Ben Bland
Welcome, everyone, to this discussion on the state of China’s economy. My name is Ben Bland. I’m the Director of the Asia-Pacific Programme here at Chatham House. As I’m sure many of you in the room and online will know, China’s investment in export and le – model of economic growth has been central to its global rise. However, it’s been clear in recent years that it’s facing growing systemic challenges, population growth has peaked, as diminishing rates of return on investments and prolonged disruptions to the domestic property market have undermined local government finances, household spending and broader domestic demand, sparking a deflationary spiral. And that’s before we even start talking about US-China tensions and the impact of the wars in the Middle East and the fragmenting global order.
At the National People’s Congress in March, China’s rubberstamped parliament, the Chinese Communist Party launched its latest Five-Year Plan, with a focus on the AI+ initiative. So, Xi Jinping is really pinning his hopes on fusing artificial intelligence with China’s manufacturing capacity to turn the nation into what he would call a “modernised socialist state” by 2035. And that’s all while pursuing a broader shift towards self-reliance across the board.
So, can China pull off this transition? Can the Chinese economic miracle find a new head of steam, or is it running out? To discuss this question, I’m really glad to be joined by a great panel of internal Chatham House experts. We have a lot of great external speakers, but it’s really nice when we have our own colleagues on show. To my left, Dr Yu Jie is a Senior Research Fellow on China in the Asia-Pacific Programme, as is James Kynge, sitting next to her. And then, we have David Lubin, who’s the Michael Klein Senior Research Fellow in the Global Economy and Finance Programme. All three have recently been in China, so I’m looking forward, as I’m sure you are, to hearing some of their personal perspectives, as well as their analytical takes.
Before we start, some housekeeping. A reminder that this event is on the record and it is being recorded. I will ask the questions for about 35 or 40 minutes and then, I’ll open it up to you in the room and online. If you’re on the Zoom, please ask your questions using the Q&A function. Yu Jie, I want to come to you first. You were recently in Beijing. I mean, how are people feeling on the streets about the state of China’s economy?
Dr Yu Jie
Thank you, Ben. I’m delighted to be here to join my colleagues and to talk about Five-Years Plan, you know, beyond just what happened in the Middle East. How they feel? I think some mixed feelings. Let’s put in this way. On the one hand, yes, the economy is clearly slowing down, and people fear, especially among young people, they could not find appropriate graduate job that they desire. I’m not talking about just a job but a job they wanted to have. But then, on the other hand, I think look what’s going on in this world. The Chinese populations also felt, well, at least it is a government intervention in terms of controlling the food price, in terms of controlling the consumer price. So, basic standard of living still being able to maintained.
So, I think it’s that mixed feeling, on the one hand, yes, you feel that the economy is getting much slower, and as indicated on this Five-Years Plan, in the next five years, the Chinese economy will grow at around 4.5%. But worse, on the other hand, I think people naturally felt the sense of economic resilience is much needed in a world that is largely chaotic.
Ben Bland
And before – yeah, before we dig into some of the deeper questions, James, what were your vibes on your trip? I think you were…
James Kynge
So…
Ben Bland
…travelling all around China.
James Kynge
Yeah, I was lucky enough to go to Beijing, Tianjin, Shenzhen and Hainan Island, and I was there for a couple of weeks. Just got back last week. My impressions were actually, very much polarised and both extreme, if that can be understood. So, on the one hand, I was truly amazed, I would say almost stunned, at the speed of China’s advance in hi-tech. This is something that I’ve been researching for quite some time now and I’ve always been concerned that I’ve been overhyping China’s tech rise. But I – and these last two weeks have taught me that I’m probably lowballing China’s tech rise and it is even more amazing than the most amazing predictions out there. So, that was one polarised opinion.
The other was that I was also amazed at how many people in China are concerned about the rise of AI and as Yu Jie just mentioned, about the impact that AI is already having on youth employment opportunities. And I can go into that in much greater detail, but I feel that this year could easily see a crunch point in terms of employment opportunities for 18-24-year-olds, that’s a key metric, perhaps in the month of August, when all the graduates have graduated and they’re trying to find jobs.
I don’t want to be too predictive. I mean, it’s hard to be predictive about China. It’s such a complicated place, but I, sort of, feel like the rise of AI and the employment impact it could have could have quite a lot of knock-on effects in terms of policy. There could even be a backlash against AI. It’s going to be interesting to watch, but it’s a big topic.
Ben Bland
We’ll come into some of that detail in a bit. For now, David, what was your mood music when you were most recently in China?
David Lubin
Well, I wasn’t there two weeks, and I was there last September, so I shouldn’t rely on my trip so much. But just to reinforce what James is saying, when you do a, kind of, wordcount of the Five-Year Plan documents, the phrases most frequently referred to are “artificial intelligence, high quality development, scientific innovation” and “national security.” And I think that gives you a, kind of, reasonable framework for thinking about how Chinese policymakers are operating these days.
There’s – you know, there’s this endless tension in debates about Chinese economic policy between what they say and what they do. We’ve had 20 years of rhetoric about rebalancing the economy, in other words shifting the focus of the economy away from exports and investment, towards a focus on consumerism and domestic spending. In the – just to put it into context, 20 years ago, in 2005, private sector consumption as a share of GDP was 40%. In 2025, private sector consumption as a share of GDP was 40%. So, in two decades of discussion about rebalancing the economy, there is zero evidence of rebalancing.
And I think, you know, the priorities that you can see from the Five-Year Plan reinforce, I think, the idea that the chances of rebalancing remain very slight. That Chinese economic policy remains a, kind of, derivative or a function of the way Xi Jinping sees China’s geopolitical sensitivities. So, economic policy is a, kind of, derivative or a subset of geopolitics and that means that, you know, prioritising net exports, prioritising innovation, prioritising manufacturing and prioritising the capture of global market share remain the central aspects of economic policy.
And although we see constant references and constant debates about the need for rebalancing, the most recent was last week or ten days ago in an article by a prominent Chinese Economist in Qiushi, the Chinese Communist Party’s theoretical journal, suggesting that, you know, endless reliance on exports is just a dead end. Yes, okay, yes, it’s another example of a Chinese Economist saying that rebalancing is important and it’s another big Chinese Economist that is, basically, going to be ignored by Chinese policymakers.
Dr Yu Jie
Hmmm.
Ben Bland
Okay, and, I mean, James, and what is the AI+ initiative and is it different, or is it just, sort of, old wine in a new bottle, as I think David is suggesting?
James Kynge
It is very different, I think, and I’d like to hark back to the title of this event, which is ‘Same engine,’ different ‘fuel’ for China. I would actually go so far as to say that the fuel is so different, it may create an entirely new engine. I think we are literally at such a historic pivot point in terms of China’s development, and it all comes down to different technologies, not only AI, but AI is going to animate all technologies in China. The official plan of AI+ mentioned in the 15th Five-Year Plan is that AI should be animating 90% of the six different sectors of the economy that are mentioned in the plan. So, basically, AI is going to be everywhere.
But the thing you notice when you get to China is that China’s already there. It’s already where it should be in 2030 in many areas, I mean, not in all areas. So, at the moment, agentic AI is already being deployed in 67% of industrial enterprises in China.
Ben Bland
And what do you mean by agentic AI?
James Kynge
I’m going to come to that.
Ben Bland
Okay.
James Kynge
Because this is another really important pivot point, a really important shift. What we’ve all been focused on, I think, until now, generally has been LLMs, large language models. So, basically, chatbots, OpenAIs, ChatGPT, etc. So, you can ask them a question, and they’ll give you an answer. What we’re shifting to now, and particularly in China with the advent of OpenClaw, etc., are these agentic AI tools, which instead of just replying to your research query, will actually execute a task for you. So, they’ll manage your calendar, they’ll write your code, they’ll book your holiday, and not just your ticket, but they’ll book every single ticket you need and every bus fare, you know, for your whole holiday within seconds. So – but they’ll also do a whole suite of industrial process tasks and all of that.
So, we’ve now shifted from an LLM-centric AI universe to an agentic AI universe, and this is why OpenClaw, everybody is talking about OpenClaw. This was invented by actually an Austrian Scientist but is now all the rage in China. So much so that the government is getting very concerned about OpenClaw. But the other thing that’s happened is that already, Chinese AI companies, like Alibaba and Drupal AI, have launched their own agentic AI competitors to OpenClaw, and Tencent will very soon, and other AI companies will as well.
So, we’re now entering the agentic AI era in China. This has two big impacts to me. One is it’ll make doing anything, really, much more efficient and the other is it will make getting a job, if you’re a human being, much more difficult, because these things do tasks that human beings currently do. So, we really are – as I said at the beginning, I think that the fuel is so different, and when I say fuel I mean AI, that I think we’re changing the engine. We’re changing the nature of the Chinese economic model with this technology. I wouldn’t – I know that sounds hyperbolic, but I really think that is the case.
Ben Bland
And Yu Jie, do you think there’s something different about, you know, Chinese consumers or Chinese businesses that means they’re more open to using these tools, when maybe, you know, here people would have more concerns about privacy or security? Is there something about Chinese culture of business or Chinese people that means they’re, sort of, more open to just trying things that could be transformative…
Dr Yu Jie
Yeah.
Ben Bland
…could potentially be very negative, as well?
Dr Yu Jie
I mean, really, to start with, the societal acceptance towards new stuff and innovation within China and perhaps much higher. You know, this is really driven back to – from – since 1950s, what is the best subject to study? Physics, science, chemistry, because these will land you a job anywhere. So, when it come to a new product or new technology, something society to be excited about. Like, for example, my mum was trying to explain to me how OpenClaw works. So, imagine a 70-year-old pensioner explain to a 30-years-old something how the AI tools works in China, and that’s the story. That’s the level of societal acceptance in terms of the attitudes towards AI.
And also, in general, I think the government presented a very positive story to showcase that how much AI can make your life more convenient, can get your business running more efficiently. So, that sense of efficiency and the positive advantage in AI within China has been well presented through the government official propaganda and various other of China’s on social media and so on and so forth. And so far, as discussed, is how much is privacy intrusion and how much data that the government’s able to collect towards an individual is less discussed? So, I think that’s where we are.
And also, judging from – if we come back to the Five-Years Plan, there’s 18 chapters, before I come here, I counted, and four chapters is about algorithm about digital economy, about AI and also about how to using this to empower the economy in general. Now, there’s one specific sec – chapter and mentioned on education. How the education mode of China, especially the particular syllabus we have now on the secondary and tertiary education, to be able to adapt to the AI-led world. So, in a way, the policymaker already thinking this idea perhaps because of this fast, rapid development of AI, will led into massive job losses. Hence, the syllabus will require to change in order to have the labour forces that will fit into the purpose for the future industries that Beijing wanted.
Ben Bland
And I hope you’re all paying attention ‘cause you’re all going to have to complete a quiz on the Five-Year Plan…
Dr Yu Jie
Yes.
Ben Bland
…before you’re allowed to leave.
Dr Yu Jie
After and before…
Ben Bland
In the either the physical or the…
Dr Yu Jie
…the exam questions.
Ben Bland
…Zoom room. David, I mean, do you have some scepticism here about the idea that the AI+ initiative is something different? I mean, from what you’re saying you’re questioning whether that can lead to a genuine change in the structure of the Chinese economy.
David Lubin
No, I don’t think, and I’m not questioning it. I think what James is saying is fascinating and I think it’s consistent with the emphasis on innovation, with the emphasis on manufacturing dominance. You know, this idea of creating a manufacturing powerhouse is – and just absolutely central to China’s ambitions. And to, kind of, put that into context, I just want to tell a little story, maybe inappropriate for an on the record briefing, but I had a meeting in April 2025 in Washington with Steve Miran, who was then the Chairman of the Council of Economic Advisers of Donald Trump, is now on the Board of the Federal Reserve. And as a, kind of, devil’s advocate I, sort of, said to him, you know, “Why are you guys so obsessed with reinvigorating US manufacturing?” You know, “What’s wrong with having a great services economy?” And he said, “You can’t run an economy, you can’t – you know, the United States should not have an economy that is based on the ability of Google to sell adverts.”
In other words, he was saying that, you know – oh, no, sorry, I’m sorry, he said, “You cannot win a war with an economy that is based on Google selling adverts.” In other words, that there – you know, and that comment to me highlighted the essential connection that a big important economy makes between manufac – you know, man – the importance of manufacturing and national security. In other words, that, you know, the manufacturing is, kind of, a core component of national security. And to the extent that AI, you know, great for productivity – you know, it may be bad for humanity, but great for productivity, you know, to the extent that AI helps China to double-down on its dominance of global manufacturing, then I think it’s a, kind of, non-negotiable part of the strategy.
Ben Bland
And James, can you paint a bit of a picture of what the AI+ is starting to look like in practice from your travels across China? I mean, what does this mean in terms of, you know, products and services and shaping the economy in daily life?
James Kynge
I mean, I think it just means that AI is everywhere, so I mean, if you’re talking about manufacturing, it just means that the production lines are all animated by AI. I mean, the Exhibit A example that I always give is the Xiaomi car factory in Beijing. You may have seen some of this on YouTube, it’s absolutely incredible, and I would certainly welcome…
Ben Bland
You won’t have seen it on YouTube if you’re in China, but yeah.
James Kynge
Well, in China, you wouldn’t see it…
Ben Bland
And yeah.
James Kynge
…on YouTube, that’s true, but there are no people on the production line. Of course, that’s robotics, it’s robotics plus AI, and I think this is one feature of the Chinese universe that I mean, perhaps isn’t emphasised enough. China had a robotics revolution, it had – you know, so robot arms have been in Chinese factories now for at least a decade, and the adoption of these robot arms has been off the charts. But now you’ve got robots connected with AI so that the entire factory is run by AI.
There is also a pilot hospital in Beijing, which I haven’t been to, and I do have some scepticism over, but apparently this hospital is entirely humanless. So, it’s all AI. I went to a – but I can see somebody expressing surprise. I’m also surprised. I would like to go there and check it out, but I’ve read this. But I did go to Xiong, which is in the southwest of Beijing, and they have an AI, kind of, exhibition centre there, it’s been open for a couple of months and it – you know, this is, again, AI plus robots.
So, these are a lot of humanoid robots, and what you notice, when you go through the door, is that China’s going from speciation of robots. So, we’re not going for a general robot that can do everything. We’re going for specifically tooled robots to do specific tasks. And to get back to the hospital example, one of the robots I saw was conducting a cruciate – is this your cruciate ligament, the one on your knee? Anyway, it was conducting an operation on a prosthetic knee in this exhibition centre. So, I don’t think it’s too pie in the sky to think that robots are going to be doing operations that are currently done by Surgeons.
There was also a robot that was kicking penalties into a goal and what I was really delighted to see was that of course, it scored nine times out of ten, but on the tenth it missed the goal and then it went, “Aww.” So, I thought it’s nice to know that, you know, robots also feel my kind of pain. But there was a robot shooting hoops in – at basketball and then there was robot Shop Assistants who were correctly choosing the right medicine to – on a shelf of medicines. So, you’d walk up to them and you’d say, “I would like some melatonin” or something, and I think I had jetlag, and it would choose the melatonin and give it to you.
And then, lastly, there was a robot that was like an all-knowing robot you would ask questions to, and it was in – it was dressed up like Albert Einstein, with the hairstyle, and you could ask it any question you like. And it was – well, I don’t know as much as Einstein, I never pretended I did, but he seemed to be telling me what – the plausible answers to all the questions that I had. So, it’s everywhere.
Ben Bland
Hmmm.
James Kynge
It’s doing, you know – well, put it this way, let’s calm down a little bit, it has the potential, it seems to me, to do almost everything.
Ben Bland
Hmmm, and Yu Jie, I know you’ve been writing about…
Dr Yu Jie
Hmmm.
Ben Bland
…youth unemployment in China for a while. I mean, how is this affecting people’s mindsets? Are people in China worried about implications for jobs in the future?
Dr Yu Jie
I think there are two kinds in here. We’re talking about two kinds of job we’re talking about in here. The jobs that most of the youngsters want to do, office workers, typical middleclass white-collar jobs, that is – was largely available because of the property booms in China a coup – back to a couple of years ago, because with the internet company, that boom in China a couple of years ago, and those jobs will be gone. But instead, and what now we have is the manufacturing job shortage within China and especially in the southern part of the country and that is required to have a highly skilled labour force. And that would require to have well trained on large language model by oneself, by the students, and these are the jobs that is actually – have a shortage right for now.
So, I see some kind of skillset mismatch between the education system, I said earlier, and vis-à-vis the jobs that employers required. So, I think we have not really come to that round yet until there’s new syllabus in place and that would enable the Chinese youngsters to learn something completely new. But having said that, and I think particular social and family structure within China also tell a very different story. Yes, people worry about it, but this now, for those who aged between 20 to 30 and who are the key youth workforces, and what is the age of the parents? The age of parents is around 60 to 70, the generation that perhaps enjoy the most generous pension of the entire generations that China ever had.
So, that would actually allow those parents, and they would rather pay off for their children to stay at home and – because they are the only child at home, and rather than pay them – rather than let them to work, suffering from, for example, mental health issues or work in the factory, and they would rather pay off their children to stay at home. So, it’s really come back to the Bank of Mum and Dad, like similarly in Europe. So, I think that’s the similar stories what I have seen so far.
Ben Bland
The children would probably rather the robots are looking after their parents, rather than them having to do it themselves.
Dr Yu Jie
And then they can go on videogames…
Ben Bland
Yeah.
Dr Yu Jie
…as well.
David Lubin
And I would say when I was…
Dr Yu Jie
Yeah.
David Lubin
…in Beijing in September, James’s former colleague at the financial Times, Eleanor Olcott, who’s the Financial Times Technology Correspondent in China, shared with me some fascinating data showing the absolute collapse in salaries of graduates from the elite universities, you know, from Tsinghua, Beijing’s University. I can’t remember the numbers, but maybe it was like in 2019 if, you know, if the graduating – average graduating salary from – for Tsinghua was $35,000 by 2023, before AI started to kick in, it had fallen to something close to $20,000.
So, there’s this extraordinary crisis for young people, which, you know, is taking place against the background in which, since the 2020 collapse in the real estate market, left household confidence on the floor. You’re now adding to that extremely low level of household confidence, that extremely strong sense of just despair, it’s – if that’s not too strong a word to use, you’re, kind of, adding to that sense of despair, a, kind of, new reason for desperation.
Ben Bland
Hmmm, and David, how do you think this all affects the, sort of, long-term growth trajectory and economic trajectory for China? Can, you know, the boosts to productivity from AI make up for these other long-term structural challenges that we’ve spoken about?
David Lubin
Well, I guess, you know, on the one hand, if you can argue, as I think James would, and I agree, that, you know, AI adoption in China’s going to be much more rapid than it is in many of China’s trading partners, that boost of Chinese productivity should leave the economy in good stead. On the other hand, kind of, offsetting that, is this huge demographic, kind of, headwind for the Chinese economy. How those two things balance out is, I think, difficult to say. But, you know, given that China, in my view, seems to be as emphatic about its current economic model, the, sort of, export and investment and manufacturing led model, as it has been at any time of re – over recent decades, it’s going to be – you know, any technology, like AI, that supports China’s pursuit of manufacturing dominance is going to leave the economy in pretty good stead.
The – you know, the question for the rest of the world is given the fact that China seems to have a, kind of, non-negotiable attitude towards its basic economic model, what should the rest of the world do faced with this common frustration about the, kind of, penetration of Chinese exports to the rest of the world? I mean, in a way, you could argue that the only thing that the US, the EU and the Global South can agree on, the only thing, is that China should import more and export less.
Ben Bland
Hmmm hmm.
David Lubin
And so – but, you know, facing – in spite of that very strong consensus, I think it’s, nevertheless, going to be the – nevertheless going to remaining – remain true that China’s view of its economic model is non-negotiable. So, what might nee – what might be negotiable is the – is exchange rate policy, and I think the only thing that, you know, that China’s trading partners may have a chance of convincing China to do is to allow the ex – is to allow the renminbi to strengthen in a more accelerated way. Whether that will do anything to change global imbalances, whether it will do anything to dent China’s increase in global market share, is another issue, but I think that’s the mo – the only practical response that the rest of the world has to deal with the, sort of, manufacturing threat that China continues to face…
Ben Bland
And obviously…
David Lubin
…today.
Ben Bland
…Xi Jinping’s creating economic policy primarily for, you know, the citizens of China. I mean, despite the advice of the Economists, including the Chinese Economists, I mean, do you think he made the right call, then, to stick to self-reliance and manufacturing, given the way the world has gone? Is he being proven right?
David Lubin
I guess it depends on what – you know, what criteria you’re judging him on. I thi – I mean, Yu Jie is probably better placed than any of us to talk about this, but my sense is that, you know, you’ve – on the one hand, you’ve got Xi Jinping wanting to, kind of, promote national rejuvenation. You’ve got a bunch of Chinese households that are a bit grumpy about the fact that they’re a, kind of, residual in that calculation. And in a way, I think one of the – you know, from Xi – I mean, if I put thoughts into Xi Jinping’s head, I think what he would like is for Chinese households to feel a greater sense of commitment to this project of national rejuvenation. That, you know, he, himself, you know, constantly emphasises his commitment to frugality, his willingness to sacrifice. I think he would like Chinese households to feel a greater sense of sacrifice, a greater sense of willingness to, kind of, subsume their own individual, personal short-term welfare into the greater good of the Chinese – the project of Chinese national renewal.
Ben Bland
And yeah, Yu Jie, do you think that Chinese people are ready to eat bitterness…
David Lubin
Yeah.
Ben Bland
…served up by Xi Jinping?
David Lubin
Hmmm.
Dr Yu Jie
Well, again, it’s a generation divide in here. I mean, if you’s – if you look into those who are in power, these are people between the age 55 to 65 and most of the senior policymakers. And of course, they have gone through the time of cultural revolution, they have gone through the time of Great Leap Forward, and hence, that sense of ETO business – ETO businesses come to very natural to them. But if you speak for those who were born in the 80s and 90s and early twen – early 2000, or, like one of us, the only story we know is the booming period of the Chinese economy. You know, that non-stoppable economic growth and all these windows of opportunity are available to us.
Now, if those windows of opportunity become less or that window may close, and what be, kind of, political discourses that the party will have to justify to its own population? So, that’s question number one. However, I think what happening and what’s going on in this world, especially given what Donald Trump has done in the past three months or four months or so, and it also make the Chinese population wonder perhaps there’s a pursuit of economics resilience, it may not be a bad idea. Just look at – give a very simple example of oil price and petrol price. Now, across the world, for now, the average increase of petrol price since the beginning of the war in the Middle East is around 25% and whereas for China, the increase is only around 10%. So, it’s relatively smaller compared with the rest of the world.
So, I mean, really, in the short to medium-term, the Chinese population may felt, right, this decision of eating bitternesses and also resolve for future, perhaps might be a right course to do, might be the right thing to do. So, I think it’s one way to another, but surely what happened in the Middle East has offered President Xi and also his team of the Politburo members, on conviction, the so-called economic self-reliance is the way forward. And hence, the economic security question need to be answered instead of the question of prosperity generation.
Ben Bland
And I think the…
Dr Yu Jie
Hmmm.
Ben Bland
…Chinese Premier, Li Qiang, told Western – mostly Western CEOs, when they went to China last month for the Development Forum, “China is a harbour of stability compared to the rest of the world.” I mean, is that right, David? You’re looking at markets in comparison. Whatever you say about China, is it actually looking pretty attractive compared to the rest of us?
David Lubin
Well, right now in the context of the Iran war, I think that’s right. China – well, in the first place, Iranian oil is still flowing to China. China is still getting access to something like 1.2 million barrels a day of Iranian oil. Electricity generation in China depends almost not at all on oil and gas. Electricity generation in China is overwhelmingly dependent on coal, on hydro, on solar and wind and on nuclear.
Dr Yu Jie
And nuclear, yeah.
David Lubin
And so, you know, yeah, you know, if factories can keep churning quite happily in the current context. I would also say that to the extent that there is an inflationary shock for China resulting from higher energy prices, well, China has a problem with deflation these days. So, if higher energy prices create an inflationary shock, China’s much better able to weather that storm than, you know, countries further to the West.
The main vulnerability for the Chinese economy, in the context of this war, is if there’s a serious negative shock to global economic activity.
Dr Yu Jie
Hmmm hmm, yeah.
David Lubin
In other words, to the demand conditions in China’s trading partners, and that would negatively affect Chinese export growth, on which Chinese GDP growth has depended very significantly in the last couple of years. So, that’s the vulnerability. But broadly speaking, I mean, nobody is a winner from this…
Dr Yu Jie
Hmmm hmm.
David Lubin
…except right now, possibly Russia, but nobody’s a winner from this. But China has – and from an economics point of view, China has reasons to feel relatively comfortable right now, I would argue.
Dr Yu Jie
I think the weaknesses and vulnerabilities also come from if we come back to this economic model, you know, the pursuit of high-end manufacturing to make China as the hub of global high-end manufacturing. But by doing high-end manufacturing, that would require to have a very resilient global supply chain and those raw materials, including rare earth and including the elements that need to producing, for example…
David Lubin
Yeah.
Dr Yu Jie
…semiconductors and shippings, and so on and so forth. And that would remain – and that would require to have a perfect global connectivity and that is the condition that China is living in. So, if we’re reliant on those so-called global connectivity, and given what’s going on in the world, and perhaps this is a vulnerable for China.
David Lubin
Hmmm.
Dr Yu Jie
Now, secondly, we’re talking about by being a export orinated [means oriented] economy of China, many that would be depending on the global shipping line is largely running smoothly. But 90% of the China’s export and import, 90% are depending on the sealine. So, a open sealine is absolutely essential for China, as well. So, I think these are the two vulnerabilities over the medium-term, even though it looks like China is now, it seems to…
David Lubin
Yeah.
Dr Yu Jie
…on the winning side of this war. But I think the longer-term vulnerability on the connectivity, that is quite pertinent.
David Lubin
I absolutely agree with that. Can I just make one point, though? That – and I’ve said it on this stage.
Dr Yu Jie
We’re not supposed to agree with each other, though.
Ben Bland
Hmmm, yeah.
David Lubin
Yeah. Yeah, I’ve said on this stage in the past, with – in the past with Ben, you know, I’ve used this phrase ‘asymmetric decoupling’, which for me, is a way of just characterising China’s broad economic strategy. And what I mean by that is the simultaneous effort to reduce China’s dependence on imports from the rest of the world through import substitution and domesticating supply chains in order – precisely in order to reduce the vulnerability that Yu Jie was highlighting. While at the same time, trying to maximise or enhance, or increase, the rest of the world’s dependence on China. And I think that if that’s a way of characterising China’s broad economic strategy, they’ve been extremely successful.
James Kynge
And I would argue that that is the China paradox, what you’ve just described. A paradox because it could result in China’s emphasis on self-reliance in the way that we’ve all been describing as actually endangering China’s stability and security in the world, because the rest of the world just can’t stand it. If China’s trade surplus is now US$1.2 trillion, as it was last year, the biggest peacetime trade surplus ever, and this year it’s obviously, well, so far, going much higher than that, there is a really valid question as to how long the countries of Europe, I won’t mention the US, because it’s clear that the US already has tensions with China, but how long countries in Europe and the rest of the world will put up with this enormous growing Chinese trade surplus and the erosion, the very rapid erosion of their industrial base.
So, I think what you’re describing is the China paradox we’re going to be dealing with probably for another decade.
David Lubin
But can I add? There’s a paradox on the other side, as well, for China’s trading partners, which is that on the one hand, you know, you see a very sharp increase in hostile trade policies directed at China coming from developing countries, coming from everywhere, but – and that suggests that – and that is partly, you know, what was informing my comments earlier about the rest of the world’s desire to see China do something to reduce the trade surplus through exchange rate re-evaluation.
James Kynge
Hmmm.
David Lubin
But one of the consequences for China’s trading partners of a stronger renminbi is all of the stuff that they enjoy from China…
Dr Yu Jie
Hmmm.
David Lubin
…becomes more expensive. And if, for example, one of the longer-term consequences of this war is to accelerate the desire of many governments to the – to accelerate the green transition for many governments, well, if buying Chinese stuff is going to become more expensive as the Chinese exchange rate strengthens, well, that’s – that, kind of, goes against what they’d like to be doing.
Ben Bland
Well, a lot of people talk about economic security now, or climate security, but very few ask the question to citizens, what price are you willing to pay…
David Lubin
Exactly, yeah, yeah.
Ben Bland
…for that security? I’m going to come to your questions in a second, but I’ve just got one more for James, which is about political control. I mean, obviously, Xi Jinping, since he came to power as General-Secretary of the party, he’s been very focused on intensifying his control, and it seems to me that, you know, autonomous systems and AI could portend some sort of loss of control to tech companies and robots and the machine and LLM. So, is that a worry for the party, or does China have safeguards in place that can keep this under wraps, or even shut it down, as we’ve seen in the past with things like online tuition and gaming and other industries, when they’ve got out of hand?
James Kynge
Great question. We’re just beginning to get a sense that it is, indeed, a worry, and I think this is most pronounced, actually, with regard to OpenClaw, which of course, is not a – it’s a foreign agentic AI. There have been statements from Chinese official bodies talking about the fragile cybersecurity safeguards within OpenClaw. In other words, it’s very easy, according to the Chinese officials, to hack the personal data of the people that are using OpenClaw. And of course, given China’s very much, sort of, ‘Fortress China’ attitude towards its entire cybersecurity, you can imagine that this is really making pulses race in – among the regulators in Beijing. So, that’s a concrete example of what you’re talking about.
I also wonder whether there’s a potential for a bigger backlash. I’m thinking back to was it 2018 when the Ant Group had its IPO cancelled at the very last moment? No, 2019.
Dr Yu Jie
2019, 19.
David Lubin
20, 2020.
Dr Yu Jie
Twen…
James Kynge
2020…
Dr Yu Jie
Was it?
James Kynge
Was it?
David Lubin
2020.
James Kynge
Okay, sorry. Anyway, the Ant Group’s IPO was going to take place in Hong Kong.
Ben Bland
Ant Group being the Alibaba, sort of…
James Kynge
The…
Dr Yu Jie
The Alibaba…
Ben Bland
…financial…
Dr Yu Jie
…e-global bank, yeah.
James Kynge
The Alibaba financial, sort of – it was, basically, going to be an online financial supermarket. It was going to do everything for you, invest in all kinds of other things, and it was going to have a massive valuation on the Hong Kong Stock Market. I was based in Hong Kong at the time writing about this, and I remember at the 11th hour, China’s regulators cancelled this enormous initial public offering, and there were similar crackdowns on Alibaba and Tencent and the other big internet platforms at that time. And I just wonder whether this is also of the same magnitude that could elicit a broader rethink by Chinese policymakers.
But at the moment, as far as I can see, the attitude in Beijing towards AI is still let it rip. You know, it’s still full-on adoption, full-on, you know, “Let’s eclipse America, let’s go farther, let’s,” you know, “increase the efficiency of our industrial companies and our service companies,” etc. So, what I’m describing is not – there is no evidence to support it. I’m just wondering whether this might happen because it seems to me that the level of the dislocation to the economy could be something that people can’t absorb.
Ben Bland
Great, thanks ,James. So, yeah, over to you now. So, just put up your hand or catch my eye if you have a question and we’ll get a mic to you. I’m just looking around the room. We’ve got one there at the back.
Adelaide
Thank you.
Ben Bland
And just do tell us who you are and any affiliation, if you have one.
Adelaide
I’m Adelaide. I’m a PhD student in AI at Oxford, and I was wondering about…
Ben Bland
I think you should be answering the questions…
David Lubin
Yeah.
Ben Bland
…probably.
Adelaide
I know nothing about policy. I was wondering about especially the, kind of, financial backing, because in China it – AI companies tend to be more state-funded, whereas in the West it’s more VC and private tech companies. And you see clashes between, like, CLOD or Anthropic and DOJ in America. So, I’m wondering, kind of, what the long-term effects of where the funding is coming from?
Ben Bland
Who wants to take that?
Dr Yu Jie
I can.
Ben Bland
Okay.
Dr Yu Jie
Or do you want…?
David Lubin
Yeah, go for it.
Dr Yu Jie
Okay. I did a bit of research on this, and soon to be published paper, in terms of, like, decision and also how this process of funding allocation being in place on innovation in general. I think it’s not just AI, but innovation in general. What we have is in – we seems to have this conventional wisdom that everything is very much state-led, very much is led by Central Government itself. However, in reality, I think it’s because of the rapid development of AI and also innovation process in itself, and then those companies become a de facto decision-maker. Because many of those policy areas are extremely new for the Chinese policymakers and what they do instead is a bottom-approach. Really started from most of the so-called champion companies or the small – whatever, the champions, you may call it, you may refer to, and then they asking for extra funding from the government.
So, I think it’s more of a bottom-up approach and sooner or later, if the company become actually big enough, they will be invited as being part of the policymaker. So, I’ll give example of these robotic companies that is happening back to Hangzhou, and the one where German Chancellor Merz, when he visit China, he visit there. And the company – the Chair of the company, now sitting at the board of China’s humanoid robotic machinery and policymaking. So, in a way, that tells you that it’s not just entirely state-led process.
David Lubin
Hmmm.
Dr Yu Jie
It’s an interaction between companies on the one hand, because they’re the pioneers of those policy, then on the other hand, when it come to funding allocation and those are familiar names come up, and the government will naturally just pivot those fundings to those companies.
David Lubin
Hmmm.
Dr Yu Jie
Yeah.
Ben Bland
Great. We have – we’ve got one here in the second row, in this block.
Adrian
Thank you. Hi, I’m Adrian. I’m a student at LSE doing IR. I was wondering about the problems with the export-led model that we’ve been discussing. So, the deflate – like, domestic deflationary pressures, as well as the risks that we mentioned about the global shipping lanes and global demand decreasing. I was, kind of, wondering how large these risk are for China to put all their eggs in one basket, as, like we’ve mentioned how they’re not stopping this export-led growth model and, kind of, what the likely outcome of these risk are?
Ben Bland
David is that one you could take?
David Lubin
You know, it would be more of a risk if the rest of the world was genuinely unhappy buying Chinese stuff.
Ben Bland
Hmmm hmm.
David Lubin
There’s evidence of unhappiness, there’s evidence of un – you know, and I was talking earlier about the increase in hostile trade policies directed at China, particularly coming from the Global South. I mean, in – by some measures, in the last few years, it’s the Global South countries that are the fastest growing source of hostile trade policy directed at China, but als – of course, also, you know, the EU and the US need no further description. So, there’s evidence of unhappiness, but it’s a very, kind of, ambigu – you know, ambivalent field, because on the one hand, as I said, I think it’s true to say that the one thing that unites the US, the EU and the Global South is that they wish China would import more. But as Ben suggested, you know, what is the price you’re prepared to pay to get that goal? And I think that it’s, yeah, it’s a difficult question to answer.
In terms of the, you know, the negative effects of Chinese deflation, it goes back to really what we were talking about earlier. So, you know, the price that Xi Jinping is paying, or the price that the Communist Party is paying in pursuit of this export-led model, in pursuit of this economic policy whose objective is to support China’s geopolitical leverage, that the price it’s paid for the pursuit of that policy in that way is that you’ve got a bunch of unhappy households. And the existence of deflation is a reflection of that unhappiness, and the problem is that it, kind of, creates its own, sort of, vicious circle. You know, the – classically for Economists, you know, the risk of deflation is that if prices start to fall, then, you know, households postpone purchases, expecting things to get cheaper, until you get this vicious circle between falling prices and falling economic activity.
The Chinese authorities have been making effort at the margin to address that problem, but – and I see it’s at the margin, that, you know, the most visible effort that Chinese policymakers have made to, kind of, you know, support this and support Chinese households has been a series of trade-in programmes, where you get subsidies to buy new computers, new TV, new electrical appliances, whatever. So, for example, house – you know, in May last year, the year-on-year growth rate of household appliances was 53% because of the subsidies available to, you know, trade in your old, you know, TV and buy a new one. But by the end of last year, in December 2025, that growth rate was 53% in May last year, was ni – -19% by December. Because all these trade-in programmes is that they’re, kind of, swapping future consumption for current consumption.
James Kynge
Hmmm.
Dr Yu Jie
Hmmm.
David Lubin
And so, it’s a, kind of, nonsense programme in my view. It’s – that’s not all they’re doing. You know, they’ve been introducing childcare subsidies. They’ve been introducing some interest rate subsidies. They’ve been introducing most recently a new social insurance programme for old age. So, there’s effort to give some fiscal support to Chinese households, but as I say, it’s marginal effort. It’s not anything that’s really, in my view, moving the needle of Chinese household confidence.
James Kynge
Just one very quick countervailing point. We’ve been very negative, I think justifiably, about the Chinese consumer and their ability to spend more, for all the reasons that David has been describing. There is one benign wind, however, that I think we should keep in mind, and that is that there is the potential that China’s property decline or collapse, or burst bubble, or whatever you want to call it, could stabilise this year. We’re already seeing the stabilisation of property prices in Tier 1 cities in China and, you know, peop – there’s no sense of when the other tier cities will stabilise in terms of their prices. But it could happen within a year, 18 months or something like that.
When that happens, I think we need to bear in mind that then China will have undergone the biggest property collapse in history and there will be a stabilisation from that, and there would be – you know, there’s a sense, very much, among Chinese households that the value of their biggest asset has declined 40/50/60% over the last five/six years, and that will now stabilise. So, there is a potential. I’m not saying it’s going to be back to the races and I’m not saying that people are going to be spending a lot more…
David Lubin
Yeah.
James Kynge
…but there is a potential for stabilisation that I think we should note.
David Lubin
And sorry, just quickly to reinforce that. That’s ab – I take your point absolutely and it’s critical to understand that one central component of this collapse in household confidence in China is exactly the collapse in the property market. So, if there is a case for making – is that if there’s a case for suggesting that, you know, we are within, let’s say 18 months or two years of the bottom of the Chinese property market, then you should, kind of, arithmetically, expect some improvement in household confidence, yeah.
Ben Bland
Right, and the next question. We’ve got one over there, and in the second row.
Sandeep Kaur
Thank you. Hi, thank you for the views on this topic, really insightful. So, my question is more…
Ben Bland
And what’s your…
Sandeep Kaur
…on…
Ben Bland
…name and any affiliation?
Sandeep Kaur
Ah, my name is Sandeep Kaur. I am from – I work as a Data Scientist at Cirium. So, I’m more from the technical and AI and deployment side of things, and my question is also on, yeah, deployment, actually. So, my question is, basically, when you say “a decentralised economic system” in China, so how that structure will help in the deployment of AI at scale when we compare it to other more market-based economies?
Ben Bland
James, do you want to…
Dr Yu Jie
Hmmm.
Ben Bland
…take that?
Dr Yu Jie
Go ahead.
James Kynge
I mean…
Dr Yu Jie
Hmmm.
James Kynge
…just briefly. My sense from, you know, the last couple of weeks of research I did in China is that this is very much a business motivated adoption of AI. It’s a case of wanting to get ahead and keep ahead of the competition, because they’re adopting AI as well and their efficiencies are growing. I mean, just to give you a sense of the speed of this thing, the number of tokens being used in China has increased 1,000 times in the last two years.
David Lubin
Maybe just say what…
James Kynge
And since…
David Lubin
…the tokens are.
James Kynge
Sorry, a token is a – effectively, a measure of compute power used in AI. And this has – you know, so, obviously, that’s a huge increase, but this has increased even faster since the beginning of this year. To give you the numbers, in December 2025, 100 trillion tokens were used by Chinese AI models and now – that’s the last month, we’re in April now, I guess that’s March, 140 trillion were used. So, in just three months you had a 40% increase in the amount of tokens and that’s down to the use of agentic AI models, which I was talking about earlier, the ones that can actually do tasks.
So, you can see that this is a complete step change and it’s down to, you know, companies, individuals, everybody realising they’ve got to get on the AI bandwagon, on the AI, whatever, train, you call it, because everybody else is doing so, and if they don’t do it, they’re losing competitive advantage, just like Yu Jie’s father was, or mother.
David Lubin
Mum.
Dr Yu Jie
My mum, yeah.
James Kynge
Your mum.
Dr Yu Jie
But on the other hand, I think also, this is a political campaign. You’ve got to remember that every single major Chinese policy initiative, it is a political campaign to start with, first, and we’re now just on the cusps of the 21st Party Congress, which come up next year in October. Now, what will be able to justify a provincial Party Secretary or Governor being elevated as a Politburo member? Looking through a record of your GDP growth of your particular provinces, look into how many investment are high in the technologies you made. So, being justified to be able to integrate themself into the AI bandwagon, that is absolutely essential as a political mission, as well. So, I think that’s also the political story added a layer of – we have seen this boom of the AI in China.
Ben Bland
Next question. We’ve got one here at the front.
Jeremy Pennant
Thanks very much. Jeremy Pennant, a member of Chatham House. With the news that Anthropic have held back their latest AI model due to – well, to the public, anyway, due to the concern of how powerful it is, can the panel give us some sense of the Chinese equivalence, where they’re at, at the moment? And with the rate of progress that you’ve been talking about, especially James, do you sense there might be any state intervention?
James Kynge
The Anthropic comparison I can’t think of a relevant situation in China, but I would like to – on the second part of your question, I would like to point to the possibility of a shortage of compute power. And I think we’re already seeing the foothills of this, because I – earlier, I mentioned Drupal AI, which is one of the AI companies that’s come up with its own OpenClaw competitor. And that is now charging a 50% increase per token, compared to, well, compared to previously. So, you are already seeing that the increased use of tokens by agentic AI it’s creating a shortage in compute power, which is meaning that companies that would like to keep their prices very low in order to boost their market share, are unable to do so because the economics just don’t work.
So, in my mind, and I don’t want to be – again, I don’t want to be too predictive because so many things change so quickly in China, predicting is really a fool’s game, but there could be a point where we get a token shortage and that could increase the price of tokens, which could slow the whole thing down. I don’t know if that quite answers your question, but I see that.
Ben Bland
And as a, sort of, linked follow-up, there’s a question from Babatunde Okuneye online. He’s asking about whether you would agree that “China’s progress on I – AI suggests that the West’s export controls, but chiefly the US export controls on AI technology have had limited impact, or even failed”?
James Kynge
Completely failed. Yeah, that’s the only way to put it. Yeah, complete failure, yeah.
Ben Bland
And could the US or US and its allies, have done anything better, or this just inevitable that if China puts its mind to something, it will find a way?
James Kynge
Well, okay, so that requires a lot longer answer. The main way in which the US tried to restrict China’s ascent up the AI ladder is through restricting the high-end chips. So, we’re talking about NVIDIA H200s, we’re talking about the latest NVIDIA Blackwells. But if you go to Tianjin, to the main market in Tianjin, there, and you lean over one of the counters and you ask them for the latest NVIDIA chips, they’ve got them. And then you say, “Oh, isn’t this breaking some smuggling law?” And they say, “These are American laws. There are no laws in China against smuggling in NVID – banned NVIDIA chips,” right? So, that’s…
Dr Yu Jie
Ah.
James Kynge
…point number one.
Point number two is Chinese AI companies, like Alibaba, Tencent, Drupal, all of the other ones, are big international companies. They have datacentres in Johor Bahru and other parts of the world that are not subject to American chip bans, or at least smuggling is extremely easy, if you call it smuggling, right?
Dr Yu Jie
Hmmm.
James Kynge
So, the – you know, this American policy to slow China down has created inconveniences for Chinese companies, but it hasn’t com – it hasn’t done more than that.
Dr Yu Jie
Hmmm.
James Kynge
The other thing I should mention, it gets a little bit technical, but if you want to go into this subject, look at what Huawei’s done in terms of creating superclusters of China’s domestic made chips, right? I’m particularly talking about the Ascend 910C, okay? So, if – basically, if you com – if you create a cluster of hundreds or thousands of these chips and you put them altogether onto a rack, you can get as much compute power as if you could using fewer, many fewer, latest NVIDIA chips. It’s the – it’s all down to the design of the cluster.
The latest clus – supercluster created by Huawei’s got 15,000 chips in it and that can outperform a similar NVIDIA rack. So, China’s actually got technology that outperforms America’s, you know, NVIDIA based compute power, and next, China’s going to a supercluster of 500,000 chips, next year it’ll be at a million chips. So, you know, so, basically, this is why I say the US sanctions on China have been a complete failure, there’s no other way to put it.
Dr Yu Jie
Well, it’s actually accelerator. It’s the…
James Kynge
Yeah, exactly.
Dr Yu Jie
…accelerator for China to developing…
James Kynge
Yeah.
Dr Yu Jie
…its own semiconductors and so and so forth. But I think one of the key question in here we should also add is, look into the physical infrastructures that would require to build the chips and semiconductors and all this high-end of manufacturing. So, you have the road that it’s enable to deliver, and you have the water system and cheaper electricity that would allow to empower the test of the semiconductors. So, I think that’s also another reason with added benefit, because of cheaper – cheaply available energy, that enabled manufacturing to be able to develop the semiconductor testaments, yeah.
Ben Bland
Oh, well, the analogue clock on the wall is telling me we’re nearly out of time. Thanks everyone for putting down your AI friends and agents for one hour. Some of you have been looking at them, I know, so not all of you, but most of you have set them aside for a whole hour to spend some time with our human experts here. So, please give them a round of applause before we’re all replaced by robots [applause].
