Oct 25, 2025 ยท 2025 #40. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
What is a Browser? What is a Bubble?
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 2
Hello everybody, it's Saturday, November the 8th, 2025. I'm probably not going to run this show until Sunday, tomorrow, November the 9th, although I think Keith on his That Was The Week show will run this today. Today we're running an interview I did with Andrew Hill of the Financial Times. Andrew manages the FT's Business Books of the Year. They're down to their shortlist. And the theme really of all six books on the shortlist is this growing economic conflict, battle, race. lots of ways of describing it between the United States and China, particularly around technology and AI, but on lots of different fronts too. It's perhaps best summarized in one of the books on the list, Breakneck by Dan Wong, China's quest to engineer the future. And John Thornhill of the FT summarized it as... a battle between China's engineers and America's lawyers. He believed that China's engineers will beat America's lawyers. And that's the theme of That Was The Week, our weekly text summary. The title is, Is China the New America? And it comes with all sorts of graphics on the juggernaut of the Chinese economy and how it's catching up. So, Keith, is China... The new America? Is there an AI race?
Two separate questions, and it's probably best to unpack them. Obviously, it's a highly emotive question, and the narrative around it in the media, as you just showed, has all kinds of points of view, many of which are strongly held, but I think mostly shallow. China is clearly the economy leading growth in the world. Its GDP growth is second to none.
Yeah, and you historicize it in your editorial. You suggest that this shift, if you like, in economic weight is equivalent to what happened in 1870 and 1918, when the US caught up and then surpassed the British Empire. So it's a historic shift.
Absolutely. And, you know, it needs to be recognized for what it is, because other Otherwise, one can descend into kind of nationalist rhetoric that isn't grounded in facts. China is big, growing in terms of purchasing power, already surpassed the US. In terms of absolute dollars, it hasn't. own 17% of world GDP. The United States owns 26%. And in terms of individual purchasing power, it's not even close. The US is, in purchasing power terms, $86,000 a year is the GDP per capita, whereas China's is about $26,000. So that's a huge gap. So there's lots of different ways to understand it. But there's no doubt China is growing way faster than the US. It has either caught up and surpassed it by some measures or is on the tail by other measures. And AI is the latest focus for that. And it's often depicted as a race. You know, I don't want to quibble with words. There obviously is competition, but A race usually implies a finish line, and with AI, there is no finish line. So it's hard to call the winner at any given moment.
Well, Ben, every economic race then can be suggested it's not really a race because there's never a finish line. Every technology morphs into something else. I mean, the Internet's morphed into AI, and AI will probably eventually morph into something else. But as you know in your editorial, Keith, In historic terms, countries do compete. I mean, that's just the nature of things and the US-Europe or US-UK British competition marked or shaped the 20th century. US-Chinese competition, maybe I'll avoid the word race, will shape the 21st century. Would you accept that?
Absolutely. No, we're moving from a century of US leadership or dominance to a century looking ahead in which the US will be one of several, with China being being the top of those several, but India and Indonesia and even Nigeria are going to be part of that picture. And so the world is changing to a more multi, I don't know if polar is the right word for more than two, but a multi-polar world with China.
And that's one of the pieces you linked to CEO and the FT description of Wang is CEO criticizes Western cynicism while Beijing loosens regulations and cuts energy costs for data centers. And actually coming back to the FT book of the year, one of the other books on the list is by Stephen uh wit um who's been on the show as well his book on and video is called the thinking machine it's an interesting book and he's an interesting author right so
uh so you know there is there is some real obviously jensen huang is a man who knows what he's talking about when it comes to ai because although you have to
assume that he said this i mean he knew that this would be catnip to the world's press. Presumably, he wanted to make headlines and make his point. What's the point that Huang is trying to make?
It's roughly the same point that David Sachs makes in the Andreessen interview that's also in this week's newsletter, which is the US, in order to compete, is going to have to follow the highly deregulated environment that the Chinese government has created in China for AI. and not put too much friction in the way. And that is complex in the US because there are now four states, all democratic states for what it's worth, led by California, seeking to regulate AI at the state level in ways that would restrict how fast it can grow, basically, because there'd be all kinds of hurdles to jump over before you could execute and implement things.
That's democracy, Keith, isn't it? I don't know whether Sachs or Marc Andreessen are particularly fond of American democracy or Jensen Huang, who is in some ways Taiwan based. Are they suggesting that democracy doesn't work in terms of economic development then in keeping up with the Chinese?
I think decentralized federal democracy with multiple centers of power, they are saying that. They're saying it's just not viable for companies to deal with 50 states. What they would prefer is a single federal set of rules as opposed to a patchwork quilt of rules. And, you know, that makes a lot of logical sense.
There isn't so much agreeing with them. It's more executing against a complex network of rules that are all different from each other. In software, that becomes a nightmare to maintain and manage. If you've ever filed taxes in more than one state, you would know how hard it is.
Isn't that the secret sauce? I mean, this is not a politics show. Isn't that the secret sauce of American democracy is that You have 50 centers, 50 states where... You and I could have an interesting conversation there.
I've always felt that it's, you know, democracy doesn't rely on a federal structure. Obviously, Britain isn't federal and it's a democracy. Democracy just relies on the people electing decision makers with an executive and a judicial branch as well. and so you know a federal like germany and america are both federal that that's a choice and america is federal but i don't know that democracy rests on that um i i would say that it is i've always thought of it as the unfinished american revolution you know the american revolution was a compromise at the end of the day with a balance of power between federal government and the states logically you kind of want a strong federal government, and America hasn't got one. Yeah,
and it's the theme of another of the books, actually, a massive bestseller this year, the book by Ezra Klein and Derek Thompson, Abundance, How We Build a Better Future. It's thinking about the future very much from a progressive point of view, but they're also, in their own way, from the left, critical of the power of states and the ability of... bureaucracies of one kind or another to hold up progress. So perhaps in my sense, I mean, I don't think we could argue that either David Sachs or Marc Andreessen are on the left, but they're in agreement, Sachs and Andreessen, with Klein and Thompson on America's failure to build.
Yeah, I think that conversation is the one where the labels right and left go out of the window, and it really comes down to economics and politics and how they interact. The role of the government, for example, in the economy is a shared conversation at that level. And I think it does belie those labels. It's interesting that Andreessen is very pro-American at the moment. He has this American dynamism view, and Sax is on the Andreessen show with Andreessen and Ben Horowitz.
And the AI czar. So he's not only an opinion carrier, but he's a person that executes policy. So what they are saying and what Jensen Wang is saying, even though it sounds the opposite, it's actually the same. They're all acknowledging that America is at risk of falling behind China. This week,
Cursor, the app that most people use for coding, dropped Anthropic as its main coding engine and replaced it with China's Quen. which is falling on the heels of Box.net having done the same. So Chinese open source models are being widely used now in American enterprise.
You are a defiant internationalist, Keith. It's no secret that you're not keen on any kind of nationalism. Should we even be using these terms, Chinese open source software, American open source? They just happen to be companies based in China or the U.S.?
Yeah, I mean... The problem is the media uses them to such an extent that it's hard to avoid getting involved in the conversation frame that way. But what is really happening is just innovation, which is now globalized. Silicon Valley is clearly investing massively in that. not always in American companies, some are overseas companies including Chinese companies like DeepSeek has American investors for example Minimax version 2 came out this week which is ByteDance owned and has leapfrogged even the other Chinese DeepSeek model So these are all just global technologies innovating and releasing things to the whole world, not just to their fellow citizens of their own country. And so what's really happening is a global uplift using AI that gets framed in these national terms.
That's one reality, but the alternative reality is the political one of nationalism, growing power of the state, not just in China, but in the United States. Another of the books on the FT book of the year, I think it's an excellent book, is Edward Fishman's Choke Points, American Power in the Age of Economic Warfare. Fishman was on the show a couple of weeks ago. He's really good. And actually, as I said to Andrew Hill, I hope he wins because he's a young, smart, energetic author. But he's simply, and he's worked within the DC, so he knows his way around DC. He's not just an academic. The reality is that we live in an age of economic warfare. I mean, that's just the nature of things, whether one's in DC or Beijing or anywhere else, Keith. So each power is using its political resources to bash the economic power of the other.
The only thing I'd tweak is the use of the word warfare. I mean, the fact that Trump and Xi met in Korea last week and made some kind of a deal shows you that it's a negotiation. They're certainly aware of each other's strengths and weaknesses. They certainly make decisions that favor their own economies that may impact others negatively. So all of that's true.
Well, that's not Fishman's point. He argues, and it's not just with China, is that America is increasingly defining itself by using political power to punish its enemies, and China may not be its principal enemy, Russia of course is the principal enemy, but nonetheless we live in an age of, for better or worse, of economic warfare.
So I would reframe that as America is trying to slow down its already visible relative decline, and is using power to do so but america can't economically speaking america can't win china china clearly is winning i don't think i didn't think you wanted us to use this word win uh well you know what i'm saying is america will not slow down its relative decline it can't it's it's an economic impossibility and what will happen is uh Right now, China is accelerating that relative decline and has been doing for about a decade. And there's nothing in the wind that tells us that's going to change. America can innovate. It can certainly be a huge global leader in AI. It can get a lot of the world's GDP growth flowing to American AI companies. But unfortunately, GDP is made up of more than AI. It's made up of shipbuilding and energy.
I mean, speaking of in the wind, the one piece that I don't think you put it in the newsletter this week, the one piece that really struck me was in The Economist. It was a leading editorial about China's clean energy revolution will reshape markets and politics and the way in which China has clearly won when it's come to clean energy. I mean, a massive investment, unimaginably massive investment in clean energy in China, whereas America has fallen back, still relying on oil and coal. And Trump's very reactionary on that front. So... Hasn't China, and I know you don't like this word, but hasn't China already won the big battle for energy? I mean, even Saudi Arabia is investing in AI as opposed to oil because they're going to need to rely on China's clean energy. And it also gives China a moral advantage because they can claim to be saving the world too.
I think in those micro economies like clean energy, you can use words like won and lost. if you define an end point and you can measure that somebody's passed it and somebody else hasn't, and in clean energy, clearly the economist numbers, and it's not just the economists, this has been well known now for quite a while, that China is very far ahead in clean energy production. The number of terawatts of power that China can turn on annually is, you know, vastly greater than the amount American can turn on in a decade so so that that's a
really very yeah it's more than just I mean I know you talk about micro economies I mean it's essential for AI I mean without without this Chinese energy revolution um the AI revolution isn't really viable is it
Well, you definitely need a lot of new energy. This week, Sam Altman talked about 30 gigawatts of energy and roughly a gigawatt being added per week at scale when he talked about why he's committed 1.4 trillion to various contracts over the last few weeks. And to be honest, 30 gigawatts is a small number compared to China's numbers. So it seems very clear that There's nothing to argue about here. China is clearly going to win in energy.
So China in the 21st century is a combination of the US and Saudi Arabia in the 20th century. It controls the energy and it also, I don't know if it controls the innovation, but it certainly controls
It controls its own rate of growth, is the truth of it. And, you know, there's a new 15 year plan that came out of the Chinese Communist Party's recent, I forget what they call it, like the plenum or whatever it's called, which kind of spells out very clearly what their plan is. It's not secret at all.
And does it speak? I mean, you've already talked about Sachs and Andreessen talking about how America wins the AI race as well as Jensen Huang's critique of American political ability to hold up progress. But does this, in particular, this clean energy revolution, does it speak of the superiority, at least in economic terms and technological terms, of the Chinese model? Someone presumably, in the heart, in the power of the Chinese Communist Party, made a call that clean energy was the thing. They've invested billions, probably trillions of dollars in this thing, which is inconceivable in the United States.
Well, look, let's give the US a little bit of credit. They're getting close. Close to what? To that style of execution. So the US Energy Secretary has embraced nuclear fusion and fission. They are talking about deregulating the permitting required to build energy. It is a highly centralized Washington DC based set of policies that Sachs is in the middle of, that is attempting to mirror that Chinese ability to plan and execute. So I don't think America is is sitting on its hands. I do think there's innovation. It's just that it pales by scale at the moment. But the right, I think Andreessen and Sachs and Trump and the people who work for Trump to give them credit in this domain are aware of the challenge and are kind of stepping up to it.
I wonder whether we can talk about less the convergence of the U.S. and the Chinese model, actually the convergence. One of the books on the long list for the FT Book of the Year is Cory Doctorow's Enshittification. But I have a new word. I call it enstatification, where increasingly the U.S. state's becoming more powerful. You talked about Trump and Sachs. They already invested in Intel, the US government. The stock is up 50% over the last month. Are we talking about an increasing convergence of the Chinese and US model when it comes to innovation and investment?
Well, obviously, the main difference is a one-party system versus a multi-party system. But I think once you put that on the side for a separate conversation, the state getting involved in planning is inevitable. And the trend is for it to get more and more involved. In fact, if AI works and creates abundance, I would argue the state has to get very involved in procuring and distributing the wealth it produces. And so you're right, there is an end-stateification. The question then becomes, which state on whose behalf with which outcomes?
Right. And the big news, the other big news of the week is, Sam Altman's always making news. He's made it clear that he doesn't think open AI is trying to become too big to fail. He denies the fact that, and lots of pieces that you link to, a journal piece about whether open AI is becoming too big to fail. I guess the implicit in it is that if he makes it too big to fail, when they fail, it will require state investment. But The increasing intimacy, if you like, of the health of open AI and the health of the US economy means that for better or worse, open AI might be too big to fail. Even if Sam himself is not consciously fashioning it, it becomes the reality and it makes... whatever happens in the economy, when there is a correction, most people think there will be one, it's probably more likely to be like 2008 than 1999. Yeah,
I would separate OpenAI, the corporate entity, from the investments and infrastructure that will be built by it. If the corporate entity was to fail, It would be a little bit like when the entity that built the channel tunnel between London and France failed, not the tunnel, the train ride that went through the tunnel, in that the tunnel remained and is now profitable. The infrastructure that OpenAI builds will be built and ultimately will serve the needs of distributing knowledge and intelligence to the world If OpenAI, the corporation, were to fail for some reason, which doesn't seem likely, Allman did a very credible job of explaining why that's unlikely, which he mainly talked about the fact that by 2027, there'll be 100 billion run rate in revenue.
Yeah, Keith, you're the ultimate. startup entrepreneur, you've raised round after round of finance. You know better than I do that anyone who talks about 2027 is just blowing smoke. Who knows? Anyone can say anything. You can say 2027, we're going to be making a trillion dollars, 10 trillion.
No, no, Andrew, you be a bit fairer. Me, I'm fair. Well, by December, which is only a month away, they will be at a run rate of 20 billion. a year. So they have to 5x that in two years. So if you do 20 times 2, which is just doubling, you get to 40. And if you do 40 times 2, you get to 80. So they'd only have to double two years in a row.
Yeah, but we've done so. I mean, there is also the cost, the fact they're not profitable. But certainly... Certainly, it's an interesting development, Keith. It certainly makes it different from the dot-com boom, doesn't it? These companies are so big now. There was no equivalent. I mean, the fact that if Pets.com or even Amazon had gone bust in 1999 or 2000, it never would have triggered any kind of concern from the US government about the general health of the US economy.
Yeah, and Sachs did answer that this week, by the way. He said that there's no question of the government ever underwriting OpenAI. So why? Because if it failed, there's at least five other companies that are in the game. So it isn't that AI would fail, it's just that a single corporate entity would fail.
he has to say that. I mean, he's not going to say that the government would step in. This has made a lot of... news in silicon valley uh i know tech crunch has got involved what's your sense of a silicon valley take on this i mean might if if silicon if let's just say open ai got into economic trouble i don't know if stuff happens that we don't know about which always does Are there going to be people in Silicon Valley who believe that the government needs to step in?
No, because it isn't the banking system that the dollar relies on. It's just an industry. And the governments don't step in. If the whole industry was failing, possibly there'd be some strategic reason for the US government to get involved. But if it's only a company, there's no reason to get involved because it isn't systemic. It doesn't really impact... whether the US is succeeding or failing.
Interesting. Meanwhile, the reality, for better or worse, is that you linked to this excellent piece in the FT about how big tech's market dominance is becoming ever more extreme. My interview of the week was with Jacob Silverman, who has a new book out, Gilded Rage, Elon Musk and the Radicalization of Silicon Valley. In terms of your narrative, Keith, of this competition, it's not a race, you say, between China and the US, how does the increasing power of big tech, of Google and Microsoft and Amazon and a handful of other big tech companies, how does that shape this?
Well, that's where there's a real race. And the race is to own the revenue that AI is producing. If you did a pie chart of the revenue from AI today, OpenAI would be getting a big chunk of it. Microsoft would be getting a big chunk of it. Google would be getting a big chunk of it. Anthropic, a smaller chunk. and then everybody else much smaller after that, including the Chinese companies who are mainly open source and therefore don't get paid for the AI. They have to get paid for services and token sales, which is usage-based. Quen is getting a lot of usage, so it's probably making a decent amount of revenue from selling tokens. And that's the real race. Now, Microsoft locked in seven years where OpenAI can't distribute its core models to anybody else in a deal last week. So Microsoft has locked in a seven-year success in owning its share of that revenue. OpenAI is shipping new products pretty much every week, and its revenue is growing super fast, so it's probably fine. Anthropic is having some struggles. It's getting dropped and replaced by Quen in a number of places, although it's pretty strong in enterprise. So I'm not worried about it. Perplexity, maybe a bit more worried about, although it did just do a deal with Snap.
Okay, I take your point on that. But my question is, what happens when In this competition between the United States and China, as China becomes larger and larger, more and more powerful as their state capitalist model, the US model is represented by a tiny handful of big tech companies. Does that change anything?
Well, those companies are sucking most of the revenue out of the AI system. China is a revenue laggard, even though it's an investing leader. And so to get revenue, you've got to globalize fast and productize fast. And I would say that because of OpenAI mainly, American AI is productized where consumers are using it and paying for it. And China hasn't done that yet. So big tech is America's savior at the moment.
So in your editorial, when you ask about is China the new America, it's not really the new America because it doesn't have a Google or a Microsoft or an Amazon, does it?
But they are still, I mean, they may not be state-owned companies, but they're intimately bound up with the Chinese centralized state, whereas these American companies aren't.
Well, I think... Andrew, I think on the margin you can argue that, but I would say the amount of time a big tech execs spending DC these days tells you that they've got their hands in each other's pockets just the same.
Basically the same. Alibaba is not state-owned. It's listed actually on American stock exchanges. and trades on American stock exchanges, mainly with American investors. So this idea of the Chinese state is a puppeteer is a fiction.
So you're saying another of the books for the FT Book of the Year is by Eva Du. Another interesting book is House of Huawei. Are you suggesting that Huawei then and...
That's a more nuanced conversation than Alibaba and Google. I do think Huawei is a global innovator in hardware, especially internet hardware. I do think it's subject to the law of China because it's based in China. It wouldn't surprise me that it's been forced to put back doors into things just like the UK government's trying to persuade Apple to put a back door into the iPhone.
Are you suggesting that, I don't know, if Trump wants Google to do something, he just picks up the phone, calls Sundar and says, look, you need to do this just as presumably happens in Beijing.
American companies can use the legal system to fight back, as Apple has done with the UK government and the American government on occasion. Take end-to-end encryption is a great example. End-to-end encryption is a constant battle. The EU is trying to get rid of it. But the companies have used the legal system, at least up until now, to resist that. But there is still this attempt by the state to get control over tech.
In other words, it's a profoundly different system. Anyway, let's end with your idealism. You don't think it's a race, Keith, and your editorial is quite provocative this week. I don't agree, but you argue in your closing thought that And in terms of this quote unquote race or competition between China and America in the 21st century, and I'm quoting you here, the real opportunity lies not in rivalry, but in reciprocity. What do you mean by that?
That is $26 trillion of annual addition to GDP. We're nowhere near that today. I mean, analysts say a lot of things. Which analysts have said that? McKinsey.
Just do a search on what is the upside case for AI's addition to GDP, and you'll find that 20% is a rough consensus on upside. But it doesn't really matter what the number is. Let's just assume there's upside. We're at a tiny fraction of that today. 20 billion is what OpenAI is saying it will do this year as a run rate. So that compared to 20 trillion is a thousand times smaller. So you've got roughly a period of 10 years that implies the potential for massive growth. If that is true, Chinese and American companies are both going to be trying to get their share of it. And as they do, they're going to be producing products that hopefully we want to use and pay for. That's why it's reciprocity.
I mean, you could have said the same in the Industrial Revolution, that it's in the interest of the English and the Germans and the Americans to be reciprocal. That's just not the way the world works.
So what does that mean in the AI economy? And I'm quoting you here. I'm quoting your closing thoughts on your editorial. The 20th century rewarded nations that industrialized. The 21st will reward those that, and I'm quoting you here, I don't know what this means, humanize intelligence, aligning technology with global well-being rather than national dominance. What does that mean?
It means that as you ship products that people pay for, a massive amount of new wealth will be created everywhere. The simple version would be a rising tide lifts all boats. And that's what it's arguing. And therefore the idea that there will be a loser, neither America nor China will lose if you turn it around. Whether one of them will win is moot, but neither one of them is gonna lose. And neither, by the way, will Europe lose. Europe will be a net beneficiary of this as well, just not as an owner of AI, but as a user generating wealth from its use. Yes.
Well, there you have it. Everyone's going to win. I have to admit that I find that unconvincing, but maybe I'm too pessimistic. Keith believes that everyone's going to win in the age of AI. We're going to humanize intelligence. So we'll all be winners, although some will be bigger winners to borrow... from Orwell than others.
Exactly. Not all winners are equal, but we will see, Keith. A very provocative conversation and editorial this week. His editorial is entitled, Is America the New China? Might be Is America the New America? Is China the New America? Might be Is America the New China? We're increasingly then perhaps in many ways converging. Have a good week, Keith. And you and I will converge again next week.