Geely sets 2027 target for Eva Cab commercial rollout

Robotaxis are easy to demo and hard to industrialize. That gap is the whole story here. Geely did not just show another autonomous-driving prototype at Auto China 2026 — it showed a vehicle built from scratch for paid driverless service, then tied it to a real rollout target through its ride-hailing arm CaoCao Mobility in 2027. That matters because the industry has been full of “smart cars” retrofitted for autonomy, but much thinner on purpose-built fleet vehicles that can actually scale. (carnewschina.com) ### What is EVA Cab, exactly? EVA Cab is Geely’s dedicated robotaxi prototype — not a consumer EV with extra sensors bolted on. It was unveiled on April 24 at Auto China 2026 by Geely Auto Group, AFARI Technology, and CaoCao Mobility. The design drops the normal driver setup entirely, with no steering wheel, no pedals, and no driver’s seat, because the whole point is Level 4 autonomous ride-hailing from day one. (financialcontent.com) ### Why does “purpose-built” matter so much? A normal car turned into a robotaxi still carries the cost and packaging compromises of a human-driven vehicle. EVA Cab tries to skip that. It uses a minivan-like layout, wide sliding doors, and a (financialcontent.com)wroom. (electrive.com) ### What changed with this announcement? The new part is not just the reveal. Geely also attached a commercialization clock to it. The company said a deeply customized CaoCao Mobility version is planned for mass production in 2027, when commercial operations are meant to begin. Other coverage around the launch says CaoCao is aiming to put thousands of these vehicles on the road first, then scale much further by 2030. (carnewschina.com) ### What kind of tech is Geely packing into it? Geely is pitching EVA Cab as a high-compute, sensor-heavy platform. Reports tied to the launch describe dual Nvidia Drive Thor-U chips delivering about 1,400 TOPS, plus 43 sensors and a 2,160-line digital lidar system with long-range detection. Som(carnewschina.com)ons. The important point is simpler — Geely wants this framed as an L4-ready fleet machine, not an assisted-driving car. (goodcarbadcar.net) ### Is this still just a concept? Not entirely. Geely says pilot robotaxi operations have already been running through CaoCao in Hangzhou and Suzhou for more than a year. That does not mean full driverless service at scale is solved. But it does mean EVA Cab is being tied to an operating network with dispatch, maintenance, charging, and rider-service e(goodcarbadcar.net)t sound less like pure auto-show theater. (electrive.com) ### Where does this fit in the bigger race? China’s robotaxi field is getting crowded fast. XPeng also used the show cycle to push its own robotaxi claims, and Baidu’s Apollo, Pony.ai, and WeRide are all chasing commercial scale with different vehicle and partnership models. Geely’s angle is vertical integration — build the vehicle, co(electrive.com)et from software-first autonomy companies. (goodcarbadcar.net) ### What is the catch? The catch is that 2027 is a target, not proof. Robotaxi rollouts depend on regulation, safety validation, mapping, remote operations, and city-by-city approvals. A flashy dedicated vehicle helps, but it does not erase the hardest part — getting unmanned service running reliably and cheaply enough to beat a human-driven ride. (ca([goodcarbadcar.net)native-robotaxi-prototype-at-2026-beijing-auto-show/)) ### Bottom line? Geely is trying to move robotaxis out of the prototype era and into the fleet era. EVA Cab matters less as a concept car than as a signal that one of China’s biggest auto groups wants commercial driverless service to become a manufacturing business by 2027. (carnewschina.com)