BEIJING – Geely Auto Group pulled the wraps off a purpose-built robotaxi at Auto China 2026 in late April, staking its claim in a race that already includes Baidu, Pony.ai, and a growing roster of Chinese autonomous-driving contenders. The vehicle, called the EVA Cab, was developed with Geely’s ride-hailing affiliate CaoCao Mobility and sensor partner AFARI Technology. It features what the companies describe as a 1,400 TOPS onboard computing platform and the world’s first 2160-line digital LiDAR capable of detecting objects at 600 meters.
CaoCao says it plans to bring a deeply customized version of the EVA Cab into mass production by 2027 and scale to a fleet of 100,000 driverless cars by 2030. Those are ambitious numbers, and they arrive at a moment when China’s autonomous-vehicle sector is moving faster than any other market on earth.
A small fleet already on the road
The EVA Cab is not a pure concept car. CaoCao CEO Gong Xin said in an April company statement that an initial fleet of 100 autonomous robotaxis began operating in Hangzhou at the end of 2025. On April 1, the company received approval to conduct unmanned road tests in the city, according to Gong, a step that signals at least one Chinese municipality has reviewed CaoCao’s safety case and granted permission to run vehicles without a human safety driver on public streets.
Hangzhou is a logical proving ground. The city of roughly 12 million people is a major tech hub, home to Alibaba’s headquarters, and offers the kind of dense, unpredictable urban traffic that closed test tracks cannot replicate. Every ride generates sensor and decision-making data that feeds back into the software stack, giving CaoCao a development loop that companies still stuck at the prototype stage simply do not have.
Gong framed the broader strategy around fleet asset management, arguing that the companies that win the autonomous-driving race will be the ones that master vehicle operations at scale, not just the ones that write the best algorithms.
Bold specs, but unverified
The EVA Cab’s technical claims deserve close attention and healthy skepticism in equal measure. Geely says the onboard computing platform delivers 1,400 TOPS of processing power, a figure that would place it among the most capable autonomous-driving systems announced to date. The 2160-line digital LiDAR, which the company calls a world first, is said to offer a 600-meter detection range, well beyond the roughly 200- to 300-meter range typical of current commercial LiDAR units from suppliers like Hesai and RoboSense.
None of these figures have been confirmed by independent benchmarks, teardown analyses, or third-party sensor-industry trackers. The “world’s first” label for the LiDAR has not been corroborated by competing manufacturers. Until outside testing validates the numbers, they are best understood as manufacturer-stated specifications rather than independently verified performance data.
Geely’s post-show statement described the EVA Cab as China’s first purpose-built robotaxi and linked it to the company’s L4 AI digital chassis. Both the pre-show and post-show announcements were distributed through GlobeNewswire, and no independent press coverage of the physical debut has surfaced in available reporting as of late April 2026.
Where CaoCao fits in a crowded field
China’s robotaxi landscape is already competitive. Baidu’s Apollo Go service operates paid autonomous rides in multiple Chinese cities and has logged millions of kilometers. Pony.ai, which went public on Nasdaq in late 2024, holds robotaxi permits in Beijing, Shanghai, Guangzhou, and Shenzhen. WeRide has also expanded its driverless operations across several municipalities.
What distinguishes CaoCao’s approach is its parentage. Geely is not a software startup; it is one of China’s largest automakers, with a portfolio that includes Volvo Cars, Polestar, Zeekr, and the Lotus technology brand. That gives CaoCao potential access to manufacturing scale, supply-chain relationships, and vehicle-engineering expertise that pure-play autonomous-driving firms typically lack. The EVA Cab is designed from the ground up as a robotaxi, not retrofitted from an existing passenger car, which CaoCao argues will allow tighter integration between the vehicle platform and the self-driving system.
Still, manufacturing muscle alone does not guarantee success. The global autonomous-vehicle industry is littered with missed deadlines. General Motors shut down its Cruise robotaxi unit in December 2024 after a pedestrian-dragging incident in San Francisco the previous year forced a suspension of driverless operations and exposed gaps in the company’s safety reporting. Even well-capitalized programs can stumble when real-world edge cases outpace software development.
The gap between 100 cars and 100,000
CaoCao’s stated path from 100 vehicles today to 100,000 by 2030 would require an extraordinary scaling effort. Mass production of the customized EVA Cab by 2027 implies that factory tooling, sensor supply chains, and software validation would all need to converge within roughly a year. No financial projections, supplier contracts, or plant-readiness disclosures have been published to support that timeline.
Reaching 100,000 vehicles by the end of the decade raises additional questions. Each new city would require its own regulatory approval, insurance framework, and remote-monitoring infrastructure. It remains unclear how many of those vehicles would operate in fully unmanned mode versus carrying a safety driver on board, a distinction that regulators, insurers, and passengers are likely to treat very differently.
For investors and industry watchers, the milestones to track are concrete and external: Hangzhou municipal transport filings that confirm the scope of CaoCao’s unmanned-test permit, any Ministry of Industry and Information Technology announcements covering the EVA Cab’s production approval, disclosed capital expenditure for factory retooling, and published safety-case frameworks that detail how the system handles adverse weather, sensor failures, and high-congestion scenarios.
What Geely’s robotaxi bet actually proves so far
The most grounded reading of the evidence is this: Geely and CaoCao have moved past the concept stage. They have a dedicated robotaxi platform, a small but operational fleet collecting real-world data in one of China’s largest cities, and a parent company with the industrial heft to scale production if the technology and regulations cooperate.
Whether the EVA Cab becomes a common sight on Chinese roads by the end of the decade will depend far less on press-release superlatives and far more on the slow, externally verifiable work of safety validation, factory build-out, and city-by-city approvals. The prototype is real. The ambition is enormous. The proof is still ahead.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
