WeRide Inc.

Q2 revenue was RMB 232 million, up 82% year over year and 103% quarter over quarter. Gross profit was RMB 87 million, up 143% year over year, and gross margin was 37.5% versus 28.1% a year ago. Net loss narrowed 1% year over year to RMB 401 million, and EBITDA loss narrowed 8.1% year over year to RMB 335 million. By segment, L4 revenue was RMB 125 million, up 47% year over year, while L2+/L3 revenue increased about 26x year over year and 219x quarter over quarter. Overseas revenue increased 164% year over year and about 170% quarter over quarter, and overseas markets accounted for nearly 40% of group revenue. As of June 30, cash and other liquid financial resources were about RMB 5.4 billion. Forward-looking commentary included an expectation that L4 projects in the second half would contribute more as deployments take place, L2+/L3 vehicles should exceed 100,000 by year-end and surpass 0.5 million cumulative deliveries next year, and the company expects positive cash flow in a single quarter by 2028 and full-year breakeven in 2029.

Tony Han framed the quarter around three themes: overseas acceleration, asset-light scaling, and a path to self-sustaining cash generation. He emphasized a “physical AI” strategy in which L4 robotaxi data and L2+/L3 data reinforce each other through the WIT and Genesis models, and he argued this creates a technology flywheel. His tone was highly confident and assertive, repeatedly stressing regulatory advantages, safety validation, and that the company is ahead in both L4 and ADAS.

Jennifer Li said Q2 was an inflection point driven by stronger overseas business, scaling of the asset-light L4 model, and rapid commercialization of L2+/L3. She highlighted the key financial metrics: RMB 232 million revenue, RMB 87 million gross profit, 37.5% gross margin, RMB 401 million net loss, and RMB 335 million EBITDA loss, plus RMB 5.4 billion in cash and liquid resources. She also pointed to operating leverage, noting total operating expenses of RMB 533 million were up 9.2% year over year, below revenue growth, while R&D rose 36% to RMB 434 million and G&A fell to RMB 69 million.

Analysts focused on resource allocation across L4, L2+/L3, and AI infrastructure, overseas scalability, ADAS competition, the risk of OEMs building robotaxi systems from ADAS data, and the durability of the asset-light partnership model. Management’s answers consistently stressed the shared data and model infrastructure across businesses, saying the same foundation models and data flywheel support all three areas. On overseas operations, they said the model is repeatable because WeRide works with local and platform partners, does not own the vehicles, and relies on regulatory permits and safety records as the real moat; they also said the company has autonomous-driving licenses in 8 countries and operates in 12 countries. On Mercedes, management said contractual limits prevented detailed disclosure, but argued that strong engineering and execution can beat better-funded rivals.