Chinese robotaxi operators find European doors opening wider

The regulatory climate for Chinese autonomous driving companies in Europe has shifted noticeably over the past twelve months, according to James Peng, founder and chief executive of Pony AI, one of China's leading robotaxi operators. Speaking to reporters on Tuesday, Peng said European authorities had become more willing to allow pilot projects, a change he described as clear and consequential for firms trying to commercialise self-driving technology beyond their domestic market.

His comments come at a moment when several European countries are racing to establish frameworks for autonomous vehicle testing, keen not to fall behind the United States and China in a sector that could reshape urban transport, logistics and automotive manufacturing. The shift Peng describes is not a wholesale welcome: it is gradual, localised and hedged with conditions. But for Chinese companies that have faced mounting scrutiny in Western markets over data security and technology transfer, even a partial opening matters.

A gradual European opening

Peng told an online media briefing that the change over the past year was tangible. Rather than a single policy decision, he described a pattern of incremental permissions: selected areas being designated for pilot operations, regulatory conversations becoming more constructive, and authorities showing greater willingness to let autonomous vehicles operate under supervision on public roads.

Europe has not been hostile to autonomous driving, but it has been cautious. The European Union's approach to vehicle automation has emphasised safety certification and type-approval standards, processes that take time and require extensive documentation. Individual member states have moved at different speeds. Germany, for instance, passed legislation in 2021 allowing Level 4 autonomous vehicles to operate in defined zones, though actual deployments have been limited. France has run pilot programmes in designated urban areas. The Netherlands and Sweden have also hosted trials. What appears to be changing is the willingness to include Chinese operators in those trials, rather than restricting them to European or American companies.

The European Commission's transport directorate has been working on a regulatory framework for connected and automated vehicles, aiming to harmonise rules across member states. Fragmented national regulations have long been cited as a barrier to scaling autonomous driving services across the continent. Peng's suggestion that the environment is becoming more open implies that some of those barriers are being lowered, at least for pilot purposes, even if full commercial deployment remains some distance away.

Pony AI's fleet expansion

Pony AI is putting numbers behind its international ambitions. The company's total robotaxi fleet reached 1,975 vehicles on 30 June 2026. Peng said the company intends to expand that global fleet to 3,500 cars by the end of the year, and that several hundred additional vehicles would be deployed on streets outside mainland China before 31 December.

The distinction between total fleet and international deployment matters. The vast majority of Pony AI's robotaxis operate in Chinese cities, where the company has built partnerships with local governments and secured operating licences across multiple municipalities, including Beijing, Shanghai, Guangzhou and Shenzhen. Its international operations are smaller and newer. Expanding outside China is both a commercial imperative and a strategic signal: Pony AI wants to demonstrate that its technology works in different regulatory environments, traffic patterns and road conditions.

The target of 3,500 vehicles by year end represents an increase of roughly 1,525 cars in six months, a pace of deployment that would be aggressive by any standard. Whether Pony AI can hit that figure will depend on vehicle procurement, local regulatory approvals and operational readiness in the cities it is targeting. The company did not specify which European locations it is focusing on.

The financial picture: growing revenue, persistent losses

Pony AI's second-quarter results, released alongside Peng's briefing, show a company that is growing quickly but remains deeply unprofitable. In the three months ending 30 June, total revenue jumped 69 per cent year on year to US$36.2 million. Robotaxi fleet sales, a subset of total revenue, surged 691 per cent to US$12.1 million, reflecting the rapid expansion of the fleet and the increasing commercial use of its vehicles.

Those growth rates are striking but need context. A 691 per cent increase in robotaxi fleet sales sounds dramatic until you consider that the base was almost certainly very small. Pony AI only began generating meaningful revenue from its robotaxi operations relatively recently, so year-on-year comparisons amplify modest absolute gains. The same applies to total revenue: US$36.2 million for a quarter is not a large figure for a company that aspires to compete with Waymo, which is backed by Alphabet's capital and operates at a far larger scale.

The net loss tells its own story. Pony AI lost US$45.4 million in the quarter, though this represented a narrowing of 14.9 per cent compared with the same period a year earlier. Autonomous driving is capital-intensive: vehicles, sensors, mapping, remote monitoring infrastructure and engineering talent all cost significant money. The path to profitability for robotaxi operators depends on achieving sufficient fleet utilisation and ride density to cover those fixed costs, something no operator anywhere in the world has yet managed at scale.

The competitive shadow of Waymo

Analysts and industry officials routinely describe Pony AI and its Chinese peers as China's answer to Waymo, the Alphabet-owned operator that is widely regarded as the global leader in self-driving taxi services. Waymo operates commercial robotaxi services in several American cities, including San Francisco, Phoenix and Los Angeles, and has been expanding steadily. Its fleet size, ride volume and technological maturity all exceed those of any Chinese competitor.

But the comparison is not purely technical. It is also geopolitical. Chinese technology companies expanding into Western markets face a level of suspicion that American firms do not. Questions about data handling, mapping restrictions and potential links to Chinese state interests have complicated the expansion of Chinese firms in sectors from telecommunications to semiconductors. Autonomous driving, which relies on detailed mapping data and real-time sensor feeds, sits squarely in that zone of concern.

Pony AI's ability to operate in Europe will depend partly on how it addresses those questions. Peng's remarks suggest he believes the regulatory mood has shifted in his favour, but a warm reception for a pilot project is not the same as permission for a full commercial service. European regulators may be willing to let Chinese operators run limited trials, watched closely, while remaining cautious about granting broader operating licences.

Why European cities want robotaxis at all

Europe's interest in autonomous driving is not purely about technological prestige. Urban congestion, air quality targets, public transport costs and the economics of taxi and ride-hailing services all create pressure to find new models. Robotaxis could, in theory, reduce the number of private cars in cities, complement public transport during off-peak hours and provide mobility in areas that conventional taxis find unprofitable.

The practical obstacles are considerable. European cities tend to be denser, older and more complex than the wide streets of Phoenix or the grid layouts of many American suburbs. Cyclists, narrow lanes, complex right-of-way rules and unpredictable pedestrian behaviour all make autonomous driving harder. Regulators know this, which is part of the reason the European approach to autonomous vehicle safety has been conservative. Opening designated pilot areas is one thing; allowing robotaxis to roam freely across a city is another.

Questions that remain unanswered

Peng's briefing left several important gaps. He did not specify which European countries or cities Pony AI is targeting for its next wave of deployments. He did not say whether the company had secured regulatory approval in any new jurisdiction, or whether discussions were still at a preliminary stage. He did not address how the company plans to handle data localisation requirements, which several European countries have tightened in recent years, particularly for companies handling geospatial information.

The company also did not break down its international revenue or fleet figures, making it impossible to assess how much of its current business comes from outside China. Given that the vast majority of its 1,975 vehicles operate on Chinese roads, the international portion is likely still small. A target of 3,500 vehicles by year end sounds ambitious, and without knowing how many of those additional cars are destined for Europe, the Middle East or other Asian markets, it is hard to judge the significance of the European opening Peng describes.

There is also the question of what happens after the pilot. European regulators have a track record of allowing limited trials while deferring decisions on permanent regulatory frameworks. The EU's legislative process for autonomous vehicle type-approval is still evolving, and final rules could impose requirements that make it harder or more expensive for Chinese operators to compete. A pilot is not a licence.