accelerate into infinity, and this was the first hint that even this rocket has to obey the laws of physics.

But the speed limits on Nvidia aren't just mathematical; they are also geopolitical. The company's plan to re-enter the Chinese market with its H20 AI chip has stalled. Here is the Wall Street Journal on the state of that deal:

China remains a major question mark. Nvidia seemed to score a major victory last month when the Trump administration reversed course and said Nvidia could sell its H20 chip to the Chinese market. But the company disclosed Wednesday that it hasn’t shipped any of those chips and might not in the current quarter “as we continue to work through geopolitical issues.” Nvidia added that it has yet to see any published regulations regarding the 15% cut the Trump administration has demanded of any sales of AI chips to China.

And then there is the problem of being too profitable. Nvidia’s gross margins hit an incredible 72.4%, a level that acts as a giant, glowing target for everyone else. This includes direct rivals like AMD, in-house chip design efforts from its biggest customers, and a host of startups all hoping to grab a piece of those profits. That margin is also, it turns out, a target for the US government, which sees it as a convenient source of revenue via its new 15% "tax." The problem with being the AI kingmaker is that everyone, from competitors to governments, wants their cut.

Tesla's Narrative

For a long time, the electric car market was pretty simple. There was Tesla, and then there were some other companies that also made electric cars but that you didn't really have to think about.

Anyway, new data from Europe suggests this state of affairs might be changing. In July, Tesla’s new car registrations plunged by a startling 40% from a year ago. During that same month, its main Chinese rival, BYD, saw its registrations shoot up by 225%. The really crucial detail here is that the overall market for battery-electric cars in Europe was still growing. This is not a story about people souring on EVs. It is a story about people souring on, or at least having other options than, Teslas.

So what happened? Well, there are a few working theories. One is that Elon Musk’s very public political life has made the brand a bit less palatable to some of its traditional eco-conscious customers in Europe. Another, more boring theory is that Tesla has finally run into a problem that Ford has understood for a century: people like new stuff, and Tesla’s core models are getting a little old.

Faced with this rather inconvenient reality, Tesla is doing what any good company does when the numbers for its main product look bad: it is trying to convince everyone that it’s not really in that business anyway. Here is Thomas Besson, an automobile sector research analyst, speaking to CNBC on the company’s new narrative:

They talk about almost everything else but the car they’re selling at a slower pace now because effectively, the age of their vehicle is much higher than the competition and the latest products have not been as successful as hoped, notably the Cybertruck.

This seems like a reasonable strategy. If the data for the thing you actually sell is getting rough, it helps to talk about the amazing, world-changing things—like AI, robotics, and robotaxis—that you might sell someday. In the meantime, the car market has become a real, messy, global fight. And Tesla’s response, it seems, is to argue that it’s not really in that fight at all.

The Scoreboard

• Semiconductor: Nvidia CEO Huang says bringing Blackwell AI chip to China 'is a real possibility' (CNBC)
• Hyperscaler: Google has eliminated 35% of managers overseeing small teams in past year, exec says (CNBC)
• Robotaxi: How Tesla and Waymo's radically different robotaxi approaches will shape the industry (Reuters)