Diet TBPN: November 6, 2025

0:00Wild time on the timeline yesterday with comments from OpenAI, OpenAI's finance chief Sarah Fryer. She was at a Wall Street Journal event on Wednesday. She put the timeline in turmoil. She did put the timeline in turmoil. The headline from the Wall Street Journal was OpenAI wants federal backstop for new investments. OpenAI, at our core, are the model company that needs always to be the state of the art. That's what we've done time and time again. GPT-5 is no exception. but even in areas like open source, we're attempting to put the state-of-the-art model always out into the world. And in order to do that, we always want to be on the frontier chip.

0:42So the question is, how long does a chip remain on the frontier? Is it three years, four years, five years, or even longer? Now, in a world where we have no compute, we're compute-constrained, we are absolutely using chips that have, like, A100 equivalents that have been around, like, maybe six, seven years at this point in time. If that's the case, financing chips gets a lot easier. If the timeline on the chip stays short, that gets harder. And so this is where we're looking for an ecosystem of banks, private equity, maybe even governmental, the ways governments can come to bear. Meaning like a federal subsidy or something?

1:22Meaning like just first of all, the backstop, the guarantee that allows the financing to happen. It's like a ton of bricks. That can really drop the cost of the financing, but also increase the loan to value. So the amount of debt that you can take on top of an equity portion for some federal backstop for chip investment. Exactly. And I think we're seeing that. I think the U.S. government in particular has been incredibly forward leaning, has really understood that AI is almost a national strategic asset, and that we really need to be thoughtful when we think about competition with, for example, China.

2:00Are we doing all the right things to grow our AI ecosystem as fast as possible? We're in a global erotica race. Are you talking to the White House about how to further formalize that kind of backstop? We're always being brought in by the White House to give our point of view as an expert on what's happening in the sector, for sure. So on LinkedIn, Sarah Fryer said she wasn't asking for a backstop for open AI's investments. She was just making the point that American strength in technology will come from building real industrial capacity, which requires the private sector and government playing their part.

2:36Every comment, every sentence uttered by an OpenAI executive will be analyzed to death. It's one of my favorite things to do. So that's exactly what we're going to do on this show. should a federal backstop exist for large-scale industrial projects? Like when there's a lot of money floating around in the economy, should the government be ready to step in if things don't line up, if things don't pencil out? And so Sam Altman actually touched on a similar concept in his conversation with Tyler Cowen that was recorded, I think, a few days ago, but also went live yesterday, which is just sort of crazy timing.

3:12As you know, both you and I were fans of nuclear power, but we also know the insurance for nuclear power plants is provided by the government. The plants might be quite safe, but people worry. They're nervous Nellies. There's a lot of parties involved. So the federal government does the insurance. Do you worry that the future holds the same for AI companies or the feds or your insurer? At some level, when something gets sufficiently huge, whether or not they are on paper, the federal government is kind of the insurer of last resort, as we've seen in various financial crises and insurance companies screwing things up.

3:49So I guess given the magnitude of what I expect AI economic impact to look like, sort of I do think the government ends up as like the insurer of last resort. But I think I mean that in a different way than you mean that. And I don't expect them to actually be like writing the policies in the way that maybe they do for nuclear. And there's a big difference between the government being the insurer of last resort and the insurer of first resort. last resort's inevitable, but I'm worried they'll become the insurer of first resort. And that I don't want. I don't want that either. It's not, I don't, I don't think that's what will happen.

4:23Let me give some more context here. In that same day, that same interview, OpenAI CFO Sarah Fryer suggested the market is overly focused on anxiety about a possible bubble in artificial intelligence and should muster more exuberance about the technology's potential. Quote, I don't think there's enough exuberance about AI when I think about the actual practical implications and what it can do for individuals. We should keep running at it. You know, Friday, everyone saw the interview with Sam and Brad. Brad asked a very simple question that's on everyone's minds. Sam basically doesn't answer the question.

4:56Nobody heard his response and thought, okay, I'm confident. And so to go into that and then say in this interview, the market is not exuberant enough. She can go out and say that she takes it back or that she didn't mean it to come across in the way that it did. but in the moment agreed with the interviewer's point. I feel like it is true that there is a government backstop. Whenever there's a trillion dollars of spend going on, there is a de facto backstop. Like this happened in 2008. There was a global financial crisis. Banks were leveraged on housing. The government came in and acted as the lender of last resort.

5:32In that moment, it was the rational thing to do. Now, the problem is that if you are identifying the backstop before you get into trouble, It creates this moral hazard. And that's why people were so upset with Wall Street. I don't see, I actually don't see a problem with the government behaving like it always has. I don't have a problem with that. I'm fine with a backstop. And I wouldn't change anything about the way the government works. And I'm serious about that. I think that the alternative is actually worse. I think the alternative, if you go back to 2008, and you say, okay, let's not bail out the banks.

6:08I think that you just extend the recession. Yeah, yeah, yeah. But if in 2005, the banks and various players were saying, we're going to do a bunch of super risky lending that we're not sure is going to pencil out. And we just want to know that if things get really bad, you'll be there to bail us out. But we're going to make a ton of money in the process. My conclusion to my post today was America has a playbook for backstops. I don't think Sarah Fryer or anyone at OpenAI is asking to rewrite the playbook. But there's something that feels very moral hazard-y about the CFO of the most important company at the center of a massive economic wave sort of telling everyone how the magic trick works.

6:45It is a problem if you say, I'm openly acknowledging the way the system works and I'm going to push it to the limit. That's a little bit rough. The A-B test of like 2008, you can kind of say is like Great Depression, where there's this big critique where people were basically mad that there wasn't stimulus earlier. Totally. And then basically by not doing stimulus, you drag out the depression like way longer than it should have. Totally. I mean, it was the foundation of the Keynesian economic model. Like this idea that it is actually advantageous to spend. I like the backstop more than we're going to try and pick a winner and do a specific project.

7:19What was the backstop in dot-com? Do you know what they did as a result of dot-com? No. Like they didn't go in and say, oh, you're in pets.com. We're making you whole. What they did was they said, there's a recession right now. We need to lower interest rates. We need to cut taxes a little bit. We need to just stimulate the economy just a little bit to get back on our feet and just start up the economic engine. So they did tax cuts and lower interest rates. And that is a backstop. And I'm fine with that. I think that's fine. If the AI bubble pops and the end result is that we get lower interest rates and some tax cuts and like different.

7:55It's very different than various players wanting to develop data centers and lenders feeling uncomfortable about not having clarity on, are these chips going to be worth anything in six years? Are they going to be worth anything in five years? Are they going to be useless after a few years? When I hear these comments from Sarah Fryer, what I'm hearing is we have a bunch of lenders that would happily finance this if there's a government guarantee that they won't lose money. And that's very different. If you get some type of setup where the government does come in and say, not just for OpenAI, but for a variety of these players, like, yeah, we will provide a backstop.

8:31Yeah. The lenders start to feel comfortable. They start to lend. What's the lender's incentive? If a lender makes money when they lend money and they have some type of structure where they can't lose money, what are they going to do? They're going to lend as much money as they possibly can. Looking back through, you know, 07, 08 and the example you gave around post.com crash, I'm not against government intervention during a crisis. Totally. Using the government to create what will ultimately just be an even greater crisis and ultimately just socializing the losses from all of this. And yeah. Tesla.

9:05Government backstop, right? Gave loans. Paid back the loans. Tesla paid back the loans. Sure. With interest. The taxpayer made money on that. And anyone who says, oh, Elon only, you know, he's just the government pays for everything. That's it. It's like they're wrong. The government made investment in Tesla and they got back money with interest. And so the government participating in private markets, it doesn't always end poorly. Sam is not asking for the government to take a position. He's like, I don't want to be Intel. I don't want them on my cap table. I don't want any of that. It's like.

9:37Sure. So honestly, I don't know what they're asking for because it is a crazy, crazy piece of comms to say, to even talk about it in the current zeitgeist. And, you know, the general public already hates AI enough. I don't know. Yeah. I would expect that something like this does happen. Yeah. I mean, that's why David Sachs came out and said, like, there will be no federal bailout for AI. The U.S. has at least five major frontier model companies. If one fails, others will take its place. The AI race looked more clearly like China designing AI weapons. And OpenAI is purely a defense tech player that is also developing AI weapons.

10:18And in order to be competitive, in order to not have the CCP take full control over the world with their sort of dark super intelligence, we needed a national champion, national lab. We needed to give them as much infrastructure, as much energy, all these things. That scenario, I'd be looking at this a lot differently. I think literally me, you, Sarah Fryer, David Sachs, we're all in agreement here that no one's saying OpenAI gets a bailout if the company doesn't do well. No one's saying that. But I think it's important to break these out. A bailout is like OpenAI is going to default and they need a capital injection in order to stay solvent.

10:58And David Sachs is clear. We're not going to have a bailout. But a backstop of loans, I would expect that to happen. I guess the bigger question is, like, should America vote actually socializing the losses of the AI buildout? Like, if you don't like AI, then you're like, no, there's no point. But if you think it's like the railroad or you think it's like the Internet infrastructure buildout or you think it's like the computing revolution, then you should be all in on that. on like, yes, put the weight of the U.S. government behind this project. I just think there's no public appetite for this.

11:33Look at the state of the average American. I think you're right. They can't afford a home. They can't even afford McDonald's. They're already worried, I might get a job, and then I might lose it to AI. You have people working in AI that are already talking about job displacement. We have some breaking news. SAMA has tweeted addressing the backstop. I would like to clarify a few things. First, the obvious one, we do not have or want government guarantees for open AI data centers. We believe that governments should not pick winners or losers and that taxpayers should not bail out companies that make bad business decisions or otherwise lose in the market.

12:07If one company fails, other companies will do good work. What we do think might make sense is government building and owning their own AI infrastructure. Go code. Then the upside of that should flow to the government as well. We can imagine a world where governments decide to take to offtake a lot of computing power and get to and get to decide how to use it. And it may make sense to provide lower cost of capital to do so. Building a strategic national reserve of computing power makes a lot of sense. If you believe in the lore of AI 2027, this is trust the plan. Jensen Wong was in the Financial Times talking about the China will win AI race with America, says NVIDIA chief.

12:49He's like China is going to win the AI race. We need to sell them as many chips as we possibly can right now. Please. They're going to win. His position is that we want to get China dependent on American AI infrastructure so that we can pull it back at any time. But saying they're about to win the AI race when he's also advocating for selling them as many chips as they can produce doesn't really sit that. We need to better define what the AI race is. Like Gurley earlier this year, I think made some good points. Like, what is the AI race? How does one win? What does that look like? What does it look like in the near term?

13:28If you're AGI pill, it's the race to a super intelligence, a system that can do any economic work. And the reason that people don't take that narrative super seriously right now is that companies like OpenAI make short form social media apps. Well, then the question is, if we're not getting 10x the results, do we need to be spending 10x as much next year? What we do think might make sense is the government building their own AI infrastructure. The project is going to happen. Building a strategic national reserve of computing power makes a lot of sense, but this should be for the government's benefit, not the benefit of private companies.

14:01The one area where we have discussed loan guarantees is as part of supporting the buildout of semiconductor fabs in the United States. I completely agree with this. I'm pro all of it when it comes to building semiconductor fabs in the United States. I've been a bit more negative. I'm pro buildout. I think the power side is an area that could make a lot more sense. Yeah. That could benefit the average American a lot more. Yeah. Which is like, hey, the government is going to help finance bringing a lot more power online. And we're going to be able to stay at the leading edge in AI, lead in AI, and your power bill.

14:36We're actually going to make your power bill cheaper over time. That's something that people could get behind. But federal backstop so you can get more videos of your cat dressed up like a doctor. I don't think people are going to be on board with that. I think it's very unclear to me how building just a national reserve of computing power makes a lot of sense. New consumer devices and robotics. Excited about that. I think it's hard for them to lean on those, right? Because when I look at the history of big tech companies launching ambitious products, oftentimes they either flop entirely or lose money for a long time.

15:16So new consumer devices. So Zuck has been doing this and just been literally just lighting money on fire for years now. You're saying we are looking at about$1.4 trillion of commitments over the next eight years. and you're saying that part of our solution to delivering on this is new consumer electronics. I personally, I don't lean on that too hard. None of these things scream cash flow to me. These, to me, seem very likely like new business lines that also lose money. I'm excited about these categories. I think they're going to make cool, novel products. But it doesn't give me more confidence that they can meet those commitments.

15:56Keep the government out of the application layer. Focus on the energy layer. Just work on driving down energy prices and let all the tech companies duke it out. And they will be beneficiaries of cheaper energy. But I want, you know, tons of American jobs building everything from natural gas to solar. That solar plant, I want 10 of those. And I want it staffed by Americans. How about that? Yeah, yeah, yeah. I'm on board with that. We could agree with that. The other thing is we're also looking for ways to more directly sell compute capacity to other companies and people. And so when I look at this, when you're simultaneously saying we're so compute constrained that they're making the AI lazy and they're not launching new features.

16:42Yeah, yeah, yeah. To me, again, this just says OpenAI is a new hyperscaler, and it's so hard to compete with the hyperscalers that have huge amounts of cash flow that they need every advantage that they possibly can. And one way they could get an advantage is through programs like Sarah suggested and then walked back. The timeline continues to be in turmoil. If OpenAI were to fail, it would have zero impact on the future of AI. Zero, he says. They have a dozen viable competitors and 10 folks who would buy the assets. Let it rise or fall based on merit. Like the fact that this is the reaction to that quote is like so crazy.

17:20The idea of like federal backstop just immediately conjures massive failure. And it just puts your company and failure right next to each other. So very, very rough. This is why traditionally CFOs are only allowed to speak fake accounting identities in public. Backstop is what the Federal Reserve calls it when they have to save 20 banks from exploding in a global financial crisis. The word you were looking for, Sarah, was partnership. I'm actually more pro backstop than partnership because I don't love the idea of the government picking winners in any particular layer of the stack, but partnership just sounds so much more positive than Backstop.

18:03It just conjures up financial crisis. It just feels like major layoffs, recession, stock market sell-off as soon as you start talking about Backstop. So a huge amount of the money being spent on AI has gone to GPUs, particularly NVIDIA, rocketing the fabless design company to a nearly$5 trillion valuation and the title of most valuable company in the world. The problem from the Carlota Perez perspective is that all of this spending on chips is relative to the sort of infrastructure she wrote about. Railroads, factories, fiber, short-lived. If there's a bubble in railroads and we're like, we think railroads are going to unlock 5 % GDP growth, it's amazing.

18:45And then you build all the railroads. Like the worst case scenario is you end up with a bunch of rails. You can still put cars on them. Like they still work. but with a bunch of gpus that like depreciate and get useless and they break like they don't stick around forever good thing john is we're gonna find out we're gonna right now we're effing around right now we're gonna find out fryer is describing a worse form of regulatory capture than anything we have seen proposed in any u.s legislation i am aware of a firm lobbying for this outcome is literally rather than impressionistically lobbying for regulatory capture.

19:20And this is going back to what I was saying before. If there is some type of program where the government is guaranteeing loans made for certain data center projects, like anybody will just get to be in that program. I'm with Dean Ball on this one. And, you know, Sarah and the team have walked it back, but they said it. Yeah. David Sachs said there will be no federal bailout for AI. But he did address your comment, Tyler. He said, that said, we do want to make permitting and power generation easier. The goal is rapid infrastructure build out without increasing residential rates for electricity.

19:57Let's hear it for that. NVIDIA down almost 8 % over the last five days. A bunch of other names are down a ridiculous amount. Duolingo is down 30%. Snap is up, right? CarMax down 18%. HubSpot down 18%. Wow. DoorDash down 15%. Crazy sell-off. What's been notable this week, companies are selling off after an earnings beat. Palantir down 10.7 % in the past five days. Whether you're going long or you're going short, do it on public.com. That's right. Investing for those that take it seriously. They got multi-asset investing, industry-leading deals. They're trusted by millions. Snap shares surged after the company announced a$400 million partnership with perplexity to incorporate AI-powered search engine into Snapchat.

20:46The deal gives Snap a new business line and puts the social media company squarely into the mix around AI chatbot. Why is Snap not monetizing their almost half a billion DAUs of young people? And this is a great way for them to do it. 477 million DAUs in the third quarter for Snap, an 8 % jump from a year ago. So they're still growing. The company also reported 943 million monthly active users. So they're almost at a billion MAUs and half a billion DAUs, which is remarkable. Wild that Snap has almost a billion monthly actives and is valued substantially less than perplexity. It was kind of odd that Snapchat didn't have a dance partner on the AI side since every other major social media platform is kind of bolted onto a lab at this point.

21:43XAI and X are obviously the most prominent example, but even like LinkedIn and Microsoft. And I don't think the U.S. government should backstop data center loans or funnel money to NVIDIA's 90 % gross margin business. Instead, they should make it really easy to produce energy with subsidies and better rules. Infrastructure that's beneficial for all and puts us at parity with China. Wow. This is an example of a person who's not into AI. Danny with 142 followers here saying, build your own effing energy sources. Stop being parasites on the grid. The way the grid is set up, the way the US economy is set up is not like we don't build separate grids for each use case.

22:25Like we didn't build a separate grid. for Netflix. Some people might say that hospitals deserve better access to electricity than Netflix, right? Like, do we really need Netflix? Like, is that life or death? No. This idea of the government putting the thumb on the scale of, and basically doing away with like what is effectively net neutrality or grid neutrality is super, super alluring to people right now. That people are going to say, why are my rates going up for the thing I don't like? It's kind of always been that way? You know, why are airline ticket prices going up? Because people are ordering things that are, you know, shipping by air like that happens.

23:05What is the case against bailouts if winning the AI race against China is existential? I'm not actually making this argument or have this view necessarily, but public money backstops bailouts seem like the natural endpoint of how AI is discussed at these levels. If winning the AI race against China is existential, America in every company should do everything possible. I feel like I should read through a little bit of this FT article about Jensen saying China will win the AI race with the U.S. In the starkest comments yet from the head of the world's most valuable company, Huang told the Financial Times, China is going to win the AI race.

23:41Huang's remarks come after the Trump administration maintained a ban on California-based NVIDIA, selling its most advanced chips to Beijing, following a meeting between U.S. President Donald Trump and Chinese leader Xi last week. The NVIDIA chief said that the West, including the U.S. and the U.K., was being held back by cynicism and we need more optimism. Huang had previously warned that the latest American AI models were not far ahead of their Chinese rivals, urging the U.S. government to open up the market to its chips to keep the rest of the world dependent on its technology. But following his meeting with Xi, Trump said last week that he did not want to let China use NVIDIA's cutting edge Blackwell chips.

24:16The most advanced, we will not let anybody have them other than the United States. We still want a fair fight. We want a free market. We want everyone who wants to buy chips can buy them. And then it's still a race. Like everyone gets a pair of cleats and then you have to actually race. And so what are you racing on? You're racing on ingenuity. You're racing on ability to actually build things effectively. We know from history that even if we give China chips, they will still spend as much as they need to in order to catch up themselves. I don't know. I do think that you can put supply pressure on the Chinese system.

24:58Sure. Maybe you could slow it down a little bit. I mean, they've been investing in semiconductors for 50 years, right? Yeah, they've had like how many five-year plans. Yeah, they're going to continue to do that. There is some level of an acceptable argument to this idea that if you continue to push NVIDIA chips there, it just makes the Huawei Ascend just a little bit less exciting. This post, I think, is pretty reflective of kind of general sentiment, specifically outside of the tech industry from Elliott. He says, if OpenAI blows up and investors are saved by the government, I would vote for Fidel Castro.