Full episode · 41 min
Gavin Baker on Anthropic's S-1, GPU Prices, and SpaceX
Gavin Baker argues that compute demand remains ahead of supply and that AI’s economics and new capital can sustain the buildout.
TBPN · John Coogan & Jordi Hays
- Gavin Baker is managing partner of Atreides Management.
The best minutes
- 12:46GPU supply is tight despite the capacity coming online7 min
- 7:56AI agents are helping investors route information faster4 min
- 34:35Public listings could unlock more capital for AI4 min
Also worth watching
- 0:18Anthropic’s S-1 shows a broader customer base than expected8 min
- 31:23Rare-earth processing, not scarcity, is the bottleneck3 min
- 38:51Baker favors giving people shares over foundation grants2 min
Skip the conference sign-up and format discussion, then Baker’s discussion of online criticism, data-center advocacy and AI’s benefits.
Summary · 3 min read
Gavin Baker is managing partner of Atreides Management. He said Atreides led Fireworks’ last funding round and Atreides and Valor co-led Crusoe with Mubadala; earlier, he invested in energy and natural resources at Fidelity.
Baker read Anthropic’s S-1 as an upside surprise, with encouraging economics and a customer base he considers more diversified than critics suggest. He argued that compute remains scarce despite plans for substantial new capacity, and that pricing, financing and deployment speed will reward operators that can bring GPUs online. He described AI agents as useful in his investment work and an existing-drug use case as evidence of AI’s potential in biology. He also expects public listings and cloud cash flow to help fund the next buildout, while urging broader public participation in AI’s gains.
Minute by minute
- 0:18Baker calls Anthropic’s S-1 an upside surprise
Baker called the S-1 an upside surprise, saying Anthropic had burned only $10 billion and that its profitability, cash generation and unit economics were encouraging. He cited an estimate that about half of Fortune 500 companies had adopted Anthropic, while its spending represented 0.1% of their expenses.
- 2:36Baker sees a broader customer base
Jordi Hays questioned whether AI spending was concentrated among a small set of companies. Baker countered that Anthropic’s S-1 showed a diversified customer base, citing 250 accounts spending more than $10 million; he also said open-weight-first providers such as Fireworks, Baseten, Together and Modal were growing faster than OpenAI or Anthropic.
- 7:35AI agents help route investment information
Baker said Grok Bot was the most impactful AI product for him and his firm so far, and that he had built 17 bots. He said it helps get the right information to the right person at the right time; he had not needed personal-assistant products because his human assistant, Susan, was exceptional.
- 11:46Compute prices are rising into 2027
Baker said GPU rental prices had risen every day in October and the spot market had tightened. Citing Silicon Data founder Carmen Lee, he said forward pricing was in contango and finding a cluster for 2027 remained difficult despite plans for new capacity.
- 14:56Spot sales reward experienced operators
Baker argued that market conditions favor operators that can finance GPU purchases with prepayments and spot sales. He said CoreWeave ran GPUs well but had locked in long-term compute contracts at rates now well below market; he said the company was seeking more spot exposure.
- 18:20Preassembled clusters could speed deployment
Baker said the DSX pod has seven racks and comes preassembled, which could shorten the months it can take some operators to energize GPUs. He argued that faster deployment would reduce working capital tied up in equipment that is not yet generating revenue.
- 21:34AI can repurpose existing drugs
Baker said AI had helped identify a 50- or 60-year-old drug that might effectively treat a rare, devastating disease affecting about 10,000 children a year. He said a friend had done this and that AI had already saved lives, but did not share the friend’s story.
- 26:59Baker argues for data center benefits
Baker said he tries to use criticism on X to test his views, citing a bearish argument that Anthropic devotes 74% of its compute to training and that lower training demand could send supply into inference. He also argued that data centers can benefit towns and blue-collar workers, tying AI investment to US reindustrialization and expanded electricity generation.
- 31:23Rare-earth refining needs cleaner processes
Baker said he tracks energy, natural resources and industrial infrastructure but has not made them a large part of his portfolio. He said the firm had invested privately in a company with a chemical process for refining rare earths, arguing that the challenge is refining them rather than finding them.
- 34:35IPOs could unlock more AI investment
Baker said SpaceX and Anthropic going public could unlock mutual funds’ buying power for private markets once their shares were public and unlocked. He expected cloud revenue growth and operating cash flow to accelerate, and said cash flow could finance more of the 2027 buildout than people expect, with external financing possibly nearing its peak.
- 39:04Baker favors direct public ownership
Baker said broad public ownership of SpaceX, including giving every child in America a share, would be better than having a foundation decide how to allocate the money. He encouraged other wealthy Americans to follow examples he attributed to Elon Musk, Michael Dell and Brad Kirschner.
Published Oct 27, 2026.
Transcript · 32 min read
TBPN, hosted by John Coogan and Jordi Hays, with Gavin Baker.
From the transcript TBPN publishes with the episode.
Next guest is Gavin Baker from Atreides Management, the managing partner. I have Heard of him. He's been on the show before. The Baconator. We're gonna bring him in to the TBPN UltraDome and catch up with him. How are doing, Gavin? Good to see you.
I'm awesome, man. Happiness is a choice. I'm only ever fantastic, awesome, superb, phenomenal.
“I was actually shocked how diversified their customer base actually was.”
Read the whole transcript
Yep. And I mean, the x cursor growth, I think, you know, is is what's really extraordinary enough just because, know, SpaceX has been insourcing the Cursor compute.
“So much of investing is making sure that the right information gets to the right person at the right time.”
So... How... So I think, like, you know, rewind. This week is, you know, Anthropic s one week, and so there's there's a lot of focus there. Last week, it felt like the focus was... Like, the general attention was towards how much supply is coming online next year, how many gigawatts SpaceX is gonna bring online. How are you... What's your what's your, you know, generally, you've maintained a sort of bullishness that that some other people on the timeline have lost here and there. But how are you processing, like, how the how the how the industry can sort of absorb all of that supply if we hit the higher end of some of the ranges being discussed? And and and is it just Jevan's paradox, you know, again and again and again Mhmm.
And again until we finally... Every everyone says like, you know, everyone likes to say it, but it's hard to fully internalize.
“But even with all of that supply coming on, it is hard for the companies in our portfolio to find compute even well into '27.”
So I have a follow-up on that. One is just one click higher from the Neo Clouds or the Neo Labs, and there's been this speculation that Neo Clouds might NeoLabs might sort of pivot into NeoClouds. Is that dynamic gonna look like they signed a multiyear offtake agreement with a NeoCloud, and then they want us to just sell into the spot market? That's what's gonna happen?
Yes. I mean, what Sarah Fryer said Yeah. And Sarah Fryer is an exceptional CFO. Yeah. Said, you know, everybody made fun of us for signing all these compute commitments and we could sell them all at a three, four, five x premium. Today. Yeah. And she was being literal. Yep. Yep. And so if you if you signed a long term compute contract Yep. Really at any point prior to May 2026 Mhmm. You were way in the money. Yep. You're buying compute for, you know, maybe Yeah. 10 or fifth... 10 or $15 a watt. Yeah. And you could sell it at a massive premium to that today. Yep. And then the point the point, you know, that Carmen made about, you know, the the forward curve being in contango is actually really important and just like... It's not just that the spot market is tight today. You have to pay vast amount even with all this capacity coming online. You have to pay a truly vast amount to bring on a cluster in mid to late twenty seven. And I think that's something that just maybe public markets are not appreciating even though, you know, it's not all of this is public.
You know, Sarah Fryer said that. Yeah. You know, Yeah. Yeah. All these inference clouds have kind of given press releases. You hear a bunch of people on x talk about how hard it is to find compute into '27. Yep. Carmen Lee stands up and says, you know, like the market in '27 is is crazy strong. Yeah. So it's all out there but I think for whatever reason it's not legible to the public Sure.
Jordan? What... Talk about your outlook on biology as a as a... As an area that that the labs are investing in. It feels very exciting and that there's, you know, tremendous potential, but I think that overall concern is around timelines. It's like there's very real costs, like, coming online, infrastructure costs, and opportunities in biology seem like they're very real, but, may take, you know, years to get these, you know, will almost certainly take years to get drugs in market and actually drive, you know, revenue that would then finance.
Yeah. I mean, we'll we'll see how long it takes. You know, Montana has an interesting new program for for new drugs, I think is helpful. But I think a really big unlock. We we have a lot of compounds and people have used AI to figure out that you can use maybe a drug that we've had for fifty or sixty years for a very rare disease that's devastating for like ten thousand kids a year and effectively cure that disease. And a friend of mine did this and it's, you know, it's it's it's his story to tell, so I'm not gonna tell it. But AI has already saved a lot of lives. And I do think, you know, the AI industry, it's, you know, they're like negative five out of 100 from like a PR. A 100%.
But they... We gotta be more like pharma, the popular kids. Yeah. Exactly. Well, no. They need to start telling these stories. Yeah. Yeah. Yeah. 100%. And by the way, are... You guys are coming to tomorrow at. We are. Of course.
We are. November seventeenth, eighteenth. We'll be there with The lineup is unbelievable. The lineup is crazy. Jensen Huang, David Solomon, Mike Rowan, Lisa Hsu, Stan Druckenmiller, so many more... So many people. Yep. But the one thing I just wanna say is Yeah.
I get, like, 25 emails or texts today saying, hey, man. Can I come to your conference? Yeah. And the answer is, you can just sign up. Yeah. You can just sign up. Anyone can come. Now, there's gonna be a security screen Sure. Because of some of the people who are coming and I do, you know, there are unannounced speakers that are, you know, comparable in caliber to some of those you reference. So we're gonna have a lot of security at the conference and everybody's... You'll be screened. But you can just sign up. You do not need to ask me. You don't need to ask Antonio. You can just come. And it's like the whole goal is to kind of democratize investing. And in in addition to those companies, know, we are... We're gonna have Lynn from Fireworks Yeah.
Chase from Crusoe Yeah. Scott from Cognition. We're gonna have Max from Lagora. Yeah. We're gonna have a lot of companies. Yeah. Base ten. Two in. Two in from Base 10. Yeah. We are gonna have a lot of these CEOs there that are gonna be on a stage and then we're gonna have, you know, dozens of other private CEOs. And, you know, it's a... It's an attempt to kind of, like, open up the aperture of, you know, information for retail investors and democratize this. So, like, if you're wondering about coming, just sign up. You don't need to ask me.
Tomorrow x summit dot com, not Gavin Baker's email. Yes. You can go to a URL on the public. Can get you can get Gavin to sign And they will email you when you sign up. Yes. Yeah. Yeah. What what what do you think will change in the content and the conversation between last year's event and this year's event? What are you hoping to get across? We did... This is the first time... This is the first one. Oh, sure. Yeah. Yeah. Oh, sorry. Yeah.
Yeah. We decided we're gonna have, you know, we're gonna have bysiders doing the fireside chats. We're gonna have... You know, management teams are gonna have the opportunity. We've... If you were an active contributor on x, Substack, Reddit Sure. You get to come for free. Yeah. And then management teams are gonna be able to do a breakout session with all of those x accounts. So if you're if you're one of the, you know, four or 500, you know, kind of, you know, financial commentators that we've given a scholarship to, you're gonna get, you know, management teams are gonna opt in and you're gonna be able to kinda have your own session with the management team to ask your questions, then you're gonna be able to go write what you learned for your audience.
And I'm... You know, I think we have a really high quality, you know, I mean, I I picked most of them. And you don't have to be bullish. You know, I want... You know, I love skeptics. I love bears. Yeah. You just you just have to be polite and nice and constructive. And you can be bearish while being polite, nice and constructive. And, you know, we're... We have loads of we have loads of skeptics. We have loads of optimists. Like, I think it's gonna be great. And then I I hope every CEO does, you know, you know, kind of a hit on your show after they're on stage or before. Yeah. I'm... Might TBPN be the episode ever.
I think so. Let's see. Yeah. Do you deal with how how do you deal with all the all the attention? Positive and negative? That work? For me? Yeah. For you. For you. Oh, I know. I just feel like I... You I... You you talked about your your mindset right at the beginning of the show. Just, you know you know, it's a choice to be happy. But I feel like when it comes... When the the the spotlight, you know, once every week shines directly on you. There's a lot of love, a little bit of hate sprinkled There's... In The whole the full spectrum of human emotions.
We we can call it what it is. I mean, there there there are a lot of there are a lot of detractors, let's say. I just say, look, Antonio, who I'm doing the conference with, he says being on x is an exercise in cultivated equanimity. I don't take any of it personally. I'm not on x to be right. Mhmm. I'm... Right now, I would, you know, I'm I'm a responsible optimist, but I'm trying to be open minded to new negative facts and constantly search them out. But I'm really on x, you know, to learn, and I... You know, all of the hate, like, I try to sift through it for, oh, wow. That is actually a very good compelling bearish argument. You know, like, there was an interesting post. I forget, it was like PE analyst or LBO analyst. And it just said, hey, you know, the fact that Anthropic is dedicating 74% of their compute to training, you could read it... On the one hand, it's bullish because they could be generating a lot more revenue.
On the other hand, 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. Hey. Wow. That's like a really interesting, you know, bearish perspective. And so, you know, I'm human. I don't... Like, I'm human, but it's also just like, I'm here to learn. I wanna get to the right answer. I'm sharing my views. If somebody shares a, you know, a helpful negative perspective, I try to thank them. I need to go, like, thank that that person. LBO analyst, I think it was. I didn't have time. But, yeah, it's it's an exercise of cultivate... Cultivated equanimity.
You can't you can't can't take it personally. Yeah. But I do... I am trying, you know, doing what little I can, you know, has like Gavin Baker, you know, I don't have millions or tens of millions of followers, but I am trying to help paint, I think, more optimistic vision for AI, doing what I can to help the American public understand that, hey, like, you actually want a data center Yeah. In your in your town. You know? One of my favorite exchanges was somebody said, well, I'll wait until they start building data centers in the wealthiest county in America. Well, Loudoun County is the wealthiest county in America. They have the highest per capita data centers.
Yep. Virginia. US East. Amazon. Yep. So just understanding that, like, I understand why, you know, some people are worried. There... There's been a lot of fear. There's... You know, that we've all seen scary movies. But a lot of awesome things are happening today Yeah. Because of AI for ordinary Americans, for blue collar Americans, for these small towns. We're re industrializing America. We're, you know, massively increasing our electrical generation capacity, which is you cannot re industrialize without power. That's job number one. So I think a lot of good things are happening, and I'm just trying to do what little I can to at least tell that side of the story, and then be pretty balanced. Like, you know, Anthropic, I called it like I saw it in the summer. They were losing market share.
Mhmm. I'm calling it like I see it in October. They're reaccelerating Yeah. And gaining share. Yeah. The s one, it was an upside surprise to me. I'm trying to call balls and strikes Sure. Because my job is to just get it right.
Yeah. Jordan And it's funny because, you know, people will get so flustered by, you know, seeing different posts on x. Sure. But imagine how boring x would be if no one ever took a stance. Oh, yeah. It was just like, well, it might be this.
Yeah. It might be that. Yeah. Yeah. Yeah. And it might be this. No. No. But have you considered that? It's over or or we're back. This is the best thing ever in the world. Because then we can have a debate. At least you took a stance. Yeah. Yeah. It's great. Yes.
“the problem with rare earths, it's not that they're rare. It's that refining them is a very dirty process.”
“I think you're gonna see a pretty continuous acceleration in the sum of the big hyperscalers cloud revenue growth, the neo clouds.”
Totally. Yeah. No. No one can complain about that. That's just, yeah, investing in the future. AI skeptics will be... In a bad place.
In a horrible place. We're gonna have to check on check on them. For sure. You're gonna have to go around.
And and look, maybe I'm maybe I'm wrong and, you know, but like, I will change my mind when the facts change and I'm just gonna try and keep calling balls and strikes. And then trying to, you know, be an advocate to help understand ordinary Americans that this technology can and should be awesome for them.
“I think that's way better for America than setting up some foundation”
Yeah. No. I completely agree. It's great. Well, thank you so We'll see very soon. We'll see you. Yeah. Tomorrowxsummit.com. Check it out. November seventeenth, eighteenth, Austin, Texas. We'll see you there. Have a great rest of your day. Cheers, guys. Good to see you. Goodbye.
