Today's current thing: Aug 28

Also: SemiAnalysis' Doug O'Laughlin on the stream today to talk Nvidia earnings

Aug 28, 2025

Happy Thursday, we are live. The current thing is Nvidia’s earnings and share price drop.

We’d love your feedback on our Substack. What should we start doing, what should we do more of, and what kinds of emails do you want to get from TBPN? Just reply directly to this email.

Here’s John’s run of show. 👇

Jensen Huang less bullish on AI than Leopold Ashenbrenner

The number that stuck out to me during Nvidia’s earnings call yesterday was Jensen Huang’s estimate that between $3 and $4 trillion will be spent on AI infrastructure by 2030. It’s massive — and in the Q&A, Jensen walked through his math:

“As you know, the CapEx of just the top 4 hyperscalers has doubled in 2 years. As the AI revolution went into full steam, as the AI race is now on, the CapEx spend has doubled to $600 billion per year. There's 5 years between now and the end of the decade, and $600 billion only represents the top 4 hyperscalers. We still have the rest of the enterprise companies building on-prem. You have cloud service providers building around the world. United States represents about 60% of the world's compute. And over time, you would think that artificial intelligence would reflect GDP scale and growth and so — and would be, of course, accelerating GDP growth.”

“$3 to $4 trillion” is big, but it pales in comparison to the projections Leopold Ashenbrenner laid out in Situational Awareness, which pegs 2030 at closer to $8 trillion and perhaps $15 trillion total “through the end of the decade” on a rough apples-to-apples basis.

While it’s a bit funny to call Jensen a bear by comparison, of course he has no reason to be bearish on an absolute basis. Nvidia has built a monster of a business. Just look at the revenue and gross margin growth.

Even with this incredible progress, revenue is technically decelerating, and despite revenue and earnings being in line with consensus estimates, the stock still dropped in after-hours trading.

The big question on my mind: how strong is the CUDA platform lock-in across various AI scenarios? If AI gets dramatically better at coding, does it become easier for AI labs to move off CUDA? How big of a risk is that? Nvidia’s advantages clearly go beyond CUDA (performance and scale are huge), but CUDA has been a major driver of margins in this boom. I’ll be digging in more here on the show.