Can AI write a New Code for a “Perma-Crisis” World: A Conversation with Primavera Chairman Fred Hu and Nobel Laureate Prof. Michael Spence

Preface

At the opening of Primavera Global Investment Conference 2025, Professor Michael Spence, Nobel Laureate in Economics, joined Fred Hu, Primavera’s Founder and CEO, for an insightful discussion about global macroeconomics, the impact of artificial intelligence, and more.

Sharing his observations on the current state of world markets, Prof. Spence noted that the Trump 2.0 era is not only impacting China, but even traditional U.S. partners are also contending with major adjustments to long-standing global economic and trade relations. At the same time, China’s significant domestic consumer economy has provided a much higher level of economic flexibility and resilience, which has continued to exceed expectations.

Both economists agree: China’s economic fundamentals remain robust, underpinned by deep talent pools, research excellence, and long-term competitiveness in critical sectors such as AI, biotech, robotics, renewable energy, and more.

With virtually all markets facing pressures around sustaining growth, balancing trade disruptions, as well as less favorable demographics trends, Prof. Spence believes that AI-driven productivity gains represent a substantial opportunity area for the global economy moving forward. He also underscored the enduring importance of human judgement in the knowledge economy as AI applications become more prevalent across many industries.

 

Speakers:

Fred Hu, Founder of Primavera Capital Group
Professor Michael Spence, former Dean of Harvard Arts & Sciences and Stanford GSB, Nobel Laureate

Fred Hu:

Two years ago, you co-authored Permacrisis: A Plan to Fix a Fractured World with Gordon Brown and Mohamed El-Erian. How do you define ‘permacrisis’? 2025 has been a volatile year, with global tariff tensions, prolonged conflict in Ukraine, and Middle Eastern turmoil. How do you see the global economic outlook compared to when you wrote the book?

Professor Spence:

The idea for the book started during a pandemic-era discussion, from a sense that the global economy was undergoing profound structural change, not just a temporary shock. Many people were skeptical back then, believing most disruptions were Covid-related and would ease after the pandemic. Even central banks bet on the same outcome, until inflation changed the playbook.

Now, many have started seeing economic security as important as national security, which requires governments and companies to adapt. Therefore, we are in a period that is hard to describe exactly.

As for globalization, what we are seeing is a partial reversal due to tariffs and other global tensions, compared to what we had after World War II. And yet that doesn’t mean globalization has stopped, it has only evolved. Biden didn’t roll back the U.S.’s China policies during the first Trump administration, and as Trump returned to office, he continued to advance these policies from different directions. So, it’s time to let go of the mindset that the world is going to revert to the old system. The old playbook is gone.

Fred Hu:

Speaking of Trump’s tariff policies, I remember back in the 1980s when I was a graduate student, Trump was already writing op-eds calling for tariffs on Japan, and he has been consistent on this front. Now his policies are not only targeting China but also hitting countries that had traditionally been the U.S.’ allies, including the EU, Canada, Mexico, and so on. I was talking to Swiss business leaders, and they mentioned even Switzerland has been significantly affected.

What’s your advice for business and policy leaders managing in such volatile, uncertain times?

Professor Spence:

Liberation Day definitely shocked the system. Trump has been looking at tariffs for a long time, but it is really surprising how aggressive the U.S. has become with its traditional allies. Having grown up in Canada and lived in Europe, it’s strange to see. There is a rising judgment among European business and political circles: The U.S. is becoming an increasingly unpredictable trading, investment, and security partner.

The U.S. accounts for 25% of the global economy, but the remaining 75% will keep going, trading goods and services, and investing. The impact of tariffs for economies deeply integrated with the U.S., like Canada and Mexico is certainly the biggest, whereas for countries in Asia, there is more flexibility. We are seeing encouraging signs of recovery in China – the impact is real but not fatal.

Fred Hu:

Although China has been a major target of Trump’s tariff policy, its huge domestic market gives it more room to maneuver compared to other regions.

Recently, China has been discussed widely by Western commentators, but what is the true situation? Today, our Primavera partners from around the world are gathered in Shenzhen and can personally experience the vibrancy of Shenzhen and the visible boom of the economy.

If you were in Beijing advising top policymakers, what would you suggest they do to boost consumption and investment to help the economy resist external shocks and accelerate its recovery?

Professor Spence:

That’s not an easy question! Maybe I can answer by thinking in terms of strengths and weaknesses.

On the strengths, China is a technological superpower. This is hugely important. For the past 15 years, both the government and the private sector have invested heavily in technology. As a result, China will be a major player

in key future industries such as biotechnology, AI, and quantum computing. This will be the critical engine for China’s long-term economic development.

The short-term headwinds are well known. I believe the main issue is a demand problem. On one hand, the aggregate demand is low. On the other, much of the demand is too skewed toward public sector investment.

I used to think government should just stimulate demand, but now I realize that you can’t take an economy investing 50% of GDP and flip it to 60-70% consumption overnight. The supply side cannot accommodate this shift; workers who build roads and bridges cannot suddenly work in service sector. Resolving this structural issue is a long-term challenge.

My main criticism of China’s economic policy in recent years is that I think the government should be more proactive in fixing household balance sheets, and policies need to be implemented more swiftly.

Finally, I think the ambiguity over the private sector’s role is a fundamental issue. Investors and entrepreneurs need clearer signals and more predictability.

Fred Hu:

While facing real estate adjustments, deflationary pressure, and external geopolitical strain, China’s policy toolbox still holds many options, but it requires clarity, decisiveness, and a willingness to act boldly. Over the past few years, the government has demonstrated it can implement policies decisively.

China’s rapid economic growth after the reform and opening-up period stemmed from a favorable alignment of timing, geography, and human factors, where people, economic inputs, and globalization converged. Faced with challenges like state-owned enterprise reform and the Asian financial crisis, China was compelled to act decisively to mitigate or counteract these economic problems. This fundamentally altered the trajectory of China’s economy and national development.

I believe that the core elements underpinning China’s high-speed growth still exist. If China continues introducing pro-growth policies, whether short-term stimulus or structural reforms for the private sector, I believe the medium- and long-term outlook remains positive.

I have often written about the role of innovation, technology, and AI in driving long-term economic growth in the U.S. and globally. We are now in the midst of an AI revolution. While the US is a clear leader, China is now at the forefront too – in AI, quantum, biotech, robotics, renewables, EVs, even space tech. As some in Washington admit, China has become a technological innovation powerhouse and an all-around competitor to the U.S. In certain areas, China is even ahead.

China is also a vibrant center for biotech innovation. A third of all global pharma licensing deals this year involved Chinese firms. Multinationals like Pfizer and Novartis have openly noted that they remain impressed by the speed of innovation and development by companies in China. If scientists in China can find effective ways to treat cancer, diabetes, and cardiovascular diseases, it benefits not only China but also the U.S., Europe, and the entire world.

The same applies to the field of AI, where some policymakers in markets like the U.S. remain highly focused on competition. We all know that since the Industrial Revolution, science has been borderless—ideas cross frontiers, and progress depends on global exchange and collaboration.

I’d like to hear you view on the AI revolution for the global economy. Nobel Laureate Geoff Hinton once warned that AI could cause potential labor market disruptions.

Professor Spence:

This is a very important topic, and I’ll share a few personal observations. Stanford University’s annual AI Index Report is fascinating and worth reading. I can draw 4-5 key points from the latest edition.

First, there are only two AI superpowers in the world: China and the U.S. Only these two countries possess all the necessary elements for AI development: talent, computing power, and so on.

Second, the cost of training and inferencing AI models are plunging, and the gap between China and the U.S. is quickly narrowing. Previously, some said the gap was 12 months; now, some say it’s three months. We now see that both superpowers possess the technology. Restrictions are imposed because they view the technology as having strategic significance, but frankly, the technology and science community often finds ways around these barriers. For instance, DeepSeek is a very important case of innovation, born out of and overcoming these very restrictions.

As time goes on, where will the impact of AI manifest? It is already evident in science. For example, we’ve seen the Nobel Prize in Chemistry awarded to Google’s AlphaFold developers, Demis Hassabis and John M. Jumper, alongside David Baker, for solving the complex problem of protein structure prediction using AI models. But at the company or sector level, productivity gains are uneven and take time to show up in data, both in the U.S. and China.

A third observation is that AI has driven striking new advancements in robotics across the board. This is starting to affect blue-collar jobs, but intelligent robotics technology is still maturing, and it will at least take about seven to ten years before it has a significant economic impact.

My final observation concerns how people interact with and understand technology, a topic I want to spend more time studying in the future.

Many people who study economics believe the world is moving toward one characterized by resource abundance. This means that most things we currently consider scarce will become less so. This terrifies them because, since Adam Smith, our entire economic system has been built on the relative values determined by scarcity. Of course, I believe that future is still far off.

In the short to medium term, generative AI is being used both to streamline and enhance labor through automation. You see news globally about business owners using AI to replace customer service and other roles. If labor displacement goes too fast, we risk a major backlash and those humans who are impacted may become a source of resistance to AI. I believe striking the right balance, with careful coordination across education, private markets, and government policy, will be a big challenge.

Fred Hu:

As with every technological revolution in human history, there is disruption and turbulence, but overall, humanity needs to adapt to maximize the upside and minimize the downside. Therefore, we should neither be blindly optimistic nor fall into irrational panic. Now, let’s open the floor for questions.

—

Q&A Session

Question 1

I recently read two data points: First, China has 250 million large language model users. Second, the domestic production rate of AI chips in China is expected to reach 40% by 2025 and 55% by 2027—that’s not far off. Given these figures, has U.S. policy failed to restrict China’s AI progress?

Professor Spence:

U.S. policy will almost certainly slow down China’s AI development slightly, but only slightly.

When you impose such restrictions, it compels the brightest minds in China to find new paths and explore methods to enhance the efficiency of models and algorithms. The result is that, without higher-level intervention, the narrowing of the U.S.-China gap is foreseeable. As I mentioned, a key reason for China’s development over the past ten to fifteen years is a profound understanding of the importance of investing in education, finance, and talent.

Fred Hu:

I’ll add that it’s become harder for entrepreneurs in China to access advanced GPUs, but that’s pushed them to work smarter, focusing on efficiency and hybrid models. Instead of using massive TPU clusters, they are developing models that are compatible with less hardware while still achieving outstanding performance. Besides DeepSeek, we see models developed by Chinese tech companies like Alibaba and ByteDance characterized by their efficiency and less reliance on GPUs compared to counterparts overseas. OpenAI has spent tens of billions of dollars training its large models since 2015, resulting in high subscription fees, whereas DeepSeek is virtually free. This reduction in cost speeds up AI adoption across industries in China.

Professor Spence:

I want to add that many AI giants have raised the amount of capital in a very short time that took Google over a decade to accumulate. Therefore, one must be cautious about the valuations of AI companies. Before investing, you must ask a crucial question: is this valuation truly justified?

Question 2

The industrial structures of the U.S. and China are fundamentally different. China’s industrial output now accounts for over 30% of the world’s total, similar to the U.S. a century ago. The U.S. industrial output has now fallen to about 10%. Given the differences in industrial structure, we see AI applications in China advancing simultaneously in industrial production and consumer sectors, whereas the U.S. is more focused on B2C applications. Is my observation correct?

Professor Spence:

That is an excellent question. Due to the structural differences you mentioned, the paths of AI development in China and the U.S. will certainly differ, but I believe these differences will decline over time. Instead of these structural disparities, I believe that business dynamism and focused economic policy for technology are the core drivers.

The future AI superpowers are the U.S., China, India, and Europe. India’s development scale is currently insufficient. Europe’s prospects are bleak. A mutual friend of Dr. Hu and mine, the former head of the European Central Bank, was tasked by the EU to assess Europe’s economic performance and competitiveness. His conclusion was quite negative. European governments have neither made the structural changes needed for the capital markets nor altered the way they fund scientific research.

While Europe may have many interesting investment projects, it is may not standout in the technology field going forward.

Fred Hu:

As incomes rise, economic structures tend to converge toward a service-led structure, mirroring the U.S. and Japan over past decades.

A major driver behind the shift in China’s manufacturing sector is technology. AI and robotics are poised to provide an offset to the demographic pressure in a country like China, which is facing an aging population.

Professor Spence:

This is a very serious issue. It’s not just about growth; it’s also about the dependency ratio in the context of a shrinking workforce and an aging population. The pressure from the dependency ratio is worrying.

In my view, given headwinds to economic growth, fragmentation, and population decline among other issues, AI-driven productivity is likely the one way to meet these pressures.

Question 3

If you had a crystal ball, what do you think would be the most optimistic scenario for the global economy in 2026?

Professor Spence:

My baseline forecast remains largely unchanged, but I believe the downside pressures have increased. The negative factor is that the U.S. economy and labor market will continue to weaken.

On the other hand, the Chinese economy is recovering, which will not only boost its own economic performance but also lift the entire surrounding region. Globally, I believe India can maintain a quite respectable level of growth, while Europe is highly likely to remain flat.

Question 4

There’s a recent saying that as AI develops, humanity won’t lack productivity in the future but will lack sensibility—the perception of value, emotion, and meaning. From an economic perspective, will this become the new scarcity?

Professor Spence:

I completely agree, though we don’t have very precise metrics for it.

AI allows for rapid transfer of certain skills and expertise, much like navigation maps. Expertise used to be a scarce system; acquiring it required time, effort, and cost. What knowledge can be accessed instantly via AI, and what cannot?

Take investors, for example. Can you let AI tell a young person how to invest? Or does it take 20 years of experience to assess people and judge the qualities of an entrepreneur?

I am certain that, at least with existing tools, there is quite a lot that we cannot transfer easily, such as emotion, judgment, common sense. These things are hard to quantify and transfer precisely, but they are real and valuable.

If knowledge becomes abundant, these soft skills only become more precious. Society and markets will need to rethink how they value the truly human.