CaoCao (Geely) — First Hangzhou Driverless Authorisation Plus China's First Purpose-Built Robotaxi at Auto China 2026

Headline Finding

Two CaoCao announcements in window reframe the Chinese robotaxi competitive structure. On 1 April 2026, CaoCao (Geely Auto Group’s ride-hailing arm) became the first company authorised to conduct unmanned road testing in Hangzhou. On 21 April CaoCao announced — and on 25 April debuted at Auto China 2026 (Beijing) — China’s first purpose-built robotaxi: a Geely-built, deeply customised vehicle with mass production planned for 2027 and a stated 100,000-vehicle driverless fleet target by 2030. The pairing matters because CaoCao is a ride-hailing operator (not a pure-play AV stack developer like Pony.ai or WeRide), and a vehicle OEM affiliate (Geely is its parent). It is the first Chinese deployment example that matches the Western “vertically integrated operator + OEM + capital” model that has been characterising Waymo and Tesla.

Detail

• Hangzhou driverless permit (April 1, 2026). CaoCao becomes the first company in Hangzhou authorised for driverless road testing. Hangzhou fleet has reached 100 vehicles with 3,600+ virtual pick-up/drop-off points already in operation prior to the permit upgrade.
• Auto China 2026 announcement (April 21, 2026). Geely Auto Group announces it will unveil “China’s first purpose-built native-developed Robotaxi prototype” at Auto China 2026 (Beijing International Automotive Exhibition).
• Vehicle debut (April 25, 2026). Purpose-built robotaxi debuts at Auto China; mass production planned for 2027; commercial operations to follow.
• Fleet target. 100,000 driverless vehicles by 2030 (CaoCao stated). For comparison: Waymo’s stated 2026 fleet expansion is “thousands”; Pony.ai’s Dubai target is “hundreds” by end-2026 scaling to 1,000 by 2028; WeRide’s Middle East target is 500–1,000 in 2026 (EV72).
• Capital structure. CaoCao trades on Hong Kong (listed July 2024). Geely Auto Group is the controlling shareholder and provides vehicle platform, manufacturing, and capital. Reportage frames asset management — i.e., the disciplined deployment of fleet capital across operating cities — as the key strategic question for Chinese AV operators, distinct from the technology-platform framing that dominates Western coverage.
• International significance — “ride-hail operator + OEM” template. CaoCao is the first Chinese operator with both a ride-hailing licence and a parent OEM in-house. WeRide (no parent OEM, Uber-distributed), Pony.ai (no parent OEM, Toyota-affiliated), Apollo Go (Baidu’s BU, no OEM) all rely on third-party hardware. The CaoCao model parallels Waymo (Alphabet capital + Jaguar/Hyundai OEM partnership) and Tesla (own OEM + own ride-hail) but is the first example outside the US.

Governance Implication

EV79 is a Ch.2 (What is CAM) item: it expands the international structural map and sharpens the Ch.1 capital-concentration argument. Three implications.

(1) The CaoCao announcement breaks the Western/Chinese model dichotomy the paper has been using. Previous Ch.2 treatments of Chinese robotaxi expansion (EV59, EV60, EV72) have treated Chinese operators as technology platforms (Pony.ai, WeRide) deploying via local transport-authority partners. CaoCao is structurally different: a ride-hailing-licensed operator with an OEM parent and an asset-management framing of the deployment problem. This is the closest structural analogue to the Tesla and Waymo models seen outside the US. The paper’s Ch.2 should reflect this: there are now three deployment archetypes — pure technology platform (Pony.ai, WeRide, Aurora), vertically-integrated operator-OEM (Tesla, Waymo, CaoCao), and OEM-stack-supplier (Wayve via Nissan, Mobileye via VW/MOIA).

(2) Hangzhou as a regulatory laboratory. Hangzhou’s first-mover authorisation of driverless testing for a domestic Chinese champion (rather than for foreign or non-local operators) creates a regulatory-capture-by-association precedent. For comparative work, Hangzhou should be tracked alongside Phoenix, Austin, and Abu Dhabi as cities where the local government’s first AV authorisation went to a deeply local operator. The Institutional ODD treatment (Ch.3) should note that the identity of the first authorised operator shapes the city’s downstream regulatory posture.

(3) The 100,000-vehicle 2030 target reframes the international fleet concentration question. A 100,000-vehicle fleet, if achieved, would exceed the combined global stated targets of Waymo, Pony.ai, WeRide and Apollo Go. Even a 25% achievement (~25,000 vehicles) would equal the Waymo target. Whether CaoCao reaches this scale is uncertain (Chinese AV fleet targets have a history of being aspirational), but the announcement signals that Chinese operator-OEM combinations now intend to compete at a scale that makes the Western capital-concentration thesis (EV75) regional rather than global. The paper’s Ch.1 Pacing Problem should treat this as a plausible upside scenario for fleet-pace, distinct from the capital-pace argument.

Sources

• GlobeNewswire, 21 April 2026: “China’s First Purpose-built Autonomous Robotaxi Prototype Coming Soon; CaoCao’s Deeply Customized Version Scheduled for Mass Production in 2027”
• GlobeNewswire, 23 April 2026: “CaoCao’s Robotaxi Hosts 1,000 International Guests, Showcasing Safe, Stable, and Scalable Operations”
• GlobeNewswire, 25 April 2026: “China’s First Purpose-Built Robotaxi Customized for CaoCao Debuts at Auto China 2026, with Mass Production Planned for 2027”
• Swace News / Miami Times Now / Winger Daily / Enviro Magazine, 16 April 2026: “CaoCao: Asset Management to Decide Autonomous Driving Race, Targets 100k Driverless Cars by 2030”
• Hong Kong Stock Exchange filings — CaoCao Inc. (HKEX: 02643) listing background
• Related: EV59 (Chinese operators international expansion); EV60 (Pony.ai Zagreb); EV72 (Chinese robotaxi Gulf deepening); EV75 (AV capital concentration Q1 2026)