The AI Cold War, Signalgate, CoreWeave IPO, Tariff Endgames, El Salvador Deportations
Chamath Palihapitiya (host), Jason Calacanis (host), David Sacks (host), Gavin Baker (guest), David Friedberg (host), Narrator, Narrator, Narrator, Narrator, Narrator
In this episode of All-In Podcast, featuring Chamath Palihapitiya and Jason Calacanis, The AI Cold War, Signalgate, CoreWeave IPO, Tariff Endgames, El Salvador Deportations explores aI Arms Race, CoreWeave IPO, Tariffs, and Tough Immigration Choices This All-In Podcast episode features investor Gavin Baker joining the besties to unpack Nvidia’s massive GPU product transition, the CoreWeave IPO, and how export controls are shaping an emerging AI ‘cold war’ with China.
AI Arms Race, CoreWeave IPO, Tariffs, and Tough Immigration Choices
This All-In Podcast episode features investor Gavin Baker joining the besties to unpack Nvidia’s massive GPU product transition, the CoreWeave IPO, and how export controls are shaping an emerging AI ‘cold war’ with China.
They debate the real economics of AI infrastructure, the non-commodity nature of running huge GPU clusters, and how AI agents could radically shrink headcount for complex projects while threatening incumbent software and services.
The conversation then pivots to Trump-era economic strategy—tariffs, deregulation, deficit reduction, and the political risks of an aggressive policy ‘experiment’—before tackling Signalgate, government communication ethics, and controversial mass deportations to El Salvador’s notorious CECOT prison.
Throughout, they repeatedly return to the tension between ends and means: national competitiveness and safety versus due process, transparency, and human rights, and how execution and communication could make or break the administration’s ambitious agenda.
Key Takeaways
Nvidia’s spike in accounts receivable is largely explained by a historic product transition, not hidden demand manipulation.
Gavin Baker argues Nvidia is going through the largest semiconductor product transition ever—from Hopper to Blackwell GPUs—akin to an iPhone upgrade where you must also rebuild the house’s electrical and cooling systems. ...
Get the full analysis with uListen AI
Nvidia’s investments in CoreWeave and other ‘neo clouds’ are about diversifying buyer power, not propping up demand.
Baker contends Nvidia would have sold the same volume of GPUs without equity stakes in CoreWeave and similar players—they’d simply have gone to hyperscalers like Meta, AWS, and Microsoft. ...
Get the full analysis with uListen AI
Running massive GPU training clusters is harder and less ‘commodity’ than many investors assume.
CoreWeave is widely criticized as a debt-laden, commodity GPU cloud with concentrated revenue from Microsoft. ...
Get the full analysis with uListen AI
AI agents may unlock entirely new, complex projects rather than just automating low-level work.
While Baker emphasizes that fully-realized agents will make compute the bottleneck and drive high ROI for Blackwell, Friedberg argues the bigger unlock is enabling small teams to execute projects that today require dozens of highly specialized experts. ...
Get the full analysis with uListen AI
China’s AI and chip trajectory is constrained in the near term but dangerous over a decade-long horizon.
Export controls on advanced GPUs are adding friction, but Baker notes they also massively incentivize China to build its own semiconductor ecosystem and innovate algorithmically—DeepSeek being a prime example. ...
Get the full analysis with uListen AI
Trump’s economic agenda is a high-risk ‘grand experiment’ balancing tariffs, tax cuts, deregulation, and spending cuts.
Chamath and Baker outline the administration’s theory: use tariffs to ‘level-set’ historic trade imbalances and incentivize reshoring, while simultaneously cutting taxes (especially for under-$150k earners), slashing regulation, and attacking waste, fraud, and abuse in federal spending. ...
Get the full analysis with uListen AI
Hardline immigration and security measures risk undermining the administration’s mandate if they trample due process and human rights.
The group raises alarm over mass deportations of alleged Venezuelan gang members to El Salvador’s CECOT supermax under the Alien Enemies Act, citing clear cases where people with tattoos (e. ...
Get the full analysis with uListen AI
Notable Quotes
“There’s never been a product transition like this in the history of semiconductors… the only precedent for this on planet Earth is the iPhone.”
— Gavin Baker
“Everybody thinks it’s easy, but to synchronize tens of thousands of GPUs where they’re melting or cables are being unplugged… it may not be the commodity that everyone thinks it is.”
— Gavin Baker
“If these agents can scale the OpEx, the actual load of making something will go down by an order of magnitude… that is incredibly disruptive because the existing incumbents cannot compete with that cost scale.”
— Chamath Palihapitiya
“Human rights are a core American value… if innocent people were sent to this prison, mistake. And hopefully it gets rectified.”
— Gavin Baker
“We’re 4.5% of the way through this second term… if they want to accomplish their goals, they need to execute at a high level and communicate clearly and effectively.”
— Gavin Baker
Questions Answered in This Episode
Gavin, you described the Hopper-to-Blackwell shift as the biggest product transition in semiconductor history—what specific operational or financial metrics would you watch over the next two quarters to validate that Nvidia has successfully managed this transition without hidden demand problems?
This All-In Podcast episode features investor Gavin Baker joining the besties to unpack Nvidia’s massive GPU product transition, the CoreWeave IPO, and how export controls are shaping an emerging AI ‘cold war’ with China.
Get the full analysis with uListen AI
You argued that running massive GPU clusters is far from a commodity business; can you walk through a concrete post-mortem from a major training run (even anonymized) that illustrates the kinds of failures and interventions that separate CoreWeave-level operators from ‘me-too’ GPU renters?
They debate the real economics of AI infrastructure, the non-commodity nature of running huge GPU clusters, and how AI agents could radically shrink headcount for complex projects while threatening incumbent software and services.
Get the full analysis with uListen AI
On AI agents and MCP, what are the first real-world, non-demo projects you’d expect a 3–5 person team to execute using agents that would be utterly impossible for them to do today—and how soon do you expect those case studies to appear?
The conversation then pivots to Trump-era economic strategy—tariffs, deregulation, deficit reduction, and the political risks of an aggressive policy ‘experiment’—before tackling Signalgate, government communication ethics, and controversial mass deportations to El Salvador’s notorious CECOT prison.
Get the full analysis with uListen AI
Your tariff-and-tax strategy relies heavily on DoJ/‘DOGE’ uncovering up to a trillion dollars in waste and fraud; what are the top two or three concrete programs or contract categories you would publicly audit first to build trust that this isn’t just hand-waving about ‘fraud’?
Throughout, they repeatedly return to the tension between ends and means: national competitiveness and safety versus due process, transparency, and human rights, and how execution and communication could make or break the administration’s ambitious agenda.
Get the full analysis with uListen AI
Regarding the El Salvador deportations, if you were designing a legally robust but still fast-track process for handling suspected foreign gang members, what minimum evidentiary standard, independent review, and appeal mechanisms would you put in place before allowing anyone to be transferred to a facility like CECOT?
Get the full analysis with uListen AI
Transcript Preview
When I wasn't able to use the R word, I would use deficiente for two people.
Mm.
Jason-
And Fahamid.
... and Friedberg's dog, Marshall Friedberg.
(laughs) Oh, come on.
I can't stand Marshall Friedberg.
Come on. He's...
That little bastardino jumps on the table, eats the fucking nuts.
The worst. The worst.
I hate Marshall Friedberg. I hate him.
(laughs)
I can't stand this fucking bastardino.
Look at the little guy. He's just a little guy.
Look at, look how he sits.
He's so cute.
He sits like a moron.
He's so cute.
He sits like an, he sits like a deficiente.
(laughs)
Look at this dweeb-
Oh my God.
... and this deficiente. Now, Nick, show them my dogs. Beautiful.
Beautiful.
Valentina Zanini, she's the breeder of breeders. Look at these two beautifully elegant... Oh. Look at Aki and Dukie.
Those are my dogs.
Look how, look how well behaved they are. You don't see them jumping on the table to eat the main course, Friedberg.
No. No.
Marshall almost ruined our Christmas dinner. This is why I'm holding a real grudge against Marshall Friedberg.
He ate Shemuf's nuts.
No, and Alison, it caught Alison off guard. She's like, "Marshall Friedberg, get off the table." (laughs)
Marshall. (laughs) I just love that he's got a full name.
(laughs)
Marshall Eugene Friedberg.
Let your winners ride.
Rain Man David Sacks.
And I said we open source it to the fans and they've just gone crazy with it.
Love you bet, honey.
Queen of quinoa. I'm going all in.
All right, everybody. Welcome back to the number one state sponsored... I'm sorry, All In podcast in the world.
Oh. Oh.
The number one podcast. Oh, sorry. Stray bullets. Here we go.
Oh.
Back on the program. Our guy, Gavin Baker is here. You know him from Atrides Management, does private and public, four billion under management. And a lot to talk about with you.
GB1. Good to have you here.
GB1 is here. Solo dolo. Core Weave IPO happening soon, uh, by the time you get this. And NVIDIA, you were there at NVIDIA. You were the one analyst, Gavin, that Jensen pulled up. You're loved by Jensen. What's it like to be loved by Jensen? Tell us everything.
Well, so first he did, um, there was, uh, Alfred fo... Was from Sequoia.
Mm-hmm.
And then they had a, a nice sell side guy whose name is escaping me at the moment. But it is technically true, I was the only public equity investor on the buy side that Jen- Jen- Jensen asked. I've known Jensen for 25 years. He's, he's kind of the same guy he ever was, like, just maybe slightly calmer.
Install uListen to search the full transcript and get AI-powered insights
Get Full TranscriptGet more from every podcast
AI summaries, searchable transcripts, and fact-checking. Free forever.
Add to Chrome