SpaceX's $2T Case, Nvidia's Shock Selloff, America Turns on AI, Trump Pulls AI Order, Bond Crisis?

Gavin Baker argues xAI is now a frontier model with Grok 4.3, and the launch of Grok Build as a runtime/harness—combined with Cursor's proprietary coding data and Colossus compute—positions xAI to rapidly close the gap and become a dominant coding AI platform.

“the newest version of Groq 4.3 is on the Pareto frontier for all frontier models. And you're either on the frontier or you're not. And the companies on the frontier are xAI with one build of Groq 4.3... Google 3.1 Pro, and then OpenAI and”▶ Listen · 53:07 ⚑

Gavin Baker is bullish on SpaceX's Elon Web Services / data center business, citing its ability to build data centers faster and cheaper than anyone else, Anthropic's $15B offtake deal as validation, and Cursor's Composer 2.5 breakthrough as evidence of the compute cluster's strategic value.

“they build data centers dramatically faster than anyone else. At a lower cost. And now that you have a clear offtake partner, and I would expect partner to become partners, there is no reason they can't start stamping these data centers”▶ Listen · 48:33 ⚑

ChamathCommentarytoo early to call

Chamath argues SpaceX at $2T is justifiable by underwriting terrestrial data center revenue alone to $100-200B by 2030, plus Starlink scaling to hundreds of millions of users, a growing AI/compute business with a unique capital and execution moat, and Elon's unique ability to create civilizational-scale innovations.

“terrestrial data centers alone are $100 or $200 billion of revenue by 2030, 2032. Just— and that means just building it. So already you're buying it at 20 times revenue just for that business.”▶ Listen · 1:03:33 ⚑

ChamathCommentarytoo early to call

Chamath argues that the feared DSA/domain-specific architecture disruption to Nvidia is actually happening inside Nvidia itself, as evidenced by their co-design programs with every major lab, making Nvidia's competitive position stronger than bears anticipated.

“the reality is that that DSA market evolution is actually happening inside of NVIDIA. That's what's so insane to me. That was my takeaway from the quarter as well, which is like, holy shit, these guys actually have domain-specific architect”▶ Listen · 1:19:14 ⚑

Gavin Baker argues Nvidia is growing faster than hyperscaler CapEx and faster than Broadcom on a comparable basis, dispelling the share-loss narrative; trading at a low-to-mid teens multiple of real earnings it is cheap relative to its position, and its $20B CPU business announcement signals unique co-design advantages that competitors can't replicate.

“I think NVIDIA is probably at a low teens multiple of kind of low to mid-teens multiple of real earnings. That'd be a buy-side consensus, not a trader's number.”▶ Listen · 1:28:57 ⚑

Gavin Baker argues Anthropic's extraordinary growth, recent EBIT profitability, and massive proprietary coding token dataset position it as one of the dominant AI platform companies with strong and accelerating fundamentals.

“The success is extraordinary. It's undeniable. I think the fact that they are now— they were EBIT positive per the Wall Street Journal in the most recent quarter— is a really important fact for kind of the whole AI narrative.”▶ Listen · 4:35 ⚑

Gavin Baker argues Cursor's Composer 2.5 Pareto-dominant benchmark result—achieved after just 3-4 weeks of RL on Colossus 2—proves the unique value of Cursor's proprietary coding token dataset and points to an even larger leap when a new base model is trained with that data.

“this jump from Composer 2 to Composer 2.5 showed us is that when you do an appropriate amount of reinforcement learning using that data... this is 3 or 4 weeks and it is Pareto dominant.”▶ Listen · 50:40 ⚑

Gavin Baker argues Nvidia's quarterly results—by validating GPU useful lives of 10-15 years via aggregated inference architectures—saved CoreWeave and other neo-clouds from the Michael Burry bear case that amortization schedules were too long, strengthening their business model.

“those older GPUs, they have a useful life for 10 or 15 years. And this means that you can finance GPUs. I think Corey's lowest financing, I can't forget if it's 6 or 7%. 6%. It's going to come down.”▶ Listen · 1:21:08 ⚑

Guest argues that US natural gas (NG1) is down this year while LNG prices in the rest of the world are up 100-200%, making America relatively advantaged as natural gas is the dominant input for US electricity and manufacturing. A closed Strait of Hormuz reinforces this relative US advantage.

“NG1. It is down this year. The input cost for electricity in the rest of the world...LNG is a very important one, and it's up 100%, 200%. And so the Strait of Hormuz is absolutely bad for everyone, but relatively good for America.”▶ Listen · 1:30:53 ⚑