Competition or Conflict? The Future of US-China: Princeton University

Hon Kevin Rudd AC
Global President, Asia Society
26th Prime Minister of Australia

Fall Symposium Closing Keynote
Griswold Center for Economic Policy Studies
Princeton University, United States

4 October 2026

The subject I’ve been given today is conflict or competition between China and the United States on economic policy and between the economies. We cannot properly explore that subject until we’ve wrapped our minds around the nature of China’s domestic economic model: what it is, why it’s there, and whether it will endure.

Given China’s economic scale, what we see in the political economy of this country, the United States, of the Europeans and, prospectively, others in the Global South is a series of mounting reactions against that particular Chinese model. That shapes very much the macro political terrain and policy terrain within which other governments operate.

When we look at Xi Jinping’s economic model, I’m taken by the foundational and fundamental changes he executed at the 19th Party Congress in 2017. A lot of this passed the commentariat by. But if you look carefully into the communiqué and report at that congress, it marked a fundamental shift in the lines of China’s overall economic strategy, which had endured for the previous 35 years.

Deng Xiaoping’s period of reform and opening: everyone in this room is familiar with it, all of its characteristics.

If you look carefully at the 19th Party Congress and go to the fine print, Xi Jinping used an ideological instrument in that particular congress to turn the previous economic orthodoxy of 35 years on its head.

Remember, these are documents intended for a party audience and for the party leadership across the country, and then to permeate through them to the economic actors in the Chinese system and, of course, through to the people at large.

What’s the nature of the change?

Back in 1982, Deng Xiaoping was emerging from the economic carnage of the Cultural Revolution, and what happened after 1976 with Mao’s death, the purge of the Gang of Four and the Third Plenum of the 11th Central Committee. These were seminal events in modern Chinese history between 1976 and 1978.

By the time you get to the Party Congress in 1982, Deng Xiaoping uses the formal ideological instruments of a Marxist party to redefine what Marxists call the zhuyao maodun, the central contradiction.

What is the zhuyao maodun in China’s economic orthodoxy? It is that which Chinese dialecticians point to as a central organising principle, both of policy and subsequently of economic behaviour.

Under Mao Zedong, the principal contradiction was class struggle, at home and abroad. That didn’t get China very far. We could say a lot more about that, but I won’t.

Deng Xiaoping, looking at China’s circumstances which he had inherited, and having himself been purged a number of times, in 1982 redefined that central contradiction in Marxist terms as a contradiction between the people’s aspirations for a better livelihood and unleashing the factors of production to give effect to that aspiration for the people’s livelihood.

It was that core ideological change which gave permission to the subsequent period of reform and opening that we saw in agriculture, industry and the services industries, and in governments from central to provincial to local levels, as the disciplines of the market and what was subsequently called the socialist market economy were incrementally unleashed across the country.

The virtue of China’s penetration of global markets, even before its entry into the World Trade Organization in the early 2000s, was to have the disciplines of price, out of neoclassical economics, begin to inform the efficient allocation of resources, both at home and for those industries exporting abroad.

If you want to look at the wellsprings of China’s double-digit growth through the 1990s and into the Xi Jinping period, it is the product of those factors and forces of production having been unleashed.

Xi Jinping looked at this, not at the 18th Party Congress when he came to power in 2012, but at the 19th Congress, after he’d been in power for five years, in 2017. What he says in that document is: reform and opening has gone too far.

He said it’s created enormous imbalances within the Chinese system. It’s resulted in unequal development, code language for an ever-rising Gini coefficient, and is therefore politically untenable for a Marxist party to sustain.

In terms of unbalanced development, there was also a view that the market, apart from producing unequal economic, financial and income outcomes across the economy, was not effectively developing the totality of the Chinese economy, given its unfolding national security circumstances.

The redefinition of the Marxist zhuyao maodun, the central contradiction, in the 2017 Party Congress effectively has the following characteristics as far as the economy is concerned: a resuscitation of planning versus the market; a resuscitation of the state-owned enterprise sector versus the private sector; and a rehabilitation of the concept of common prosperity, gongtong fuyu.

If you look carefully at the unfolding policy and political dialogue since then, the headwaters ideologically belong in what happened at that congress, when he formally redefined the Dengist orthodoxy determined back in 1982.

If you look at simply one of those changes, the difference between the market determining the efficient allocation of resources as opposed to planning, that is where we come to the core elements of Xi’s aggregate economic model today, as it impacts the rest of the world, including the United States.

Firstly, what you see in the period after 2017 is the emergence of Xi Jinping’s doctrine of national economic self-reliance, zili gengsheng. This was a term last used during the Cultural Revolution, when China had to be self-reliant about anything and everything, but had been pushed way to the side during the entire Deng Xiaoping economic experiment.

This self-sufficiency and self-reliance doctrine becomes wrapped up in Xi Jinping’s concept of comprehensive economic security. The securitisation of the economy and economic policy has its antecedents in the thrust towards national self-sufficiency and self-reliance, in turn anchored against what Xi Jinping perceived to be challenging external security circumstances.

Furthermore, against a future scenario involving Taiwan, one in which China would become, in his calculus, economically isolated, it was critical for China to be self-reliant.

In understanding what it means for China to have a more planned economy, this self-sufficiency, self-reliance and economic security dimension lies at its core.

Second, beyond China’s own self-reliance and self-sufficiency, there is also a predisposition evident in the policy documents postdating 2017 to cause the rest of the world to become progressively dependent on China.

For those of you who’ve studied the dual circulation model, shuang xunhuan jingji, if I was to provide a lay summary of what it is in practice, it is along these lines: to cause China to become increasingly independent of the world in terms of what it essentially needs through its own supply chains, while simultaneously causing the world to become progressively more dependent on China.

It is not just self-reliance at home, for which North Korea exhibits a horrendous example of what that could look like. The other part of the orthodoxy is for China to be increasingly active in leveraging its size, the size of its markets in the world at large to cause, through supply chain management and construction, the world to become progressively dependent on China. Look at critical minerals and rare earths.

This was consummated in the formal articulation of the dual circulation economy concept in 2020, and has been the orthodoxy ever since.

The engine room of the dual circulation economy is, of course, China’s techno-industrial strategy.

You’re all familiar in this room with Made in China 2025, launched back in 2015. Everyone looking at the launch in 2015 said they’ll never do it. Well, 10 years later, against the 10 core technology goals which China set for itself, any rational and objective marker of the exam or term paper would say the following: somewhere between seven and a half and eight out of 10, tick.

There are two or three things still missing: chips being one, civil aviation and engine design being another, and one or two other things besides. But against where China was in 2015 and the other technology priorities identified in that strategy, it is a remarkable feat in terms of state-driven techno-industrial strategy to deliver the outcomes we have seen.

That concluded last year, but this is simply the starter. The entrée is still to come. It’s called China’s medium- to long-term technology plan, 2025 to 2035, where, surprise, surprise, those things China did not quite reach in the first plan are identified as the core priorities for the second, including semiconductors, civil aviation and jet engine design. Thirdly, nuclear fusion. Fourthly, quantum.

In other words, looking out to the techno-industrial frontier, where China not only needs to catch up, but then to be leading in the emerging technologies of the future. Quantum and fusion, in their view, go to the heartland of what it will take to power computers and compute power, as well as the energy sources necessary to do that in, hopefully, a sustainable way.

Of course, running a techno-industrial strategy of this order of magnitude, even within China, requires finance. So how do you do it?

The sources of finance for this massive piece of industrial strategy are partly out of state finance, but China’s budgetary circumstances are severely pressured, not least because it has a dysfunctional taxation system writ large. There are multiple professional reports on the structure of the Chinese tax system.

The other source of capital, primarily delivered through credit through the Chinese banking system, is leveraging this phenomenal savings ratio on the part of Chinese consumers into the techno-industrial strategy, at minimal interest rates. It’s a cheap source of domestic finance.

That’s one source of capital. The other source is China’s net exports, and that brings us to the question of how that’s impacting the rest of the world.

This model, driven by industrial policy, techno-industrial strategy and state subsidy, largely financial subsidy, occasioned by recourse to budget but primarily by drawing on the capital available in the banking system, and then turbocharged by an annual trade surplus the likes of which has never been seen in world economic history, in absolute or relative terms, is ultimately what those of us who have looked at it for some time have concluded to be an unbalanced and ultimately unsustainable model, simply in terms of financial cost at one level.

On top of that is the equal and opposite reaction it is generating around the world.

If you look at the drivers of Chinese economic growth today, the unbalanced nature of this growth model is reflected in the following.

Private domestic consumption, as of today in the Chinese economy, is flatlining. Its net contribution to growth in the given year is not where it should be for a low- to middle-income economy.

Second, private residential construction, normally a driver of growth in any emerging economy, is not just flatlining. It’s been in recession for the last six years, since Liu He imposed his three red lines on the further expansion of residential housing credit, and Xi Jinping announced that houses are built to live in, not to speculate on.

Furthermore, in Xi Jinping’s view, this was not a place for speculative capital. Capital could be productively used in what he terms the real economy. Go back to technology and manufacturing.

If private domestic consumption is not growing as it should be as a net contributor to the growth equation, private residential construction is heading into its sixth year of cumulative recession as a sector, thereby burning a huge amount of personal wealth held by Chinese consumers over a long period of time, further depressing consumption because consumers don’t have the money they used to have.

Then how about that third driver of growth, private fixed capital investment? Again, flatlining, because private entrepreneurs, by and large, have read the headline messages: we’d like more common prosperity, gongtong fuyu. We’d like a more resurgent state-owned enterprise sector. Thank you very much.

For those of you in the platform economies, the pingtai jingji, we reserve the right to take a golden share in your corporations at a time of our choosing.

If you were an American sitting in that system, you may not feel entirely encouraged. And guess what? Our Chinese entrepreneurial class, with some notable exceptions, are themselves somewhat discouraged as well.

On the measures of private business confidence, which are hard to assess given the nature of the Chinese system, but reflected primarily in what we can disaggregate to be the private fixed capital investment numbers, and that’s tricky in itself, this too is flatlining.

What’s generating growth in the Chinese economy? The officially published data comes out of the National Bureau of Statistics, which in recent times has come under methodological pressure and challenge from the party itself. Therefore, the believability of the numbers for the international economic audience has also come under challenge.

The two pistons working in this five-piston engine are state investment, leveraging budget and credit out of the private savings banks into the techno-industrial strategy, plus net exports.

Net exports, if you accept the official data at face value, represented one-third of all growth last year in the aggregate Chinese economy. I suspect significantly more than that once you interrogate the data further.

We have this techno-industrial strategy, driven by a largely mercantilist approach to exports, all within the rubric of a more statist approach to economic management and Xi Jinping’s overarching doctrine of dingceng sheji, top-level design, as the ultimate planner-in-chief.

People look at this strategy unfolding and say we need to convince China to change course because it’s unsustainable for themselves in terms of the financial inputs. It’s also unsustainable externally because of the U.S. reaction over 10 years now through its own wave of tariff protection and industrial policy intervention, and the possibility, slash probability, that by year’s end the Europeans will move in the same direction.

Look at the carnage in German manufacturing every month: 10,000 jobs being lost.

Then the third subregion in the world, at a different level of manufacturing, is Southeast Asia, where the Chinese trade surplus is the largest of all three. That’s where the big surplus lies of the $1.2 trillion. Next comes Europe, and then the United States, shrinking because of the tariff measures which have been imposed.

Why is it argued that this strategy is ultimately unsustainable? Financially, domestically, it’s unsustainable because the impact on budget and the state savings banks is itself unsustainable over time, as the appetite increases when so many of the corporations being subsidised by soft credit in this direction are themselves not profitable. Perhaps one-third of firms in the export sector.

Then you add the other element of unsustainability: the rising political reaction from abroad, as political leaders respond to the voter base saying, “I don’t want to be deindustrialised to the extent that I have all my supply chains industrially now dependent on China, not to mention job losses.”

Look at the Alternative für Deutschland result in the most recent Länder election in Germany, in Saxony-Anhalt, where AfD, the fascist party, now runs a German Land for the first time since the war.

These forces are at work. But anyone who thinks that China’s current policy settings are an accident, or are there because Chinese policy leaders haven’t got the right policy advice, is not actually looking at the evidence.

China knows precisely what it’s doing. It believes this strategy of theirs is right, and they intend to do it because the endpoint is to turn China into the indisputable techno-industrial power of the world. That very much comes back to the psychology of Xi Jinping that I referred to before.

When we look to the projections that by 2030 China’s proportion of global manufacturing will rise from one-third to something like 45 per cent, we are getting to threshold points in the global deliberation on these questions where it becomes unsustainable.

Artificial Intelligence

The trends I’ve sought to describe here are turbocharged by the new dynamic called artificial intelligence.

If you read the Chinese literature on this, Xi Jinping has an entire ideology around what’s called new quality productive forces, xin zhi shengchanli. It’s a complex concept. It comes out of the Marxist theory of labour and their view of how productivity can be yielded in a planned economy.

Essentially, it has the following characteristics.

First, the rapid deployment and diffusion of artificial intelligence across firms in the Chinese economy will turbocharge China’s productivity growth. There is reason for them to have some confidence about that, given the pace of diffusion: by one set of analyses, some 53 per cent of firms already taking up AI, against an average of 25 per cent here in the United States.

With the state, the good old state, under top-level design from Xi Jinping himself, saying: for the next five years, you all have diffusion targets of around 73 to 75 per cent. So you don’t get much of a choice. You’re going to be out there adapting and diffusing, whatever business you are running.

If you read the Chinese literature on this, it’s stated in these terms: artificial intelligence turbocharged by this wider concept of new quality productive forces, with data itself now formally categorised as an entirely new factor of production. Land, labour, capital, enterprise, data.

The new quality productive forces are about how this data is to be harvested across the totality of the economy, within each firm and across all sectors. That’s very much the productivity agenda, which the literature says will not only catch up with the United States, will not just keep pace with the United States, but will leapfrog the United States.

That’s what the literature says in their own internal discourse, in my favourite fun journal called Qiushi. It’s the Communist Party’s theoretical journal for super nerds. If any of you are suffering insomnia, I can send you some back copies, or even worse, my analyses of back copies, which will send you even further to sleep.

I even gave a lecture the other day at the Australian National University, my alma mater in Canberra, entitled “New Quality Productive Forces”. You should have seen them go to sleep. It was wonderful to watch.

The second big advantage which Xi Jinping sees as coming from the AI revolution, aka the new quality productive forces revolution, is not just boosting the productivity of firms. He also talks about this representing a paradigm shift for the Chinese economy writ large.

What does he mean by that? The language he deploys is that now we will be actively, through artificial intelligence, deploying “the visible hand of the state” in directing “the invisible hand of the market”.

Kind of an interesting concept if you’ve grown up on Adam Smith, like many of us have in this room. Effectively saying that this represents a brand new challenge to Smith, The Wealth of Nations, price discovery through supply and demand.

Remember Smith standing on the docks in Glasgow, counting all that stuff coming in and out of the country, and saying, “Eureka! I’ve worked out how prices are set.” That essentially constituted the price-based orthodoxy of, let’s call it, neoclassical economics for the last 250 years as it’s unfolded.

Of course, there’s a Keynesian challenge, but the Keynesian challenge is within the overall framework of a neoclassical orthodoxy still anchored in price.

What the Chinese literature in their theoretical journals is now saying is: artificial intelligence, if it can efficiently allocate resources within firms to lift productivity growth, can therefore do it across firms, across sectors and across the total economy.

In other words, it becomes the silver bullet for state planners writ large.

I’m simply describing to you the level of attachment to this proposition. I’m not defending the attachment.

This is alive and well in this massive level of political attachment to the AI revolution, the fourth industrial revolution, not just achieving leapfrogging of the West, the U.S. and the rest on productivity, but transforming the entire economic paradigm.

They’re not likely to step back from this, if I read the literature carefully.

Donald Trump said, most recently in his colourful response to Dario Amodei and Dario’s article on the future of artificial intelligence, “We Must Pace the Frontier”: “He who wins AI wins, and we’re not going to have any regulatory slowdown. Thank you very much.”

Before anyone gets misty-eyed in this room about what the Chinese strategy towards AI is, China’s strategy towards AI is basically the same as Trump’s: he who wins AI wins.

If you look carefully at the State Council directive which Xi Jinping had issued 10 years ago, in 2016–17, on the centrality of AI for the next industrial revolution, he has been ahead of this curve for a very long time. He sees it as transformational in the grand strategic competition between China and the United States.

If you go across the AI stack: energy, China ahead; compute, China behind, probably by four or five years; frontier models, China behind, not by four or five years, but by five or six months, maybe; diffusion, China way ahead, based on the numbers I ran through before.

Competition across the stack is acute, but there are only two powers in the world in this domain. There might be a bit of Dutch lithography coming up the side on elements of compute and chips. We all know that. But this is a two-horse race.

What we see is this enormous belief by the Chinese leadership that this represents a silver bullet, a secret weapon in executing China’s ultimate competitive race with the United States, continuing to be embraced and, in my judgment, turbocharging the pre-existing techno-industrial policy settings I described before, within the framework of the dual circulation economy.

That is, a self-reliant China at home, and a world increasingly reliant on China abroad.

As I move to conclude these remarks, we’re left with the question: will China change this approach?

I’m a realist, and those of us who have grown up in the Antipodes, as I have, are sort of born realists. I’m the son of a farmer. You plant something, it grows. You don’t put water on it, it dies. You feed the cow, it produces milk.

That’s the school of Anglo-Saxon empiricism from which I come. So I’m a complete loss to the French academy because I’m not of the school which says: that’s fine in practice, but will it work in theory?

I have the reverse approach.

As I said, these policy settings are now hardwired into the Chinese system. If people believe that polite policy, political and interpersonal discourse is going to cause a moment of Buddhist self-revelation in Xi Jinping’s mind to say, “Oh God, I got that wrong. It’s time to change course”, knowing a little bit about the Chinese system since I first started to study this language 50 years ago this year, I don’t think that’s likely.

What could change it?

As I said before, the domestic pressure on the financial sustainability of this entire techno-industrial strategy is ultimately placing huge pressures in the system.

The external pressures coming out of this emerging tariff policy and industrial strategy reaction, not just from the United States but progressively from Europe, will conceivably, as we get into calendar year 2027, bring unprecedented challenges to bear on what is already a Chinese domestic economy addicted to overproduction.

It is producing these through what the Chinese admit is involution, neijuan, selling product on the global market at sub-cost levels. Not just sub-market levels, but sub-cost levels. They are the two forces very much at work.

What will happen? The locus of control and the evolution of this debate lie with the Europeans and then the Southeast Asians. There are three ginormous subsets of China’s net export performance with the world.

The United States is coming down rapidly in terms of the Chinese trade surplus, notwithstanding modest changes in trade policy in Washington last week, in sectors described as non-sensitive. They represent perhaps $30 billion of the total trade number. Not insignificant, but certainly not overwhelmingly significant.

The Europeans, however, represent a very large chunk. The centre of gravity in Brussels, Berlin and Paris is shifting so rapidly on this question that people should watch very carefully what happens by December.

Southeast Asia is somewhat different because they are less predisposed towards taking on China directly. They’ve studied the history of the tributary states. They know what happens when you misbehave: China arrives and is very unhappy. So we’ll see what happens in Southeast Asia.

AI Regulation

To conclude, what does all this mean for what was debated last week in D.C.? Not just on the general bilateral relationship, not just on the general economic relationship, but specifically on AI and the possibility of negotiating guardrails between the two, on what they describe as dangerous AI.

You’ve seen the literature. You’ve seen something of the U.S. view, despite what the president has said: we’re proceeding with a deregulatory environment because that turbocharges American performance and our competitive performance in the world. Remember, he who wins AI wins.

Beneath that, across others in the U.S. administration, both in the White House and certainly in the Treasury, is a deep and emerging view that autonomous agents produced by frontier models, including proprietary models and not just open-source and open-weights models, now pose three sets of risks.

First, agentic AI acting autonomously against infrastructure, social, economic, military and the rest, of which we’ve had a number of semi-real-world and real-world examples in the last three to six months.

Secondly, such frontier-model agentic AIs finding their way into the hands of rogue states.

Thirdly, also into the hands of rogue actors.

At Asia Society, where I work, a few weeks ago we were involved in a Track 1.5 with the Chinese Communist Party and sat opposite Wang Huning from the Politburo Standing Committee; also Ma Zhaoxu, the senior vice foreign minister responsible for last week’s visit; and the International Department of the party and its own personnel directly engaged in the security dimensions of the AI debate within China itself.

What I’d say to a group like this is as follows.

China, at least as articulated then a few weeks back, thinks that the problem of dangerous agentic AI being produced by frontier models and getting out of control is overstated.

Secondly, to the extent that it exists, it’s primarily an American problem.

Thirdly, they have a level of self-confidence about their own cyber defences that, in my judgment, exceeds the reality. Because they’ve honed those cyber defences against their principal concern, which is algorithmically turbocharged release of unauthorised information within China that would destabilise the Chinese leadership, they have got reasonable defences.

But I’m not sure those defences extend to water infrastructure, power infrastructure, general civilian infrastructure, financial services infrastructure, let alone the infrastructure in and around the People’s Liberation Army itself.

There are already anecdotal stories emerging about how parts of those systems have been penetrated, and not by nation-state actors, just by agentic AI going for a wander on a Thursday afternoon.

The takeout message in Beijing is that we do not share what we perceive to be a new American sense of urgency. In fact, worse than that, we Chinese believe this is just a regulatory attempt by you guys to impose a set of handcuffs around us to keep us back, given that in many of the categories of the AI stack we are behind.

The dialogue track which Secretary Bessent announced last week on AI is useful. It’s yet to be populated either with personnel or with a concrete agenda. I’m glad that the track now exists because it didn’t really exist before.

The challenge moving into the next meetings, both in Shenzhen for the APEC summit next month, and then the G20 down at Mar-a-Lago in December, is that we will need to see this track generate real proposals in terms of effective guardrails.

First, the proposition that agents and models should not be released until they’ve been properly tested against existing cyber defences, and if there is a problem, for release to be deferred. For that to be agreed as an industry standard, a national standard and then a bilateral standard: that’s a big ask.

Secondly, if that can’t be done, how do you calibrate collective cyber defences against the unknown, emerging, turbocharging, swarming, recursive-self-improvement-driven agents?

That’s the work they have cut out for themselves in the next six weeks.