2026 GZERO Summit in London

Where LSEG and Eurasia Group convene global leaders to address the geopolitical forces shaping markets, economies and business.

On 8 October 2026, LSEG and Eurasia Group hosted the GZERO Summit in London for the second time, convening senior leaders from across the public and private sectors to explore the geopolitical forces shaping markets.

Against a backdrop of pivotal European elections and a closely watched US midterm cycle, discussions examined the evolving political, economic and strategic landscape, spanning:

  • Competitiveness and trade
  • Technology and innovation
  • Politics and governance
  • Defence and security

The Summit explored the implications of these shifts for business and markets, with particular focus on transatlantic relations, energy security, global trade and the growing impact of AI on economies, markets and society.

Featured perspectives

The Summit brought together perspectives from leaders spanning business, technology, policy and geopolitics, including:

  • David Schwimmer, CEO, LSEG
  • Ian Bremmer, President and Founder, Eurasia Group and GZERO Media
  • Emily Prince, Group Head of AI, LSEG
  • Corentine Poilvet-Clédière, CEO, LCH SA, LSEG
  • Rishi Sunak, Former Prime Minister of the United Kingdom
  • Radosław Sikorski, Deputy Prime Minister and Foreign Minister of Poland
  • Jake Sullivan, Former US National Security Advisor
  • Mark Esper, Former US Secretary of Defense
  • Lord Parker, Former Director-General, MI5
  • Maggie Fanari, CEO, J. Rothschild Capital Management
  • Mariangela Zappia, President, Italian Institute for International Political Studies
  • Dominic Barton, Strategic Counsellor, Eurasia Group and Chairman, Rio Tinto

Watch our videos

Hear Emily Prince discuss how a customer-centric, end-to-end approach can unlock new ways of building products and accelerate innovation.

AI has the capability to dramatically decreasethat information premium that smaller companiesare paying forbecause you could, through AImake the analysis of thousands of companiesand papers and perspectives and so onmuch more efficient.I'd love to say cheaperI don't think we know that yet, rightbut at least much more efficient.And all of a sudden, the absolute rich diversityof the economic landscape of Europe could be much more visibleand much more attractive.So I think it could compress basically the cost of complexityand Europe should really engage in AI as in analysing market kind of AIto get basically that capital coming their way.

Hear Corentine Poilvet-Clédière discuss how AI can reduce the complexity of analysing smaller companies, making Europe’s diverse opportunities more visible and attracting capital to the continent.

Hear Corentine Poilvet-Clédière discuss how enabling investors to discover, analyse and hold small- and mid-cap companies can help unlock opportunity across European markets.

How much progress Europe makes will be a function of its political will. As it builds scale, resilience and competitiveness. Europe needs to avoid the false choice between resilience and openness. AI has the capability to dramatically decrease that information premium that smaller companies are paying for. I've resisted quite a lot this temptation into is AI about efficiency or is it about value because I think what that speaks to is that there's a divorce between efficiency and values. Given particularly the compute gap between the U.S. and China given where they are, where they stand at the frontier vis-a-vis the United States, has been a very smart strategy. The paradox of the moment we’re in that it is both, I think, the most dangerous moment of our lifetimes but also the most transformational.

Explore highlights from the GZERO Summit, where leaders examined resilience, AI and the forces reshaping Europe and global markets.

AI as a general purpose technology they don't come along that often. We're lucky to be living at a time when we will get the benefits of this and if we get that right I think it'll be transformational not just economically but also for society and really raise the floor for humanity as well.

Rishi Sunak reflects on the transformational potential of AI and the opportunities it could create for economies and society.

From the kind of perch I have. Anthropic and Microsoft in particular. The things that have really struck me are the following. One is just how quickly the technology is spreading. And if you look at how long it took kind of previous big technologies to reach 100 million people, I think that's quite helpful. So it took 70 years for the telephone to reach 100 million people electricity. It took 50 years. Then fast forward a bit the PC, it took 15 years. The internet? It took seven years. Right. ChatGPT two months. Right. So this is clearly something that's happening much faster than we've experienced before. And then the second thing is how quickly the technology itself is improving. And it's hard for us to kind of grapple with exponential growth. I always find we had to get a crash course in it during Covid, but effectively you should think about the capability of these models doubling consistently every six months. Right. That has now been happening for some years. And I have to say, I always thought that would slow down. But if anything it is sped up. And that is, you know, you think about something doubling every now potentially four months, not even six months before you know it. Something quite significant is happening.

Rishi Sunak explores the accelerating pace of AI adoption and advancement, and what sets this technological shift apart.

Don't know if any of you have been around long enough to remember when risk registers were sort of fanciful lists of things that nobody ever quite believed would ever happen but you know you have to have a risk register. These days, they feel more like a menu of things that are actually going to happen but you're not sure in what order and quite what shape they'll hit you but you're pretty sure they're going to. So the resilience is these days is much more about are we, as an organisation or a business, configured to ride these surges and shifts and adapt fast enough to recover fast enough to be back to market exploiting the new opportunities or are we knocked over and drowned? And those questions, I think, everybody is dealing with in businesses reach really deeply to how we structure Are we valuing optionality above optimisation these days which is a hard challenge because it can mean spending money on things you might not need.

Lord Parker discusses how businesses can build resilience, adapt to disruption and recover quickly in an increasingly unpredictable world.

In MI5, the the hour by hour, day by day, week by week core work is making tough calls based on a need to try to pick out which three snowflakes in this blizzard add up to something we need to be concerned about. Some of the hardest decisions were about what not to do. So as I was talking earlier about surging to new and sharp threat spikes and shifts and so on it's a zero sum game because an organisation like MI5 is fully deployed operationally all of the time, there is no sort of reserve force sitting back at barracks waiting for some new work to come along. So surging to something which is a spike means either dialing down something else or stopping it all together for a period and those are tough calls to make because stopping something over here might involve actually, and likely does involve, a higher risk that some harm will happen out of our sight in order to deal with the threat over here. I found those quite tough calls to make and generally it went well but it does have a high risk.

Lord Parker reflects on the challenge of making tough decisions and balancing competing risks when resources are finite.

China has been smarter in a lot of their leadership over the past years than you would have anticipated when you first became National Security Advisor. What do you think that they are most getting right, right now compared to what you might have expected ten years ago? I think now they've shifted to AI and they've said America is the Iron Curtain, we're the open weight model provider and beyond that we will also provide the training and the tools and the financing you need to be able to leverage this new technology in your country on your terms, not on anybody else's. I think that strategy, given particularly the compute gap between the U.S. and China given where they are, where they stand at the frontier vis-à-vis the United States, has been a very smart strategy. And I think more broadly, what China has recognized is that to they recognized two things that I don't think were fully appreciated. I certainly didn't fully appreciate them a few years ago. First, that you can convert scale in frontier industries into innovation. So for a long time, China was largely copying or fast following Western innovation. But they've built such scale and such a massive production innovation loop in in key industries that they are now pacing the innovation, clean energy and battery technology. Is there a great example of this? And that is leveraging process engineering breakthroughs that allow them to push the envelope on lithium ion battery technology. You name it. They're about to do the same thing in biotechnology through sheer scale. They are a getting a huge amount of the American biotech industry to do their testing and, and, and work in China. And then they're converting that scale into now making their own breakthroughs in biotech. I think imperiling the lead of the United States in a critical industry. So that's one thing. And then the second thing they've recognized is the control of upstream. Inputs are vital to national leverage, power, influence, resilience. And this, this, they have made a 20 to 25 year bet on and it has put them in a very strong position to the point where it poses huge dilemmas to American and Western policymakers, even if one resolves the heavy rare earths and magnets issue, you just go down to the next item on the list. Be it pharmaceuticals, be it, you know, certain forms of precision electronics, etcetera, etcetera, etcetera, where they have real choke points up and down the supply chain in ways that put them in a position of real advantage as we saw them exercise that advantage in the trade war, the short lived trade war with the United States. And that's not the way that the Americans looked at the Chinese five, ten years ago. If if that's the direction of travel, should U.S. policy shift from a more decoupling towards managed interdependence, where the Americans are actively trying to have leverage over the Chinese in return? Yeah. So first, before I answer that, I should say I painted a slightly incomplete picture. I do think that China has been smart in some of these critical areas. They also have badly, badly mismanaged their own domestic economy in ways that have created real drag. So she Jinping is basically betting. We talk about a K shaped economy in the United States with, you know, the top few percent doing extremely well and everybody else suffering and struggling. China has its own version of the K shaped economy, which is kind of the real economy where most people live and operates doing terribly. And then the kind of high tech new productive forces economy. And she is hoping that the bottom of the K will muddle through enough that the top of the K can pull. But to be fair, that's not mismanaged. That is an intentional strategy on the part of the Chinese leadership. The top of the K is is the part that is going well. The bottom of the K does reflect, I think, massive mismanagement. And and I don't think they've that I am not convinced that this bet they are making is going to is going to pull them. Fair enough. So we will have to watch and see on that. But but to your question about whether the U.S. should shift to manage interdependence, look, I have never been in favor of decoupling the U.S. and Chinese economies, largely because I think it is completely absurd to think about that ultimate result. It's just not plausible. So we really embraced the concept of de-risking, not decoupling, which Ursula von der Leyen first put forward. And then the G-7 really kind of took as the convergence of the Western approach to dealing with China. So what is de-risking mean? It does mean managing interdependence, and it does mean having our own leverage, which we do have in a number of key areas, and trying to build up more forms of leverage. But it also does mean looking for those pain points where your dependance is so great, and the ability for China to make that dependance bite is so fast that you're just exposed. Have you rare earths as an example of that? It's true that if we resolved that there's this long list of other issues, as I just said, but some of those other issues don't have the kind of, you know, if they cut it off for three months, the economy shuts down quality to them. So I do think we need to identify that set of vulnerabilities and dependencies and say that's intolerable. We can't live like that just as China doesn't want to live like that. But more broadly, I do agree that we, on the one hand, need to make sure that we have cards to play, we have leverage on our side and on the other hand, maintain intense diplomatic interaction with China to keep this thing from spiraling into a kind of tit for tat trade and economic war or technology war that gets out of control. So I'm not trying to straw man. I guess I'm saying over the past years, the momentum has been towards de-risking as an intentional strategy and a bunch of places where concerned that you've got national security vulnerabilities. On balance, would you say that trajectory needs to shift towards more strategic interdependence to ensure that there's leverage on both sides? I don't know that it needs to. I, in fact, think we need to accelerate our efforts in certain key areas to risk, to de-risk. So no shift, no letting the foot off the gas go faster, do more. And frankly, I think the entire set of industrialized democracies need to do that, because the alternative is going to be an outcome in which the whole world is dependent on China, and China is dependent on no one. But at the same time, I also agree with you that there has not been sufficient emphasis on those areas where collectively the West or whatever you want to call it, has built up its own capabilities in its own leverage in ways that create some disincentive for China to play these cards. Now, these two things are very related, of course, right there. They're connected. So I guess what I would say is no less emphasis on the de-risking prong, but more emphasis than we have seen on, I guess, what you would call managed interdependence. Great. Now that brings us to AI, where the Americans have both feet on the accelerator. There's a big question about whether the Chinese have approached appropriately would ever consider cooperating on pacing the frontier. Such an approach has not yet happened. Should such an approach happen and absent governance. Where do you think the geopolitics of AI are likely to head in the next months? Even so, look, my my view on this is really quite simple. And you and I have talked about this a lot. Well before hugging face, which I do think really caught the world's attention. You had hundreds of engineers from these companies putting out a public letter saying this technology is dangerous. It is moving too quickly. The gap between safety and capabilities is too great. We need to slow down. They use the word pacing because it's like a nicer word than slow down. But pacing means slow down usually. Yeah, yeah. So I mean, in this context, it certainly does. Pacing implies a, a a run that you can sustainably continue. Right. So it is it's artful from that perspective. It's artful. It is artful. But at the end of the day, what are they saying? You talk to them behind closed doors. We need time. We need to put time on the clock because we need to use that time then to figure out these alignment problems, which, by the way, this isn't as simple as just build some more guardrails into a model. The engineers literally don't know what exactly the combination of steps that are required to ensure that a given model is sufficiently aligned that it can be safe for release. They don't know that yet. They're still trying to work their way through that kind of science. Things like mechanistic interpretability like monitor ability of the chain of thought, reasoning of the models, these sorts of things. So they're saying, hey, we need because there are race dynamics both among the American labs and then between America and China, we need government to come in and say, we're all going to find a way to pace, whether that's the amount of compute used or how you use AI to train further AI or whatever the case may be. What the Trump administration is basically said is we're not interested in that, okay? This is a race with China. We have kind of tools to deal with it. If something goes wrong, like we're not going to regulate. And when China hears that first when they hear the frontier, they think, Slow China down. That's is all a trap to try to hold them back, which, amusingly, is not what these companies is saying at all, is not. Yeah. But from China's perspective, when you are behind, even narrowly behind, narrowly but meaningfully behind, and you hear the leader saying, okay, everyone's got to slow down now, it's not unusual for you to think, hey, they're just trying to lock in their lead. They want a moat, right? Yeah. And that is basically China's view. The problem is that from America's perspective, if a pacing agreement is okay, the U.S. and China get to parity and then we pace together. Well, that's the U.S. giving something up. But from China's perspective, if we pace with the US persistently ahead, that's China accepting a permanent second class status. Yeah. So this creates a huge structural problem between the two sides in terms of coming to an agreement on this. And do you think the Chinese share the sense of concern, of vulnerabilities, of lack, of being able to understand the damage these models can do, and how are they approaching it, given that they're completely unwilling to work with the United States? So first, I think they believe the US is ahead. So this is more the U.S. problem than it is ours. Second, they think our ability to control. Misalignment or misuse of AI systems is greater than the Americans ability because of just our ability to manage the internet in the country. I do to to a certain extent. I mean, I think the US internet is a much more Wild West situation than theirs. Do I think so? It's better than ours. Do I think it's good enough to handle the kinds of risks that are coming? I absolutely don't. But their mindset is a bit this is more America's problem than ours at the moment. But shortly after Dario Ahmadi came out with his Pacing the Frontier essay, there were two statements, almost like a split screen out of Beijing. One from the Foreign Ministry was basically dismissive of Dario. This is all the Americans talking a bunch of nonsense. You know, they essentially want to control and dominate and regulate and choke AI kind of classic party rhetoric. The second was an unusual essay written by the head of the Ministry of State Security. Basically, the number the number two spy chief in China, the headline of which basically was this technology is gaining capability to the point where it threatens party control in China. And therefore we need to act to make sure that does not happen. Now, when China talks about loss of control, they're much more focused on loss of party control over over the country. When the US talks about loss of control, it's much more human control over machines. But it kind of drives towards a similar place, which is in an elevated level of concern among people watching this close to the technology in both countries, saying we have a huge risk here, that risk has to be managed. And yet I would say in Washington, that doesn't translate into a view that government needs to do it. It translates into a view, okay, there's risk. The company should be taking care of that risk. You want to pace the frontier, go ahead and paste the frontier. You do it. Whereas in China, I think there's probably a greater willingness of the government to go to the labs and say, you're going to work with us and we're going to figure this. So if you were back the National Security Advisor right now and you were coordinating the next summit between the president and chairman, she would your top priority for what we need to engage on be this very issue. And secondarily, how would you engage the private sector in the United States and or China? So so my view on this, again, it just may be too simple. I may be missing something, but the vibrating intensity, not just of the CEOs who have their own reasons to do what they do, although I don't fully by the cynical take that they're just saying these things for IPOs or whatever. The actual people building this technology are saying this is highly risky, very dangerous. You can set aside the extinction issues, just the disruption and the ways in which this could flow through our economy and society and deeply negative ways in the very near term, like now, not years from now, but now, I would say the United States, even without an agreement with China, should be taking steps to impose a degree of slowdown, to buy some time to figure out the best way to regulate this unilaterally. We have a lead over China. That lead is somewhere on the order of seven months, but it's really probably larger than that because much of the way China keeps paces through distillation. So if we slow down, there will be some impact, at least some on slowing them down as well. I would use that lead right now for safety, and then I would go to China and the rest of the world, and I would say we've acted. So this is not just a matter of us coming to you and saying, you do it and then we'll do it. We've acted. Now you've got to act now. China then has two choices to do something cooperated defect or defect. Yeah. And then if they defect, then we are on the horns of a huge dilemma. You know, Ted Cruz, senator from the United States was posed the question like, if this is truly a dangerous technology, why does it really matter who builds it? Because if it's going to harm us all. And he said, I'd rather be killed by an American killer robot than a Chinese killer robot. He said that I appreciate the honesty. He said that it works as a civilizational right. I do not find that very persuasive. If I'm being killed by a killer robot, I don't really care who it is. So we would be on the horns of a dilemma then. But I think our best chance to get to a bilateral agreement with China has two features. One, America moves first and two probably. It's not some formal treaty a la arms control. Probably it is more of a set of understandings that involve parallel actions in both places that cannot be tightly verified, but can be generally monitored. And by the way, I'll add one more thing. It is uniquely unfortunate this is happening right now. Yes, yes. Add one more thing. Okay. My one more thing is for those in the US who say, we can't slow down because of China and if we slow down, it will give China an advantage. There's one aspect of a slowdown right now, or even a pause on further training of advanced systems that actually could work to America's advantage in the competition, not just with China, but generally. And that is, our labs right now have relatively lax cybersecurity compared to where it should be, given the gravity of this technology. So one way of putting this would be to say, a pause on on further training runs until such time as you have reached a cybersecurity level that we think is sufficient, and that will help your China bona fide as well, because I don't want to be Jake running for president on the we're going to slow down unilateral platform. Right. Like that's not a that's not a good platform. So well first Ian, you know it's interesting the politics on China and the US are changing to a certain extent. And I think on this question, the level of concern among the American public is I is extremely high. I'm not actually sure the like your week on China answer to the hey, we got to slow this thing down will resonate in May. I think this is an open question, but yes, I would like to have an argument that's rooted in security and competitiveness. And my argument would be we do the pause for some number of months. We use that time to build our defenses so that it's not worth it to have a lead in AI. If at the last minute, a foreign competitor can come in and steal the model weights right as you're getting to the finish line and they've just caught up. So let's fix that problem. And oh by the way, that also happens to bias the time we need to put a better safety framework in place. And oh by the way, it also allows us to go to China and test whether this is possible without, in my view, closing off the ultimate decision of a president a few months down the line to say, okay, China's not coming along. If we want to go back to the race, we can always go back to the race. So we've worked out the beginnings of at least a very short term policy between the two of us here. Tell the audience, please, that your P doom number is lower than ten. Just tell them that it is lower than it is lower than ten. We knew that. I thought we'd end with something optimistic. Everyone, please join me in thanking the great Jake Sullivan. Glenn. So, Ian Bremmer and Jake Sullivan, thank you so much, ladies and gentlemen. Lunch is now served in the foyer. We'll see you back here in about an hour as the G0 summit continues.

Jake Sullivan explores how China’s approach to AI could shape the evolving global technology landscape.