TBPN

Jensen on Dwarkesh Rocks the Timeline

April 16, 2026

Happy Thursday.

We’re live on YouTube and 𝕏.

The current thing in tech and business is Dwarkesh Patel’s combative interview with Jensen Huang.

Today’s lineup

Semafor Editor-in-Chief Ben Smith at 11:40 AM
Vital Lyfe Founder & CEO Jonathan Criss at 11:50 AM
Eigen CEO Paul Scherer at 12:00 PM
Actor, Comedian, & Investor Kevin Hart at 12:15 PM
Global Brand Equities Founder & CEO James Morrissey at 12:15 PM
Factory.ai Co-Founder & CEO Matan Grinberg 12:50 PM
Ulysses Co-Founder & CEO Akhil Voorakkara at 1:00 PM
Expo Co-Founder Charlie Cheever at 1:10 PM
Slash Co-Founder & CEO Victor Cardenas Codriansky at 1:20 PM
Cognition Head of Product Growth Theodor Marcu at 1:30 PM

Daily Op-Ed, by John Coogan

Jensen on Dwarkesh Rocks the Timeline

Jensen Huang, the CEO of NVIDIA, spent over 90 minutes in the ring with Dwarkesh Patel. It was electric. The key question was whether or not Nvidia is a car. I’m only half-joking.

Nvidia has been a market leader for years. During the gaming boom, Nvidia was the gold standard for rendering PC graphics, but there was decent competition from AMD. Once the AI boom kicked off, the CUDA ecosystem significantly sped up the development of AI systems and training of AI models. Basically, CUDA is a programming model that enables GPUs to accelerate demanding workloads by parallelizing computation. This created a CUDA moat, because developers were way more productive and the biggest bottleneck to progress was allowing AI researchers to quickly test ideas and scale their experiments up to whole fleets of GPUs and eventually entire datacenters.

But recently, the biggest cost center for AI labs flipped from researcher time to compute capacity. This creates a much larger economic incentive to figure out a way to drive down the cost of the chips used to train AI models. I wrote about this back on Tuesday, October 21st, 2025:

“Not every link in the supply chain can be completely commoditized. Nvidia has an insane amount of power, having ramped full-year revenue from $27B in 2023, to $60B in 2024, to $130B in 2025, all while growing net profit margin from 16% to 56%. All the hyperscalers (and OpenAI) are sort of incentivized to form a bit of an anti-Nvidia alliance to commoditize the accelerator market and drive down those margins a bit.”

Today, the AI chip market is starting to look much less monopolistic. AI coding agents can make it easier to write software that works on non-CUDA chip stacks, and the teams behind competing chips have plenty of resources and economic incentives to bring performance in line with Nvidia. For the past few years, Nvidia has looked like SpaceX’s launch program, an incredible technology with very few viable alternatives. While the products have not degraded (quite the opposite actually), increased competition makes the category look more like the car market than the rocket launch market.

Because the AI buildout does not seem to be slowing down (as long as power can continue to be brought online and data centers continue to find towns that approve of them), demand for chips will presumably grow, but every chip designer (and AI lab) has to be praying that those net profit margins come down. The market did not react negatively to this discussion in the short term, although the price has been basically flat since last August.

We send this email every weekday, right as our show goes live. Subscribe to get it in your inbox.

Headlines

Dwarkesh-Jensen reactions:

➞ Dean Ball: “Dwarkesh/Jensen reveals how inconsistent and un-battle-tested AI acceleration talking points are”…

➞ Tenobrus: “jensen in the dwarkesh interview isn't "wrong" per se. he simply does not care about the truth”…

➞ Sriram Krishnan: “Every person here’s reaction to the Jensen + [Dwarkesh] podcast can be extrapolated *directly* from whether they believe in the frontier labs achieving short timelines for AGI/ASI”…

➞ MetaCritic Capital: “The fundamental problem between Dwarkesh and Jensen is that Jensen doesn’t believe in AGI. Which is insane, but that’s still his underlying assumption”...

Grace Kay: Cursor plans to use xAI's infrastructure to train its Composer 2.5 coding model, according to people familiar with the matter

Anthropic releases Claude Opus 4.7, “our most capable Opus model yet”

Bloomberg: Elon Musk’s lieutenants have reached out to chip industry suppliers for his envisioned Terafab project, early steps in an audacious and likely arduous attempt to break into the production of cutting-edge chips

Bloomberg: Alphabet Poised for $100 Billion Windfall on SpaceX Investment

Bryan Johnson announces Rx platform