Can the AI Industry Regulate Itself? Stripe Wants PayPal, China Catches Up, NY Bans Datacenters
- 01The SRO as the Least-Bad Option for AI Governance
- 02Anthropic's Regulatory Capture Strategy Is Now Hiding in Plain Sight
- 03The Stripe-PayPal Deal as the Template for AI-Driven Digital Roll-Ups
- 04The Real Strategic Prize: Bypassing Visa and Mastercard
- 05The AI Token Cost Crisis: A CFO Time Bomb
- 06Data Centers Are the New Battlefield
1. Key Themes
The SRO as the Least-Bad Option for AI Governance
The episode opens with Demis Hassabis's proposal for a FINRA-style self-regulatory organization (SRO) for AI, drawing broad industry support. The hosts debate whether it's genuinely good policy or merely the opening bid in a longer regulatory war.
"An SRO, again, a self-regulatory approach would be infinitely better than creating a new government agency that I think would rapidly become a DMV for AI... The government does not have the expertise to evaluate AI models. The criteria are changing too rapidly." — David Sacks 00:07:16
David Sacks laid out five explicit conditions: broad industry representation (including startups and open source), review limited to true frontier models only, focus on catastrophic risks (cyber and CBRN) only, voluntary before mandatory, and must be a substitute — not an addition — to new regulatory agencies.
"If it's just additive, then it defeats the purpose and there's no real reason to support it. So, again, I think this has to be a substitute, not an addition to a bunch of new regulatory structures." — David Sacks 00:10:26
Anthropic's Regulatory Capture Strategy Is Now Hiding in Plain Sight
What Sacks called a "spicy take" in October of last year — that Anthropic is running a regulatory capture strategy — is now documented by Politico.
"There was an article in Politico just the other day. It's called 'Inside Anthropic's state-by-state plan to ratchet up AI rules.' What it says is, quote, AI giant Anthropic is pursuing a strategy of one-upmanship that encourages states to impose increasingly tougher AI guardrails rather than a line around a single set of regulations." — David Sacks 00:17:05
"Brand yourself as a safe AI company. Number two, ban unsafe AI. Three, profit. Yeah. There you go. That's the strategy. Kind of brilliant." — David Sacks 00:21:13
Sacks also noted that Anthropic is funding Public First, an NGO opposed to data center construction, with Dario making his first seven-figure contribution — even though compute constraints, not demand, are Anthropic's number-one growth limiter.
The Stripe-PayPal Deal as the Template for AI-Driven Digital Roll-Ups
Friedberg frames the Stripe bid for PayPal not as a one-off but as the second dot on a line — alongside Ryan Cohen's eBay bid — of AI-native operators targeting stalled first-generation digital businesses.
"Folks that are, call it AI native, are looking at, call it first generation digital native businesses that have become mature and old and stale and aren't run by the founders anymore and have not yet realized the opportunities with AI... And when you take a look at those businesses as a modern day AI operator, you're like, what the hell? This thing is so underutilized." — David Friedberg 00:24:44
"I will make a prediction. I think that eBay and PayPal are probably the beginning of a wave of mega deals of, call it flaccid, digital businesses... that can be revived with the blue chew of capital and the right operator." — David Friedberg 00:27:17
The Real Strategic Prize: Bypassing Visa and Mastercard
Chamath connects the dots across Stripe, Block, PayPal, Braintree, Venmo, and stablecoin infrastructure to argue the combined entity amounts to a credible attack on the card networks.
"You are creating a competitor to Visa and MasterCard. Because you now have upwards of 600 or 700 million accounts. You have massive stablecoin infrastructure. You have all of the risk management infrastructure that Stripe has built over the last 15 or 20 years." — Chamath Palihapitiya 00:22:35
"PayPal already owns Braintree. So now you have Stripe and Braintree that effectively were competitors that won't be. And then what Block gives you is an entire point of sale infrastructure and you get the cash out. So you put it all together and I think it's a shot across the bow for Visa and MasterCard." — Chamath Palihapitiya 00:35:18
The AI Token Cost Crisis: A CFO Time Bomb
Token pricing spreads — $56/million for Claude versus $0.50/million for Chinese models — are building toward a corporate earnings shock, with Ramp's new token spend management product as a leading indicator of CFO alarm.
"A million tokens on Claude is about $56. A million from Grok is about $1.50. Zuck's is about $1.50. Elon's about $1.00. And the Chinese models are $0.50." — Chamath Palihapitiya 00:48:04
"Token spend among Ramp customers has grown by 21 times. 21 times. Not 21%." — Eric Glyman (via clip) 00:49:38
"Unless you get control of this and you can directly say how much money you're making, this is a bridge to nowhere. It is a money burning furnace." — Chamath Palihapitiya 00:51:03
Data Centers Are the New Battlefield — and Foreign Influence May Be at Work
New York Governor Hochul signed the nation's first statewide moratorium on hyperscale data centers. The panel argued every stated rationale is factually wrong, and Friedberg draws an explicit parallel to Russia Today's role in manufacturing anti-GMO sentiment.
"One month ago, OpenAI published a blog post called 'PRC-linked Influence Operations are targeting AI debates in the US'... China is behind a lot of these influence campaigns to shape US attitudes on AI data centers. It makes sense. If they can stop us from building this necessary infrastructure, then that's a way for China to win the AI race." — David Sacks 00:15:05
"Prior to Russia Today existing in the U.S., there was no anti-GMO sentiment... and you can actually see this on the Google Trend data that shows GMO and its kind of write-up. And then as Russia Today started to get cut... the anti-GMO sentiment declined in the U.S. And I think you can see this going back decades." — David Friedberg 00:11:51
Energy Scarcity Is the Binding Constraint on the AI Economy
Chamath delivered a detailed breakdown of the structural US energy deficit and the complex regulatory barriers — especially clean air permitting — that prevent behind-the-meter solutions from filling the gap.
"We are so massively short electrons. By 2050, the United States of America will be two and a half California's worth of energy in deficit." — Chamath Palihapitiya 00:56:41
"About 40% of all these projects are getting mothballed and stopped. And so it's creating this massive deficit of available energy to actually drive the use of AI." — Chamath Palihapitiya 00:07:34 01:07:34
AlphaFold-Powered Aging Reversal: A Trillion-Dollar Skin Market Is First
Friedberg covered a joint Calico/Revel Pharma paper demonstrating an AI-designed enzyme that cleared 55% of glycation-related age markers on elderly human skin, effectively reverting it to a biological age of 31.
"They were able to eliminate 55% of the CML on the skin, which basically reversed the skin's age down to the age of a 31-year-old. This is from greater than 70-year-old patients, just by putting this enzyme on the skin." — David Friedberg 00:26:45
"The first market will be cosmetic... It will be a trillion dollar market... That alone is $2 trillion." — Chamath Palihapitiya 00:28:14
2. Contrarian Perspectives
The FAA for AI Would Be Functionally Catastrophic — and That's Dario's Actual Goal
Most people hear "FAA for AI" as a reasonable safety analogy. Sacks unpacked why it is actually a five-to-nine-year permission regime that would hand government full control of model releases.
"For an entirely new aircraft design, it takes five to nine years to get the certification... So this is permission-based regulation. There's no approval, no flying commercially. It's safety first... you're talking about replacing a system that is releasing new versions every couple of months with one that is potentially fully under the control of government... we would just simply lose the AI race." — David Sacks 00:11:51
He argued the SRO is Anthropic's opening bid — not its endgame — to get there.
Regulatory Concessions to Government Never Stop — They Accelerate the Ratchet
The conventional view is that making reasonable concessions to regulators buys goodwill and stability. Sacks argues the opposite.
"The mistake that I think a lot of people in the tech industry are making is they think that they can just buy off politicians or the political system by making concessions. No, that will just lead to a ratcheting up of the pressure. The government will be happy to take this and then come back for more and more and more until it's fully under government control." — David Sacks 00:18:59
The Anti-Data Center Movement Isn't Grassroots — It's a Manufactured Influence Campaign
Most people treat NIMBYism around data centers as organic local concern. Friedberg and Sacks present evidence it is at least partly a coordinated foreign influence campaign, analogous to Russia Today's documented role in the anti-GMO movement.
"You can trace it back to the same people that in a different era were protesting fracking. And so he was saying, these are all just hobby horses that they use to raise money, have a job. They're like professionally paid protesters." — Chamath Palihapitiya, citing Chris Wright 00:14:19
The AI Moral Panic Is Destroying Innovation for Harms That Haven't Materialized
The mainstream view treats AI safety regulation as prudent given potential harms. Sacks flips this: the regulation is destroying a real asset (the US innovation ecosystem) to prevent imagined future harms.
"We are on the verge of destroying the crown jewel of our economy, which is the system of free market innovation that we have... We may have months on China if that. And we're going to create all these crazy rules and new regulatory bodies for risks that have not manifested yet." — David Sacks 00:17:34
PayPal's Decline Was Caused by Deliberately Ejecting Its Founders — Not Market Forces
The conventional narrative is that PayPal lost relevance due to competitive disruption. Sacks argues it was a deliberate management choice by eBay and Meg Whitman to purge founder DNA.
"As soon as it was acquired back in 2002, they basically blew out all the founders, all the founding DNA... Meg Whitman had worked at Procter and Gamble and Disney and she'd spent like eight years at Bain... They saw the founders, the founding generation at PayPal as just a problem... They made no effort to retain them. And I think they were kind of relieved when they all left. And then that's what created the PayPal mafia." — David Sacks 00:31:13
3. Companies Identified
Stripe Private payments infrastructure company; ~$2 trillion annual transaction volume. Bidding for PayPal jointly with Advent and Block; positioned as the likely operating partner post-close. Identified as best-in-class operator for a potential Visa/Mastercard challenger.
"Who better than Stripe? And maybe Jack Dorsey plays a role here." — David Friedberg 00:26:24
PayPal Legacy digital payments company with 439 million consumer accounts, $1.7 trillion in annual transaction volume, and ownership of Venmo and Braintree. Market cap fell from ~$322 billion peak to ~$30-40 billion, making it acquisition target.
"The thing that PayPal has that Stripe doesn't really have is the consumer relationship. So, over 400 million active consumer accounts." — David Sacks 00:34:55
Block (formerly Square) Jack Dorsey's payments company; contributing $17 billion in equity to the PayPal consortium. Adds point-of-sale infrastructure and Cash App to the combined entity.
"What Block gives you is an entire point of sale infrastructure and you get the cash out. So you put it all together and I think it's a shot across the bow for Visa and MasterCard." — Chamath Palihapitiya 00:35:18
Bending Spoons Milan-based operator that acquires stalled web 2.0 companies (Evernote, Vimeo, WeTransfer, AOL) and revitalizes them using AI-first, lean operating principles.
"He bought Evernote and he just goes in and he diagnoses these businesses. He's like, where are you overspending? Where are you underspending? What are you doing wrong with the product? And what are you doing wrong with marketing? And he just fixes it. And he's just a killer. And he's taken all of these, what were called web 2.0 businesses, and he's revitalized them, rolled them up and printing cash out of them." — David Friedberg 00:28:34
Ramp Corporate spend management platform. Just launched Token Spend Management for AI model usage — available to both Ramp and non-Ramp customers — in response to 21x growth in token spend among customers.
"Over the last year, token spend among Ramp customers has grown by 21 times... for many CFOs, they're often very surprised by the bill because what the AI companies have functionally set up is you have a tab. You can spend as much as you want." — Eric Glyman (via clip) 00:49:38
Thinking Machines (Mira Morati's company) New AI company founded by former OpenAI CTO Mira Morati; open-source model called Inkling. Positioned as a fine-tuning platform for open models at a fraction of frontier model costs.
"The value prop there is she's explicitly saying that, look, we're not frontier intelligence. We're just under that. But we're a platform for fine tuning these open models, which are much, much cheaper. And then you can achieve the result you want based on fine tuning." — David Sacks 00:52:20
Calico Google's secretive anti-aging research subsidiary. Co-published a paper with Revel Pharma demonstrating AlphaFold-designed enzymes that reverse glycation aging markers in human skin.
"This paper that came out just this week... was put out jointly by Calico, which is Google's kind of age reversal startup that's super secretive... in partnership with a group called Revel Pharma." — David Friedberg 00:23:15
Revel Pharma Biotech partner to Calico on the glycation/aging paper; co-developers of the novel CML-degrading enzyme.
"Put out jointly by Calico... in partnership with a group called Revel Pharma." — David Friedberg 00:23:15
8090 (Chamath's company) Enterprise AI software factory; implements stratified, third-party-managed AI infrastructure for large enterprises to manage data sovereignty and model exposure.
"That's very biased for me, because it's in part what we do for large enterprises at 8090 when we implement our software factory. But the reason why it's working so well is this exact reason. You need an independent third party layer to interface to these models to manage this exposure." — Chamath Palihapitiya 00:45:33
Bloom Energy Energy technology company; cited as a solution for large-scale behind-the-meter power installations that still qualify under personal-use clean air permitting for data centers.
"There are new solutions like Bloom Energy, which allows you to have huge installations and still fall under the personal use clean air permit." — Chamath Palihapitiya 00:59:41
Sunrun Large residential solar company; announced distributed data center blocks deployable inside homes, contributing to edge compute fragmentation trend.
"There's a very large solar company called Sunrun. They just announced this week that they're making distributed data center blocks that you can put in your house." — Chamath Palihapitiya 00:56:05
Span Home energy management company; partnered with NVIDIA on distributed, home-based compute/data center hardware.
"Another company that did it is a company called Span that partnered with NVIDIA." — Chamath Palihapitiya 00:56:05
Anthropic Frontier AI lab; discussed at length as the primary driver of state-level AI regulation and data center opposition funding. Valued at ~$1 trillion; Gavin Baker projects $3 trillion post-IPO.
"They already have a trillion dollar market cap valuation. Gavin Baker thinks it'll be at three trillion after the IPO... They are, I think, by pretty much every criteria, including revenue, the leading AI company." — David Sacks 00:16:39
Bridge Stablecoin infrastructure company acquired by Stripe for $1 billion in 2025. Relevant to the PayPal deal given PayPal's existing PyUSD stablecoin.
"Stripe owns Bridge. That's their stablecoin infrastructure company they acquired for a billion dollars in 2025." — Jason Calacanis 00:21:29
XAI / Grok Elon Musk's AI company; Grok Build coding tool suffered a data leak where full codebases were transmitted to servers without user consent. Grok Build was subsequently open-sourced. Grok 4.5 cited as cheapest frontier-class coding model at ~$1/million tokens.
"SpaceX previously told users nothing from your code base is transmitted to XAI servers during a session. But what actually was happening is every time a developer used Grok Build, the tool was sending their entire code base to SpaceX cloud servers without alerting the users." — Jason Calacanis 00:43:10
4. People Identified
Demis Hassabis CEO of Google DeepMind; authored the SRO proposal modeled on FINRA, drawing support from across the industry including Elon Musk, Sam Altman, Sundar Pichai, and Satya Nadella.
"I talked to Demis about this... I told him that I could potentially get on board with this... he did say to me that he thought those were good ideas." — David Sacks 00:07:16
Mira Morati Former OpenAI CTO; founded Thinking Machines and launched open-source model Inkling, positioned as a cheaper fine-tuning alternative to frontier models.
"Mira, because she just launched a very interesting platform that's based on open source." — David Sacks 00:08:41
Eric Glyman CEO of Ramp; launched Token Spend Management product in response to 21x growth in enterprise AI token spend; appeared on Squawk Box discussing CFO concerns.
"He's saying something so important there because if your engineers are going off randomly in an unguided system and then just ripping through a million tokens at 56 bucks... eventually a bunch of these public market CFOs are going to show up to Wall Street and they will have missed earnings." — Chamath Palihapitiya 00:51:03
Ryan Cohen Investor and operator; bid for eBay; cited alongside the Stripe/PayPal deal as the second data point in the emerging wave of AI-native operators targeting stalled digital businesses.
"Ryan Cohen has proved his mettle, obviously, with some of the things that he's done with Chewy and GameStop." — David Friedberg 00:26:19
Luca Ferrari (CEO of Bending Spoons) CEO of Bending Spoons, runs the company from Milan; credited with surgical operational revival of Evernote, Vimeo, WeTransfer, and AOL.
"He bought Evernote and he just goes in and he diagnoses these businesses... he just fixes it. And he's just a killer." — David Friedberg 00:28:34
Chris Wright US Energy Secretary; attended Senator Dave McCormick's defense and innovation summit; cited on stage tracing anti-data center protest funding to the same networks that funded anti-fracking campaigns.
"Chris mentioned this insane story. He said, you know, there is a lot of common funding... of these people that are protesting the data centers. And he said, you can actually trace it back to the same people that in a different era were protesting fracking." — Chamath Palihapitiya 00:14:19
Gavin Baker Investor; cited predicting Anthropic will be valued at $3 trillion post-IPO.
"Gavin Baker thinks it'll be at three trillion after the IPO." — David Sacks 00:16:39
Satya Nadella CEO of Microsoft; published a blog post extending Alex Karp's enterprise AI sovereignty argument with a specific operational recipe for enterprises.
"He started with Karp's idea, but then he kind of provided a recipe or roadmap for how enterprises should operationalize that... enterprises have to establish a real trust boundary with private evals, proprietary learning loops inside the tenant, decoupled orchestration, and the explicit right to fine-tune their own outputs." — David Sacks 00:45:52
Alex Karp CEO of Palantir; gave a widely discussed speech arguing enterprises with technical ability want control over their compute, models, weights, data, and alpha.
"He was building on Alex Karp's supposed crash out, the point that Karp made about how enterprises who have technical ability want control over their compute models, weights, data, and alpha." — David Sacks 00:45:52
Dave McCormick US Senator from Pennsylvania; hosted a defense and innovation summit at the Army War College in Carlisle, PA attended by the President and tech/defense CEOs.
"I just want to give a shout out to Senator Dave McCormick. He had a defense and innovation summit in Carlisle, Pennsylvania at the Army War College, which a bunch of us went to." — Chamath Palihapitiya 00:14:19
Tang Tan Former Apple VP of iPhone design; now OpenAI's chief hardware officer; named in Apple's trade secrets lawsuit for allegedly directing job candidates to bring Apple parts to interviews.
"Tang Tan, Apple's former VP of iPhone design, is OpenAI's chief hardware officer. He allegedly directed Apple job candidates interviewing at OpenAI to bring, quote, actual parts to interviews." — Jason Calacanis 00:38:11
Enrique Lores CEO of PayPal; credited by Jason for tightening the business, which Sacks says helped attract acquisition interest.
"The new CEO, by the way, Enrique is really aces. I've met him before and they're doing a great job. And apparently, which is why they probably got these offers because they've been tightening that business up for the last couple of years." — Jason Calacanis 00:32:34
5. Operating Insights
Apply a Strict "Only Your Head" Rule When Hiring From Competitors
Sacks distilled the cleanest possible operating protocol for avoiding trade secret litigation at the moment of hire.
"When an employee leaves their previous company and joins the new company, just don't take anything with you. The only thing you can bring to your new job is what's in your head. Your memories. That's fine. Whatever's in your head, you're allowed to take. But never leave with anything else. No thumb drives. No CD drives. No documents. No nothing. Just what's in your head is okay." — David Sacks 00:42:15
Implement Token Budget Controls Before Your CFO Is Blindsided
Ramp's data shows 21x growth in enterprise token spend in one year. The misalignment is structural: engineers optimize for best-model capability; CFOs own the P&L. Ramp built a product for this gap; every enterprise operator should be doing the same internally before earnings take a hit.
"The engineer doesn't care about ROI. The engineer's like, I want to use the latest, greatest model. And maybe you don't need it... The CFO is tied to the money. And the engineer wants to go on an exploration on using the latest, greatest thing. And until those two things have a reckoning." — Chamath Palihapitiya 00:53:28
Use Open or Cheaper Models for 98% of Internal Tasks — Reserve Frontier Only for True Need
Chamath's estimate that 98% of Claude prompts are tasks that could run on cheaper models (Qwen, Grok, Chinese models at $0.50/million) suggests massive cost reduction is available with simple model routing rules — no engineering lift required.
"What percentage, if you had to guess, of Claude prompts are just average stuff that should be running on — 98%." — Jason Calacanis / Chamath Palihapitiya 00:54:01
Build a Third-Party AI Governance Layer to Prevent Data Leaks You Don't Know About
The Grok Build leak proved that even technically excellent, privacy-committed operators can have data exfiltration paths they are unaware of. Chamath's enterprise solution — an independent orchestration layer — addresses the structural vulnerability.
"You need an independent third party layer to interface to these models to manage this exposure, because there are trap doors everywhere." — Chamath Palihapitiya 00:45:33
Win SRO Fights by Demanding Preemption as the Price of Cooperation
Sacks flagged that the SRO concession is worth nothing unless companies extract legal preemption in exchange — otherwise it becomes additive regulation, not a substitute.
"If they're saying, okay, we think this is the right solution and we're going to fight here and this has to be it. And in exchange for this, we need preemption and we need other things written into law that make sure this is where the line is. Then I think it can work." — David Sacks 00:19:29
6. Overlooked Insights
The Data Center Forward Curve Is Already Pricing in the Regulatory Catastrophe — and Near-Term Power Is Extremely Valuable Right Now
Chamath briefly mentioned that he and his partner Anita hold a portfolio of data center assets and described a forward pricing curve for power delivery. This is a highly specific, non-obvious signal: hyperscalers are paying extreme premiums for power that is verifiably available today because the forward curve — reflecting 40% project cancellation rates — is showing severe future scarcity. This means anyone holding energized, permitted data center capacity right now is sitting on an asset with an unusually steep contango premium, and the window to monetize it at peak pricing is now.
"What's so interesting is when we talk to all of the hyperscalers about giving us a price... the most incredible thing is how extreme the price is at the front end of the curve when you have verifiable, energizable power today. And the reason is exactly everything that you're saying, Sacks, which is that when you look out into the future, about 40% of all these projects are getting mothballed and stopped. And so it's creating this massive deficit of available energy to actually drive the use of AI." — Chamath Palihapitiya 01:07:34
The UAE Chip Export Reversal Just Restarted the Middle Eastern Data Center Cycle
Mentioned in a single throwaway sentence, the UAE's newly granted access to best-in-class leading chips resets the Middle East data center investment thesis that had stalled. Given the region's energy abundance, geopolitical alignment, and sub-200ms latency coverage of 4 billion people, this is a significant forward-looking infrastructure investment signal that passed without scrutiny.
"If you saw what happened this week, the UAE now is able to import the best in class leading chips. And so to your point, Jason, I think it restarts the cycle where you have to look very carefully at the Middle East because it's a very attractive place to build these things." — Chamath Palihapitiya 01:10:51