Viewpoint: the US and China risk technological stagnation

Barely a day goes by in Brussels without someone looking enviously at US or Chinese technology. From artificial intelligence to semiconductors to electric cars, the continent seems convinced it needs to adopt the supposed dynamism of the US, the industrial planning of China, or a mixture of both. 

But the future of innovation in the US and China is in serious jeopardy, says Carl Benedikt Frey, a Swedish-German economist based at the University of Oxford whose recent book How Progress Ends makes the case that oligopoly and repression respectively have stymied the two countries’ inventive prospects. 

We are not “doomed,” Frey told Science|Business, “but on current trajectory, it doesn't look particularly good, either in China or the United States.” 

A central message of Frey’s book, which sweeps from 11th century Chinese water clocks to modern day artificial intelligence, is that economic and technological progress is far from inevitable.

Indeed, for most of human history, it has advanced at a snail’s pace, or not at all, blocked by vested interests, from guilds in Qing China to the Roman emperor Vespasian, who banned a machine for transporting columns for fear it might create unemployment. “Progress is fragile,” Frey said. “Technology leaders can stagnate and fall behind.” 

Across the US, Europe, and even China, which still remains much poorer per capita, productivity growth has fallen over the past couple of decades. This slowdown is “global,” Frey says. 

The US once thrived due to decentralisation and competition, his book argues, from the individual patenting inventors of the 19th century, to the Silicon Valley in the 1980s, which helped create modern computing. 

Crucial to its success, the US has gone through bouts of antitrust, periodically limiting the dominance of over-mighty companies to allow new players with fresh technology to emerge. For example, pressure on IBM during the 1960s prevented the company extending its hardware dominance into software, opening the door to the rise of Microsoft. 

But now the country is once again in the grip of corporate centralisation, Frey says. His book is stuffed with statistics showing that “creative destruction” is waning in the US, with less reshuffling between companies, fewer start-ups challenging incumbents and a greater proportion of inventors working for existing firms. 

Companies engage in “killer acquisitions,” buying up budding rivals simply to shut down the competitive technologies, and spend vast sums on lobbying regulators to entrench their positions. 

Cronyism

Epitomising this centralisation are companies such as Google, Meta and Microsoft, which have captured historic levels of political influence and money to block new technologies that might threaten their business models, according to Frey.  

Since Donald Trump’s inauguration, during which plutocrats such as Mark Zuckerberg, Jeff Bezos and Elon Musk had a front row seat as the new president was sworn in, this trend has developed “on steroids,” said Frey. “I think the book turned out to be more timely than anticipated, unfortunately.”

Trump’s “elaborate” system of tariff exemptions, for example, “invites cronyism,” said Frey. This means a US business environment in which “you better have a CEO or a management team that is politically savvy, rather than technologically [savvy].” Compounding this problem of cronyism, the US government has taken shares in key companies such as chipmaker Intel, for example. 

Other straws in the wind support Frey’s argument that White House connections now trump genuine innovation in the US. Many of the country’s incumbent tech giants, such as Microsoft, Google, Amazon, Meta and Apple, are donors to Trump’s gold-plated new White House ball room, for example.

Earlier this month, the Financial Times reported that the country’s mega-rich were “flocking” to move to Washington DC to take advantage of what it called “Donald Trump’s proximity-dependent policymaking.” Zuckerberg, for example, snapped up a $23 million DC mansion last year. One nuclear fusion start-up with links to the Trump administration reached a $20 billion valuation last year, despite not having any revenue. 

AI to the rescue?

The one big exception to this narrative of US ossification is OpenAI, which burst into the upper echelons of the US tech scene in 2022 after the launch of ChatGPT. It deployed large language models (LLMs) based on AI transformer techniques that were originally developed, but not deployed, by Google, possibly because the tech giant feared LLMs would cannibalise its core search business. 

“I'm not saying that it's impossible for a new firm to challenge incumbents, but it's getting harder,” said Frey, who pointed out that the OpenAI breakthrough happened under the Joe Biden presidency, not Trump. 

For an academic with AI in his title – Frey is officially the Dieter Schwarz Associate Professor of AI & Work – he’s remarkably pessimistic about the potential of LLMs. 

They “show no signs of original thinking,” he writes, and recalls similar predictions of imminent human-level machine intelligence in the 1970s that came to little. Despite the hype, US productivity growth remains low, and the excitement over LLMs has closed off other interesting avenues of AI research, Frey says. 

Some research funders in Europe agree, with money now flowing to non-LLM forms of AI. 

Chinese control

If US innovation is being stifled by oligopoly, in China it’s the suffocating power of the Chinese Communist Party that’s the problem, Frey says. 

Under the increasingly tight control of President Xi Jinping, party organisations are now embedded in the majority of private Chinese companies, says Frey. Chinese universities have cut down on collaboration with US scientists, and imposed stricter ideological restrictions on academics. 

This risks the same decline as once-leading German universities, which from the First World War increasingly appointed professors based on their patriotic and monarchist credentials, rather than scientific merit. 

Beijing is attempting massive moonshot innovation projects in areas such as biotechnology and robotics. However, it is relying heavily on state-owned enterprises to achieve its goals, rather like the Soviet Union did, Frey writes. 

China is catching up with the US on levels of R&D spending. However, as Frey repeatedly writes, during the Cold War, the Soviet Union spent more on research and boasted more scientists than the US, but still fell behind due to its rigid, centralised system that provided few avenues for new inventions to take off. 

Frey is not the only voice worrying about declining innovation in China. Others have pointed out that its venture capital market has collapsed, and the country is now producing unicorn challenger firms at a much lower rate than it did before 2021. 

And Europe? 

US and Chinese stagnation might sound like an opportunity for Europe. However, Frey is far from sanguine about European innovative power. 

EU regulations such as the AI Act and the General Data Protection Regulation create a high relative cost of compliance for new entrants, he says. European universities don’t have the autonomy to match US academic salaries. Limits on hiring and firing workers make it hard for European companies to change direction quickly. 

Frey says he would like the EU to adopt the Danish “flexicurity” model, where it’s easy to fire workers, but the state cushions their loss of income. “When technological shifts happen, you need to be able to pivot quite quickly,” he said. 

Europe does have a few key advantages, however. The first is that its governments and regulators have not been as deeply captured by lobbyists for incumbent companies. 

“Authorities in Brussels are more isolated from some of the kinds of political pressures and lobbying that they will see in the US,” he said. Europe’s quality of life shouldn’t be underestimated as an advantage either, he added. 

Is diversity strength? 

To some extent, Frey’s work challenges the Brussels orthodoxy that Europe must coordinate its investments in R&D, eliminating seemingly wasteful duplicate funding between EU member states. 

His book points to pre-unification Germany, where states competed to endow universities with money and freedom, whose technical prowess then allowed Germany to lead in areas such as chemicals during the second half of the 19th century.

Related articles

• China set to overtake US on R&D spending this year
• Viewpoint: DeepSeek’s success masks China’s weakening innovation model

The more research funding in Europe is centralised, “the more you depend on the people [. . .] assessing funding applications,” he said. 

Having “different people making different kind of bets” has “its own benefits,” he added, although this also comes with the risk of “localism,” where funding only flows to ideas nearby, rather than the best. 

Worst of both worlds

Different cultures and fragmentation have, at times, aided innovation in Europe, he said. Thinkers and inventors persecuted or blocked in one European country were able to escape to another in a way that was impossible in early modern China, Frey’s book argues. 

However, the EU is currently stuck in the “worst of both worlds,” Frey thinks.

Its single market isn’t integrated enough to allow European companies to seamlessly grow beyond national borders. But at the same time, EU regulatory harmonisation means that member states “don't really have the space to experiment that much with different institutions” that might aid innovation. 

No perfect model

Frey’s book doesn’t offer any easy answers, however. One overarching message is that there is no perfect model for maximising technological and economic progress. 

A loose, decentralised system, where inventors have lots of avenues to create new products, is best for pushing forward the technological frontier. One example is the US in the 19th century, where individual inventors patented an extraordinary range of new ideas. 

But actually deploying this new technology takes centralisation, either through concerted state action or centralised industrial conglomerations. Industrial integration is how Prussia in the 19th century managed to rival the UK, despite the British initially leading technologically. 

Instead, societies need to switch between these two models depending on whether they have reached the technological frontier, he says. 

The conundrum for Europe is that it’s at the frontier in some areas, such as car making, pharmaceuticals, chemicals and machinery, in Frey’s view. But it lags in digital technology and services. “Europe too could grow by catching up,” he says.