Anirudh Suri

Full Transcript

The Old World Order is Dead. What Comes Next? | Ian Bremmer, President, Eurasia Group

TRANSCRIPT

Anirudh: Hi, and welcome to The Great Tech Game podcast, where we bring you some of the world’s top thinkers at the intersection of technology, geopolitics, and global affairs. Today I’m excited to have Dr. Ian Bremmer with me. Dr. Bremmer is the founder and president of the Eurasia Group, one of the world’s leading global research and advisory firms, and also GZERO Media, a digital media company providing intelligent and engaging news coverage of international affairs.

Ian helps business leaders, policymakers, and the general public make sense of what’s going on around them through speeches, commentary, and, for some of you who might have seen it, satirical puppets. I’m excited to have Ian Bremmer here with me, because he’s someone I’ve followed for a long time now, and he provides an independent view of what’s going on and a lot of what’s happening behind the scenes.

So today I’m excited to have you with us, Ian. Welcome.

Ian: Thanks, happy to join you.

Overall Assessment of Global Geopolitics

Anirudh: I want to start off our conversation today with a big-picture question for everyone watching. Global affairs today are hard to make sense of. The last few years have seen geopolitical churn, geoeconomic battles pan out, and hot wars emerge very quickly and often unexpectedly. We’d love to get your view on what’s really at play here.

Ian: We all know that there are economic boom-and-bust cycles. We experience them frequently. We have terminology for it: recession and depression. We’ve got monetary and fiscal tools to respond. Geopolitics, in other words what’s happening in the world, also has cycles, but they’re long cycles. Because they’re long cycles, people don’t necessarily understand that they’re in one when it’s occurring.

What’s happening right now is what I call a G-Zero world, not a G7, not a G20, but an absence of global leadership. That means we’re in a geopolitical recession, where the underlying balance of power no longer fits with the rules of the road, the institutions, the architecture, and the policies that are supposed to govern it. There are three reasons for that, in increasing order of importance.

First, back when the Soviet Union collapsed, Russia was not brought into the West. NATO did not include Russia, the EU did not expand to include Russia, and the G7 became a G8, but the Russians were excluded from all the important meetings. They are angry about that. They blame the West and the United States. So now we have a nuclear-armed Russia acting like a revisionist, rogue country, with its closest ally probably North Korea, and after that countries like China and Iran, and they’re effectively at war with the West. That’s the first reason.

The second reason is that China was brought into the global system, particularly the global economy, when the Americans welcomed its integration into the WTO. The presumption on the part of the West was that as China became more powerful and wealthier, its political and economic system would align more with the Western advanced democracies, that it would become Western. Today China is much more powerful than the second most powerful country in the world. It’s building its military, its economy is potent, and it has some of the top technology in the world, beating the US in many areas. And yet it’s not remotely Western. It’s still state capitalist, still authoritarian, and most countries in the West, particularly the US and Europe, mistrust it and see it as a competitor and an adversary. That’s the second reason.

The third reason, the most important one, is that a majority of Americans increasingly do not support the leadership role that America fashioned for itself after World War II, in this last emerging geopolitical boom cycle. What do they oppose? They oppose the US acting as the global sheriff; they want other countries to pay for their own defense. They oppose the US promoting a global free-trade architecture; they want more industrial policy, subsidies, insourcing, and reshoring. They oppose comparatively open borders, legal or illegal, and want much more restrictive border security. And they don’t really support the promotion of democracy or the rule of law around the world. In fact, many Americans aren’t convinced that they have rule of law or democracy in their own system in the United States. So Americans are actively refusing to play the global leadership role that underpinned the world that so many watching this right now have understood over the past decades.

I’d argue those are the three things that brought us into a G-Zero world. President Trump is a symptom of those developments, not the cause. He is certainly a beneficiary and an accelerant, and he has some unique qualities in terms of his unfitness for the role that are causing additional geopolitical uncertainty tied specifically to him, as opposed to the G-Zero broadly. But those things will go away once he goes away. The broader points we just discussed are structural; they are not going away.

The final thing is that when you’re in a geopolitical recession and want to avoid a geopolitical depression, which is World War III, and I think all of us recognize that we don’t want that, there’s a lot of conflict, forced migration, and hardship, but there’s also a lot of opportunity. I don’t believe we’re heading for World War III. There are really only two ways to avoid war in a geopolitical recession. The first is that you rebuild, strengthen, and reform your existing institutions so they better reflect the changed global order and balance of power. Or you build new institutions, for climate, for AI, reflecting new powers that are stronger than they used to be, or a new category of actors that weren’t even geopolitical actors in the past. That’s the only way out. We are starting to do some of those things, but clearly nowhere near fast enough to give anyone confidence that we’re about to emerge from the G-Zero.

When I first recognized that we were heading into this geopolitical recession, it was back in 2012, and I really wanted to be wrong. It felt very overdetermined to me, and I certainly didn’t want it to last this long. But who cares what I want? That’s not relevant analytically. So here we are, and that’s the best way I can describe the world we’re in. I think it’s important to start with this if you’re having a big conversation, because so many of the headlines, so much of the conflict and uncertainty that people are upset by and anxious about, come from those three structural factors. You can look back and say, this is what caused this geopolitical recession, this is why leadership is falling apart, this is why we are eating the seed corn of the institutional resilience that exists, and very little is being built up.

Anirudh: I’ve talked a great deal with my friend Yuval Harari over the years. I think it’s great that you laid out that big-picture view of where you see us in the cycle. I think the point about being in cycles and not realizing it is quite apt, even from my perspective. I want to double-click and deep-dive into a few points you’ve made.

Historical Analogs for Today’s Situation

Anirudh: The first is that you mentioned we’re in a cycle, and often we don’t realize we’re living through a certain cycle. That prompts me to ask: given the vantage point that you have, what historical analogs do you find helpful when you think about this? People have mentioned pre–World War I as a possible analog, both economically and geopolitically. What are some other historical analogs that come to mind?

Ian: The funny thing is, we just had the US-China summit, where President Trump went to Beijing, and Xi Jinping himself used a historical analog. He talked about the Thucydides Trap, which Graham Allison from Harvard, a friend of mine, has written extensively about: the idea that in roughly fifteen or seventeen prior instances, the world has found itself in a similar environment. I’ll explain what this means. It comes from Athens and Sparta, the rising Sparta and the declining Athens, with Athens trying to hold on to its order in its region, which was the world as far as it was concerned at that point, and Sparta challenging it. Most of the time, when you have a rising power and a declining power like that, it leads to war. So Xi Jinping is saying, we want to avoid the Thucydides Trap, and the part he’s implying, which is why it’s a convenient frame for China too, is that there are only two powers that matter, the US and China, and the US is declining while China is rising. If the Chinese are patient, strategic, and far-sighted, they will avoid the Thucydides Trap, and the declining power needs to appreciate their forbearance. Then Trump, of course, said the US was in decline under Biden and is doing great now under Trump.

But that’s not really what’s happening. The world order is not about the US declining and China rising. China is rising, of course, but America’s allies are declining. Look at Europe, Canada, Japan, South Korea: most of those countries have a precipitous decrease in their demographics. They are well behind the Americans in growth, productivity, and efficiency gains. They’re not investing adequately in technology or in defense, though that’s starting to change in some of those countries. And it’s the Americans themselves who are saying they don’t want to do this; they are rejecting their own global order. So it is not remotely a Thucydides Trap.

China, in my view, is facing some very serious problems. It’s rising in absolute power and making major investments in many technologies, but it’s in trouble. It’s not growing very fast, unemployment is too high, and it has many sectors of its economy that are wildly unproductive and indebted with bad debt that isn’t being addressed. And the demographics are disastrous. So, in some ways, we are in uncharted territory, and I suspect this is part of why you wanted to talk to me: we are about to see an entirely new group of geopolitical actors, which hasn’t happened since the beginning of the Westphalian system. These are non-state actors that act geopolitically: technology companies, AI companies, that are about to transform not just geopolitics but society and humanity as a whole, because of AI.

As much as we entered this geopolitical recession with traditional actors and a more traditional understanding, wanting to avoid a Cold War, imagining bloc behavior, wondering whether the world would be multipolar, whether regionalism is developing, what Europe is going to do, how India is increasingly relevant economically but not yet in other areas, if we get out of this geopolitical recession intact, it will be with an entirely additional set of new geopolitical actors at the table. The institution-building will need to be different, the response will need to be different. We will be creating a new architecture. I find this incredibly exciting. The field of political science and international relations is nowhere near prepared to handle this right now. Just look at the Pope’s latest encyclical and you get a sense of how unprepared we all are. But it’s wildly exciting, because we get to be present for, and play a role in, a new way of organizing humanity.

Anyone living in the world today, if you’re anxious, part of it is because you don’t trust your leaders. When the most powerful leaders on the planet today are Xi Jinping, Trump, and Putin, men in their seventies, Trump basically eighty, incredibly arrogant, convinced they’re right about everything, out of touch with a lot of what’s happening with young people today, and with no interest whatsoever in giving up power, you can understand why people would be anxious. Nobody would want the world run by people like that. You wouldn’t pick them to be responsible for what will or won’t happen to society and the global order over the next five or ten years. You wouldn’t want your kids raised in a society where those are the people calling the shots. That’s the situation. So the fact that it’s not going to last, and that people want something else, I think is a genuinely promising thing.

Anirudh: There’s so much to unpack in what you just described. What I want to double down on next is the point you made about new geopolitical actors emerging, by which I assume you mean the large tech firms and similar actors. Related to that: for the longest time, the mainstream discourse has framed the US political and economic system against the Chinese one. You often hear that China is a state-led model, the US is a market-led model, and the Europeans are characterized as a regulation-led model, a regulatory superpower, as many call them. Setting Europe aside for a moment: as China has risen with what’s seen as a state-led capitalism model, or “capitalism with Chinese characteristics,” and its industrial policy has yielded great results in the form of leadership in many critical tech sectors, it seems the lesson the US is drawing is almost the opposite of what one might have expected. Instead of China becoming more democratic and more Western post-WTO accession, it seems the US is moving closer to the Chinese model rather than the other way around. You’ve spoken about America now possibly being a state-capitalist economy with American characteristics. How are you seeing these two models, the Chinese and the American, evolve, and remain distinct?

Ian: I don’t want to put Europe aside, for two reasons, one of which we’ll come back to: as AI starts changing society, we may have a lot more to learn from Europe than in an environment where all we care about is growth and the almighty dollar. But that’s a different question.

There are these three different models. The European model is one where the government acts independently from the private sector as a regulator meant to ensure a social contract is met for its people. The problem is that while it has regulatory superpower, that comes at the expense of growth and entrepreneurship, and it’s very hard to provide for the social contract, even if you’re aligned and capable of doing it, if you don’t have the growth, because people will get angry if you don’t have the money to spend.

The Americans and the Chinese are both driving growth, but historically in very different ways. The United States does it through the private sector capturing the regulatory environment through what is effectively a coin-operated democracy. Most Americans who think the US isn’t a truly representative system believe there are two tiers of justice, two tiers of law, two tiers of power: one for those able to lobby and pay for the policy outcomes they want, and one for everyone else. And it turns out the president of the United States today can do that himself, for himself and his family, in ways that are unprecedented. In China, the state captures the corporations and determines what is and isn’t patriotic behavior, where money should be invested, and what is out of bounds.

Now, it’s certainly true that the US government is leaning into tariffs, which means less free market, but is that really so different from how the US system worked before, or is it still the private sector calling most of the shots? It’s leaning into subsidies, but how much of that is really driven by government versus by private-sector actors capturing the government for their own purposes? Certainly the DOGE experiment, with Elon Musk running it while directly embedded in the private sector, implied the latter, though that failed quite quickly.

So I think it’s premature to say the Americans are taking big lessons from China. The equity stakes the US government is taking are still small compared to the degree of regulatory capture that lets companies drive the outcomes they want, even when those aren’t necessarily in the national interest, whether that’s Nvidia or the big AI companies ensuring that even a limited amount of regulatory oversight for testing their models doesn’t get presidential approval. It’s too early to say otherwise, but it is very clear these are the two models driving growth. In China, I’d still argue it’s the state determining winners and losers, and the state will be the principal AI actor. In the US, and for governments working with the US, it’s actually the companies, and the individuals running them, that are functioning as sovereigns in the geopolitical space.

I experience this constantly as an analyst. I’ll meet with the CEO of a major AI or tech firm and get more geopolitical insight from what they’re doing, who they’re talking to, how they’re deploying capital, and what their business model is, than I get from talking to the head of state of a fairly major country. That’s pretty remarkable.

Anirudh: Let’s talk about them. The tech CEOs and the tech firms are two distinct analytical entities. The tech CEOs, as individuals, especially if they’re founders, are massively rich and influential. Tech firms have capabilities now that are increasingly critical to states, Ukraine being one example, and arguably Iran as well, among other instances. You’ve spoken about the word “sovereigns.” How do you see this? Are they becoming co-sovereigns with states? Are states still dominant, in your view, but losing control over time? Or do you see them as still ultimately subservient to states and state interests, doing the best they can within the bounds of what a country’s leaders want?

Ian: Setting China aside, because it’s a different system, we’re talking about the rest of the world outside China and outside countries functionally captured by China, and that’s still most of the world, all the advanced industrial democracies, India, most of the Global South.

I’d say it this way, because it’s changing very quickly. We can still talk about a global security order, which is mostly not about tech companies yet, though it’s starting to change. That order is still dominated by the United States. China is spending a lot of money trying to catch up, and it’s certainly an important regional player in defense, but it’s not really a global one outside its own region, where it will also be challenged by other countries spending a lot more going forward, India being one example.

Then you have a global economy that’s also changing, where tech companies are becoming much more important. In the US, tech companies are the most important drivers of growth and of new infrastructure spending right now; everything seems aligned toward it, and it’s causing a lot of disruption. But if you think about the global economy overall, we’re still talking principally about sovereigns, central banks, and ministers of finance, and that world is becoming very multipolar. You can’t talk about the global economy today without talking about the US, China, Europe, and increasingly Japan, India, and a few other places.

But the digital order, everything that is online, is increasingly determined by technology companies. So if you have three different competing orders, a security order that’s mostly unipolar though changing, an economic order that’s mostly multipolar and shifting, and a digital order that’s mostly techno-polar, then the question becomes: what does that look like in five or ten years? It appears the digital order is not only moving by far the fastest, but is also becoming far more powerful, including in other areas of geopolitical power.

For example, the war in Ukraine has changed a lot recently, and the Ukrainians are doing much better now. Some of that is because they’re spending a lot of money and learning a lot about new technology and how to fight with it. But one of the most important developments is that Elon Musk decided to cut the Russians off from Starlink. Literally with the snap of his fingers, and this was his decision, not the US government’s; no government made that decision, and he made it with no oversight. I’m glad he made that particular decision, it’s a good one, but we’re not talking about whether I like it, we’re talking about the significance of the fact that someone like him can make that call at all. Suddenly the Russians went from being able to effectively fight a drone war against the Ukrainians to being a year behind, which is a long way behind in unmanned vehicles and drones. Now the Ukrainians are taking territory and launching weapons deeper into Russian territory, and the Russians can’t do the same in return to the same degree.

Anirudh: Could Elon have done what he did without the consent of, or some kind of backing from, the US government? And if he had done the reverse, would the US government have stepped in?

Ian: He did do the reverse a couple of years ago, when the Russians announced, by their own constitution, that several pieces of Ukrainian territory were now Russian territory, and that meant Starlink allowing Ukrainian generals to use it in those occupied territories, which Russia considered its own, was acting illegally, in their view. For a period, Starlink was shut down there. Ukrainian generals were suddenly flying blind with their drones, and people were killed and territory was lost as a consequence. The American government had nothing to say about it. My view at the time was that the right thing to do would have been for someone else to take that risk, not for Musk personally. The Department of Defense should be the one making those decisions: you provide the technology, they buy it, but the decision of where it is or isn’t switched on should be made by states. I lost that argument.

So the answer to your question is clearly no. It’s a very important question. That’s a specific example, but it’s not just about how defense is changing. It’s also about what it means to be a human being. The reason I’m a political scientist is that, at some level, the thing that interests me most is how people interact with people. Political science is fundamentally the study of how large-scale human interactions work, how we organize them, how we govern them, and those things arise because human beings exist in a certain way. We’re social animals, competitive, familial and community-oriented, but also individualists, with different cultures, backgrounds, and, to a lesser extent, genetics.

I firmly believe that within the next ten years, AI will fundamentally change what it means to be a human being, that we will be different in terms of our social interactions, as each of us individually has an AI device, or a set of agents, that engages with us, knows all our data, is programmed by us, by companies, or by governments (perhaps in China’s case, or some other places), knows us better than anyone else, and is never shut off. That will fundamentally change who we are as human beings, and it will have a fundamental impact on the kind of governance models we will want and need, and that we will create. That’s going to change geopolitics completely.

In other words, I believe the digital order is set to become the dominant geopolitical order in a comparatively short period of time. God willing, short of getting hit by a bus, you and I, and pretty much everybody watching, will experience this. That is extraordinary. We’re not ready for it. But it means these technology actors, who right now are only dominant within their own particular sandbox, matter a great deal if the whole world becomes that sandbox, and that’s where I think we’re heading. That doesn’t mean the technology companies will remain dominant forever, or that we won’t see backlash, bottom-up from individuals and top-down from states, or that the technology will keep developing the way it’s presently developing, in a way that gives a small number of people who control the large language models and the data all the power. It doesn’t have to continue that way over the next ten or twenty years. Those are all interesting questions. But the best we can tell right now is that we are heading toward techno-polarity, and I think that’s a very big deal.

Anirudh: When I wrote The Great Tech Game, I talked about how tech was shaping the geopolitics of our times, becoming a key instrument of power but also a real shaper of the geopolitical objectives of countries themselves. I had a chapter on this exact relationship between governments and tech firms that I titled “The Tug of War.” Five years ago, there was a real battle between the two, states trying to exert influence through antitrust and other tools of government power to keep tech firms in check. Today, as you’re describing it, I’d actually say that sandbox is quite large, much bigger than perhaps you’re describing, because these firms are capturing national-security domains, impinging on currency as a domain, and increasingly encroaching on all the traditional domains of the nation-state. So I don’t think that box is actually that small. Yet five years ago that relationship was very different, and today they seem much more hand in glove, almost an alliance between the two. Do you expect that to last? Is it social backlash that you think might trigger that relationship souring, or do you think this kind of alliance is here to stay for at least the next five to ten years?

Ian: It’s amazing, because five or ten years feels like the absolute maximum we can even think about here. I’d say we should be asking about a year or two, given the Trump timeline, though I know you shouldn’t put it that way. I wrote about the G-Zero in 2012, and here we are in 2026, and I don’t think, even with a lot more experience and the same tools, you can make a fifteen-year-ahead prediction with confidence today. It’s too fast, so we don’t fully know.

For now, the backlash I’d focus on is the social one. I do think that by 2028 there will be at least one major candidate, assuming there are two and it’s both traditional parties, and neither blows apart, though who knows, whose campaign is heavily driven by a populist response to AI. That’s not baked in; Gavin Newsom, for instance, probably wouldn’t run that way if he got the nomination. But I think most of the people likely to run would. The logic is there because people are scared and angry, and they’re also wildly anti-corruption and anti-privilege, and they connect the two. Even though Trump himself is much more focused on World Liberty Financial and his family’s crypto ventures than on AI specifically, voters won’t necessarily differentiate between that, Elon Musk, and the longer-term controlling companies that are more traditional industrial-type powers, like Google and Microsoft and the rest.

So I think a backlash is coming, and part of the open question is whether it’s an overall backlash, wanting to bring these companies down because people don’t like them, or whether it’s much more focused on specific things: redistribution and taxing usage so intelligence can’t simply be treated as a utility only wealthy people can access, or a real constraint on the pipes and plumbing, meaning you simply won’t give these companies the energy or let them build the data centers they’d need to keep scaling large language models the way they’re presently all-in on doing. Google, for instance, is reportedly selling shares to raise another eighty billion dollars for data centers, something I hadn’t seen announced before just a few hours ago. So there’s certainly potential for backlash on that front.

But it’s also possible that some of these “trees grow to the sky” assumptions about LLMs, logarithmic gains in efficiency alongside rising costs, simply don’t hold: that the costs are too high to pass on profitably to consumers, whether industrial, governmental, or individual, and that it’s the fast followers who become capable enough that the big dominant companies implode. If that happens, we’d see a far more level playing field, with many competitors both in the US and around the world capable of disrupting the biggest companies, and the next generation of geopolitical actors would be far more decentralized. That’s possible, and I could see it.

In that scenario, we wouldn’t be talking about states, but we also wouldn’t be talking about four or five individual actors worth a trillion dollars each acting like super-governments. It might be something quite different. But if you made me bet, my bet for the next five years would be that the benefits of scale, the informational advantage of identifying what’s coming before anyone else and either buying it or squeezing and starving it out, plus the ability to collect extraordinary amounts of data and use it effectively, will be a huge advantage for China, making its government more efficient as a tech-empowered state-capitalist system, and will also be a huge advantage for a handful of tech giants, the hyperscalers, in the US. Everyone else will be a taker, not a maker, of those advantages.

I’m sure you’ve read “AI 2027,” which implies that the OpenAI model is the one that wins and the Chinese have a separate track. There’s a lot of uncertainty around that, but as of 2026 I still think that’s probably the right baseline bet, if you’re forced to make one.

Middle Powers and Sovereign AI

Anirudh: One of the interesting things that came up during the India AI Impact Summit a few months ago was this question of who the ultimate consumers of this backlash are: ultimately, the people. You’ve made a great analogy between economic and geopolitical recessions, and I think ultimately, for tech firms just as for governments, their consumers are their constituency. Is it possible that among the different tech firms leading the AI wave, some will take a much more pro-consumer position to avoid backlash, and might emerge as, let’s say, more “welfare-AI” firms than others? We’re already seeing a bit of this positioning between OpenAI and Anthropic.

Ian: I’m not convinced that will actually happen. For now it’s been positioning, not really backed by real differences in strategy. The analogy here is that twenty or thirty years ago, a lot of people believed that if China didn’t become much more open as a political system and more responsive to the demands of its people, there would be extraordinary social backlash. That did not happen. China has actually been able to be wildly repressive through the creation of a surveillance society. Look at the Uyghurs, millions of people who have been completely suppressed and forcibly integrated into Chinese society, and it’s now barely discussed, because the state was able to fully contain the issue through repressive and coercive measures.

It’s quite plausible that tech companies, which need to keep raising money to succeed, could have some of those same advantages in grabbing more and more of the attention, the eyeballs, and the data of the people on their platforms. You said maybe one of them will become more consumer-focused. I’d like that to happen, but companies fundamentally don’t think first and foremost about the long-term question of how to make sure the population at large likes them. That’s what governments are supposed to do, which is why, when the regulatory function falls apart, you have a problem, because companies don’t suddenly step in to fill that role.

Middle powers are also increasingly worried. For a long time they participated in the consumer internet era without worrying much about its geopolitical, national-security, or even economic implications for them. I think a lot more countries outside the US and China are now worried about what this techno-polar order means for them, and whether they’ll have any influence over what are, effectively, foreign tech firms running their societies, economies, and possibly their national-security infrastructure.

Anirudh: So the question emerges, and you’ve spoken about this: is sovereign AI, this attempt by many middle powers today, the UK, European countries, Japan, Middle Eastern countries, and others, to build something that gives them strategic autonomy or decision-making power, really a myth, as you’ve suggested? Or are there genuine paths emerging for other countries, companies based in India, Singapore, Japan, the UK, or France, to become influential actors in this techno-polar order?

Ian: If what these countries are trying to do is build a competing tech stack to what you see in the US and China, it’s a lost cause, because they will never be able to compete with or catch up to what the Americans and the Chinese are doing. And if it turns out the US and China are building extremely expensive models so that fast followers can take advantage once things become commoditized, then middle powers waste a lot of money on things that will be near-free to use in three to five years anyway. So that’s a bad idea in that form.

But that doesn’t mean sovereign AI as a whole is a lost cause. It just means you can’t think about it holistically; you have to think about which specific components of sovereign AI an individual country can actually compete on. If you’re in the Gulf, for example, a big component of sovereign AI is that you have incredibly cheap energy and very little zoning regulation, so you can make it possible for companies to build extraordinary models there. In return, you should think of this as an alliance with the company investing in your country, the kind of alliance where both sides think about what benefits accrue to each. That gives you leverage, and that leverage should allow you input into the business model of the company operating there.

Imagine Amazon became a consumerist, extractive, surveillance-oriented AI superpower in the US, but was also building massive data centers in the Gulf. If I were, say, the leader of the UAE, I wouldn’t necessarily want that AI to run the way it runs in the United States. I’d want it to run differently, according to my national priorities, my sensibilities, and the priorities of my citizens, over the long term. There’s no reason an alliance between, say, the UAE and Amazon couldn’t have that kind of push and pull built into it.

I think middle powers would be much better off if they found ways to coordinate, which would give them more leverage against these companies. If it’s just one country making a deal, you’re in a much weaker position than if it’s a group, which is why Norway is now thinking about the EU, and why the UK Treasury is treating reversing Brexit as something close to an inevitability: because in a G-Zero world, scale matters for influence, and not just against Trump, but with the tech companies too.

When I see the French doing a huge deal with SoftBank to fund fifty billion dollars of data centers off French nuclear power, that’s a great thing for a middle-power strategy on one hand, because France has uniquely built up all this nuclear energy, and it’s comparatively inexpensive and available without causing massive citizen backlash. That puts France in a much better position to have real influence over what kind of AI it develops for its citizens, and potentially to export that model to the Francophonie, the large number of French-speaking countries around the world. That’s a big deal.

But if France does that alone, rather than with the rest of the Europeans, where does that leave other EU members trying to negotiate similar deals with AI companies? In a much weaker position. That’s the challenge: we are nowhere close to having the kind of geopolitical governance we need to make these alliances work effectively, and as long as that fragmentation persists within the G-Zero, technology companies will end up in the stronger position.

Anirudh: Briefly, on everything we’ve touched on today, from the geoeconomic to the geopolitical to techno-polarity: where is India positioned today, and is it getting its strategy right?

Ian: I do think that after the US and China, India has done some of the most impressive things on AI. Having a huge population with an extraordinary amount of data is, by itself, a piece of leverage in attracting AI companies. There’s scale in India that makes it matter. That said, I worry about a lot of potential displacement. I already see American banks quietly talking about shutting down their offshoring operations in India, because AI can displace those roles, and also because Indian labor is extremely mobile and competitive, so companies are constantly churning through and fighting to retain people. There are also issues of data security: wouldn’t you rather have all of that in the United States, or with your headquarters? So I worry that even though Indian labor is comparatively inexpensive, that displacement risk is a real problem for India.

But I also see how much digitizing the Indian government’s benefits systems has allowed for growth without blowing out the budget, and there’s a lot to be said for having an AI-empowered government and AI-empowered ministers. I think India is increasingly oriented toward being a “legacy government at scale,” one of the countries best set up to deploy agents usefully. Part of this isn’t just about geopolitical power; it’s also about which companies will themselves be effective, and which governments will be effective. The UAE and Singapore are wildly impressive in terms of AI diffusion for their populations and utility for their governments. The US government is not; most people in the US government don’t use AI effectively, and that’s genuinely bad, because it will undermine the quality of government decision-making. We haven’t talked about this much, because it can feel banal, just looking at how the Department of Commerce or Education uses AI, who cares? But you should care a lot, because that differentiation may end up mattering a great deal. If you were a stock-market analyst, you’d be looking closely at which companies are getting this right; why aren’t we doing that with governments? There’s a whole set of interlocked interests here. We need to start thinking about governments more the way we’ve traditionally thought about companies, and about companies in the tech and AI space more the way we’ve traditionally thought about governments. That shift is not happening yet.

Rapid Fire

Anirudh: On that note, let me move into our rapid-fire segment, since we’re running out of time. Which region is flying under the radar right now, but that you expect will become a major geopolitical risk and player in the next couple of years?

Ian: The Gulf isn’t exactly flying under the radar, but it’s the region where the trajectory is going to change the most, on both sides.

Anirudh: Is there any risk that’s being overhyped right now, given that you publish a risk report every year? And, conversely, a risk we’re not discussing today that you think will emerge as a major one by 2027 or 2028?

Ian: Taiwan is wildly overhyped, and in my view not a near-to-medium-term problem at all. Underhyped is everything about new technologies disrupting not just drones but financial markets, and the cyber risk that comes once more powerful models become widely available and other actors start developing their own. All of these are underappreciated.

Anirudh: When do you think the world is likely to see a democratically elected tech CEO as the head of a state?

Ian: I don’t know, certainly within the next few years, given how many countries there are out there. The more interesting development, though, might be a tech CEO becoming head of something that is functionally a state but is a territory within an existing state, without being democratically elected.

Anirudh: Every year you release your Top Risks report for the year ahead. If, halfway through this year, you had to predict the top three risks for next year’s report, what would they be?

Ian: I wouldn’t tell you, but I appreciate the question.

Anirudh: In the contest between China and the US, what is each side’s biggest asset?

Ian: America’s biggest asset is creative destruction, and the opportunities and innovation that come from it. China’s biggest asset is the longer-term strategic orientation of its system.

Anirudh: You’ve had an accomplished career at the intersection of geopolitics, political science, and several other fields. For our younger viewers considering a career in political risk, or at this same intersection, how would you advise them, and how do you personally cut through all the information overload we’re living through today?

Ian: There are three things you need. First, you need to focus on content. You can’t succeed in this field unless you genuinely care, first and foremost, about understanding the issues. You have to read a lot, digest a lot, meet a lot of people, and engage. It’s about content, knowledge, understanding, and challenging yourself, recognizing that the world changes fast and your thinking needs to change with it. That’s number one.

Second, you have to learn to communicate that effectively across different formats: long form, short form, video, writing, in person, on a stage, one on one. That’s genuinely important.

Third, you need an understanding of, and a willingness to learn, how organizations actually work. You have to be able to deconstruct and reverse-engineer institutions, and engage with the architecture around you, because those are the power structures that will determine whether you succeed or fail. The order I’d prioritize is content first, communication second, and reverse-engineering the architecture third. But all three are essential if you want to succeed.

Cutting through the noise runs through all three of those. It comes down to curiosity, recognizing your own biases, and a genuine humility about the fact that what you want has nothing to do with the analysis of the world around you. If those things feel uncomfortable to you, you probably shouldn’t be in this field. You should do something else.

Anirudh: Lastly, what books would you recommend for people interested in the topics we’ve discussed, and what guests would you recommend we have next?

Ian: You’re going to love my next book. It’s coming out next year, almost finished, and it’s about a lot of the topics we’ve discussed here: geopolitical cycles over the long term, and where things go next. I’ve been thinking about it for three years now, in some ways for my entire career, and I think it will land well with people who enjoyed this conversation.

As for other books, I’d say right now Mustafa Suleyman’s “The Coming Wave” on AI, which is about two years old now but is still essential reading. “Superintelligence” by Nick Bostrom, and really all the foundational work in this space. And the “AI 2027” piece I mentioned earlier, which came out about a year and a half ago, is as good a place as any to read about how AI is going to transform society in ways we don’t yet fully appreciate.

For guests: if you haven’t had Yuval on yet, you should, he’s wonderful and thoughtful. Jeremy Allaire, the CEO of Circle, is extremely thoughtful on this material, and would be wonderful if he’s willing to talk with you. Amy Webb has also been really good on her own models of thinking about this; I haven’t spoken with her in a while, but I think she’s excellent. There are a lot of good people out there; the field is becoming more populated.

Anirudh: This has been fascinating, a great set of recommendations. I have to go, but thank you for being on the show.

Ian: Thank you.