Key Highlights

  • More than 70 Chinese physical AI firms have set up operations abroad, with another 20 preparing to follow.
  • A ten-kilometre taxi ride in Wuhan costs about 3.40 dollars, compared with 54 dollars for a similar ride in St Gallen, Switzerland.
  • China's overall service exports doubled to 510 billion dollars a year between 2015 and 2025, after adjusting for inflation.
  • A Chinese court ruled in April in favour of a worker who sued his employer for losing his job to AI.

Chinese Robotaxi Firms Outpace American Rivals in Global Expansion

When driverless taxis begin operating in the cantons of eastern Switzerland early next year, the technology behind the wheel will be Chinese, following a partnership between Swiss public-transport company PostBus and Apollo Go, a robotaxi operator owned by Chinese tech giant Baidu. This project is one of dozens that have emerged over the past two years in which passengers across countries including Singapore, Slovakia and Saudi Arabia are riding in vehicles powered by Chinese hardware and data.

China's physical artificial intelligence firms are expanding overseas at a pace far exceeding their American counterparts. More than 70 Chinese companies have established operations abroad, with another 20 or so preparing to follow, according to research firm EqualOcean. Many, including autonomous logistics group Neolix, already operate across dozens of countries, while American rivals such as Alphabet-owned Waymo have focused primarily on cracking the domestic market and are only beginning to expand internationally.

Thin Domestic Profitability Pushes Firms Toward Foreign Markets

Chinese firms are eager to export partly because profitability at home remains limited. Within China, these companies have been able to scale their projects and reduce the unit cost of expensive hardware, while taxi operators have collected real-time data to improve operational efficiency. Yet the underlying economics still do not work well domestically, largely because an oversupply of cab drivers, often newly unemployed workers turning to driving, keeps fares persistently low. Baidu, which said in 2025 that its robotaxis in the central city of Wuhan had broken even, has not repeated that claim since, while rival Pony.ai claims to have achieved breakeven in Chinese cities.

Foreign markets appear considerably more attractive, particularly where developers can deploy their low-cost Chinese technology and data in places willing to pay substantially higher fares. A ten-kilometre taxi ride in Wuhan costs around 23 yuan, or approximately 3.40 dollars, compared with a fare of 44 Swiss francs, or about 54 dollars, for a similar ride in St Gallen, Switzerland. This fare arbitrage likely explains why Chinese companies have increasingly shifted focus toward international expansion rather than domestic operations, with Baidu no longer publishing updates on its Chinese fleet performance while continuing to share information about its foreign ventures.

Broader AI Services Face Similar Domestic Profitability Challenges

Providers of AI services beyond robotaxis face comparable struggles achieving profitability within China. While some consumers in Western markets pay for subscriptions to services such as ChatGPT or Claude, Chinese consumers have generally been unwilling to pay for chatbot access, and Chinese businesses are notably conservative with technology spending, allocating less than a tenth of what American counterparts typically spend on IT budgets.

The Chinese government has its own reasons to encourage greater AI exports. The stability-focused Communist Party wants to avoid any perception that AI is displacing domestic jobs, given the risk of social unrest, even as it works to close the broader AI capability gap with America. Both AI service providers and their customers may fear legal repercussions if job losses become publicly linked to AI adoption, and while Western companies have openly promoted, sometimes with exaggerated claims, the payroll savings AI can generate, Chinese firms have remained notably quiet on this front, with insiders suggesting some technology companies are indeed reducing programming staff, among the first roles affected by advancing coding tools, without publicising the trend.

Regulatory and Legal Pressure Reinforces the Export Strategy

Robotaxi deployment within China has not yet resulted in significant job losses, but even the suggestion of replacing human drivers with autonomous vehicles has triggered small protests, prompting some local regulators to stop issuing licences for new self-driving fleets. In April, a court ruled in favour of a worker who sued his employer after losing his job to AI, and a national oversight committee has separately stated that companies should not deploy the technology with the explicit aim of replacing human workers.

One academic observed that AI service exports have become the primary channel through which China can externalise the economic and social costs associated with the technology's rapid domestic expansion. Like Chinese goods, AI services may still face resistance abroad given the regulatory and compliance complexities involved, a point one Chinese AI industry executive has raised publicly. However, unlike China, many Western markets face native worker shortages, including among taxi drivers who are frequently immigrants, meaning Chinese AI services could in some cases be more readily welcomed rather than resisted.

Setting Global Standards Through First-Mover Advantage

Exporting AI services offers China multiple simultaneous benefits: companies generate revenue, the party avoids domestic discontent tied to job displacement, and the broader focus on foreign markets allows China to preserve its export-led growth model even as global goods markets become saturated with Chinese products and more countries erect trade barriers. Overall Chinese service exports doubled to 510 billion dollars annually between 2015 and 2025 after adjusting for inflation, considerably outpacing the roughly 40% growth in the far larger category of goods exports over the same period, with AI-related services, spanning robotaxis, cloud computing and chatbots, expected to comprise a growing share of this expanding trade. Chinese President Xi Jinping said at a Shanghai conference on July 17th that AI development should be a symphony of international cooperation, endorsing open-weight systems and positioning China as a provider of AI as a global public good.

China could plausibly emerge as a leader in AI-embodied exports over time. Alongside the services themselves, Chinese firms are also effectively exporting technological standards, the widely accepted protocols governing how various components and systems interoperate, an area historically dominated by Western countries but often undefined for newer AI services such as robotaxis. When Chinese companies become the first to establish robotaxi operations in a given country, they effectively become the de facto standard-setters, according to one academic, suggesting that as American competitors seek greater international expansion of their own autonomous vehicle technology, they may increasingly need to adapt their systems to align with Chinese-established norms already in place.

Conclusion

China's rapid overseas expansion in robotaxis and broader AI services reflects a confluence of economic necessity, given thin domestic profitability, and political calculation, given the Communist Party's sensitivity to AI-related job displacement concerns at home. Whether China can translate this early export lead into lasting technological standard-setting influence, particularly in markets where regulatory frameworks for autonomous AI services remain undefined, may prove an important factor in determining the broader trajectory of global AI competition in the years ahead.