Oct 27, 2026 · 41 min
Anthropic’s S-1 sharpens the case for an AI capital boom
Gavin Baker on Anthropic's S-1, GPU Prices, and SpaceX
The episode connects enterprise AI adoption and rising inference demand to GPU economics, industrial constraints, and the prospect of broader public ownership.
- 1Anthropic’s S-1 suggests enterprise AI adoption is broader and more commercially mature than headline spending concentration implies.
- 2GPU scarcity, forward contracts, and cloud repricing are reshaping the economics of AI infrastructure and infrastructure financing.
- 3Potential IPOs from Anthropic and SpaceX could redirect substantial capital toward AI and widen public participation in its growth.
Don't miss
Baker connects possible Anthropic and SpaceX IPOs to a much broader distribution of AI ownership and capital.
The brief
Gavin Baker joins John Coogan to examine Anthropic’s S-1, focusing on profitability, cash generation, and Fortune 500 adoption as evidence of broader enterprise demand.
The discussion moves from private-market growth at Fireworks and Cursor’s inference demand to specialized AI agents, including GrokBot’s role in research and information flow.
GPU rental markets are tightening, while fixed-price contracts and forward commitments expose the financing risks facing neoclouds and cloud providers.
Biology and drug discovery offer a major but slow-moving AI opportunity, with progress in drug repurposing competing against long development timelines and infrastructure costs.
Baker’s broader bull case links AI to reindustrialization, power generation, and constrained resources rather than treating models as a standalone software story.
The closing argument is about ownership: IPOs from Anthropic and SpaceX could unlock new institutional capital and give far more people direct exposure to AI’s gains.
What was said on this episode
33 statements · 22 positive · 6 negative · 1 mixed · 4 neutral
Anthropic’s S-1 was an upside surprise.
“I do think the Anthropic S1 was pretty clearly an upside surprise.”
Listen at 0:48
Anthropic’s profitability, cash generation, and unit economics are encouraging.
“the profitability, the cash generation, the unit economics, they're all very encouraging.”
Listen at 0:55
About half of Fortune 500 companies have adopted Anthropic.
“half of the Fortune 500 has adopted Anthropic.”
Listen at 1:09
Enterprise AI diffusion remains very early.
“very early”
Listen at 2:10
Anthropic has a highly diversified spending base.
“it shows a very diversified spending base for anthropic”
Listen at 4:52
Anthropic’s October ARR is likely to be strong.
“I'm pretty skeptical that they would have put the blank space for the October ARR in the S1 if they didn't know it was going to be pretty good.”
Listen at 5:58
Markets should move away from monthly ARR reporting.
“I do think it'll be great for the market to get away from these monthly ARR things.”
Listen at 6:13
Fireworks’ valuation multiple was falling daily during fundraising.
“the valuation was going, you know, the multiple was effectively going down, you know, every day.”
Listen at 7:31
SpaceX has been insourcing Cursor’s compute.
“SpaceX has been insourcing the cursor compute.”
Listen at 7:56
GrokBot is the most impactful AI product for Baker and his firm.
“Rockbot is the most impactful AI product to me and to the firm that we've had yet.”
Listen at 8:51
GrokBot has been transformative for professional use.
“grokbot has just been transformative”
Listen at 10:43
GrokBot effectively routes information to the right people at the right time.
“grokbot is so good at that”
Listen at 10:55
GPU rental prices have risen daily during October.
“the price to rent essentially any kind of a gpu has gone up every day”
Listen at 12:57
Portfolio companies struggle to secure compute through 2027.
“even with all of that supply coming on it is hard for the companies in our portfolio to find compute even well into 27.”
Listen at 14:37
GPU financing conditions may concentrate capacity among competent operators.
“I hope it concentrates capacity and supply in the competent operators”
Listen at 15:10
CoreWeave operates GPUs very well.
“CoreWeave runs GPUs well. They are very, very good.”
Listen at 15:24
CoreWeave’s long-term compute contracts are now far below market.
“They just sadly locked in a vast amount of compute and long-term contracts. that are now way, way under market.”
Listen at 15:28
Spot-market GPU sales can produce sub-one-year paybacks and large prepayments.
“If we are willing to sell into the spot market, we can have paybacks well under a year, get big prepayments.”
Listen at 16:24
CoreWeave is increasing spot exposure and reducing long-term offtake agreements.
“CoreWeave, I think, is very aware of this. They're trying to get more spot exposure. They're trying to do less of these big long-term off-take agreements.”
Listen at 16:41
CoreWeave, Crusoe, Nebius, SpaceX, and hyperscalers have proven execution records.
“the companies i named i think they're gonna I mean, the future is always uncertain, but these are companies with a proven track record of execution.”
Listen at 19:32
Bringing a compute cluster online in mid-to-late 2027 will remain very expensive.
“you have to pay a truly vast amount to bring on a cluster in mid to late 27.”
Listen at 20:58
AI has already saved many lives.
“AI has already saved a lot of lives.”
Listen at 22:51
Anthropic dedicates 74% of its compute to training.
“Anthropic is dedicating 74 of their compute training”
Listen at 28:11
Reduced training demand could redirect substantial compute supply to inference.
“if all this compute is being dedicated towards training and for any reason somebody decides you need less compute for training, that's a lot of supply that could hit the inference market.”
Listen at 28:21
AI is contributing to American reindustrialization.
“We're re-industrializing America.”
Listen at 30:19
AI is driving a major increase in U.S. electricity-generation capacity.
“We're, you know, massively increasing our electrical generation capacity”
Listen at 30:21
Rare-earth constraints primarily involve environmentally dirty refining, not scarcity.
“the problem with rare earths, it's not that they're rare. It's that refining them is a very dirty process.”
Listen at 33:51
SpaceX and Anthropic IPOs will unlock substantial mutual-fund buying power.
“actually going to free up a huge amount of buying power”
Listen at 35:49
Hyperscaler and neocloud revenue growth will continue accelerating.
“I think you're going to see a pretty continuous acceleration in the sum of the big hyperscaler cloud revenue growth, the neoclouds.”
Listen at 37:41
The peak of external AI financing may be approaching.
“I think we're getting close.”
Listen at 38:18
Baker will revise his AI views if the facts change.
“I will change my mind when the facts change”
Listen at 38:39
Directly giving people money is better than allocating it through a foundation.
“I think that's way better for America than setting up some foundation”
Listen at 40:05
Everyone should receive a stake in capitalism.
“Give everybody a stake in capitalism.”
Listen at 41:09
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.