China Mobile and CATARC form 120 million yuan vehicle data venture

China's automotive data authority and its largest mobile carrier have put 120 million yuan (about $17 million) into a Tianjin joint venture aimed at the data layer of connected and autonomous vehicles.

The tie-up pairs CATARC's role as the industry's standards, statistics and testing centre with China Mobile's national 5G and computing network. It lands as the government's vehicle-road-cloud integration pilots move into a deeper phase and as Beijing tightens scrutiny of auto data exports and computing infrastructure.

A 60-40 split, with CATARC in control

Corporate records database Tianyancha shows CATARC Digital Technology (Tianjin) Co. was recently established in Tianjin's Hexi district with registered capital of 120 million yuan. CATARC Data, a CATARC subsidiary, subscribed 72 million yuan (about $10 million) for 60%. China Mobile Tianjin, a unit of China Mobile, subscribed 48 million yuan (about $6.8 million) for the remaining 40%.

Feng Yi, general manager of CATARC Data, is the venture's legal representative, giving CATARC's side both majority ownership and day-to-day control. The business scope is limited to core infrastructure: data processing, storage support, and big data and artificial intelligence public data platforms.

An industry data brain plus a carrier's pipes

CATARC, or the China Automotive Technology and Research Center, reports to the State-owned Assets Supervision and Administration Commission and sets much of the technical rulebook for the sector. Its wholly owned CATARC Data handles national auto industry statistics, energy consumption and dual-carbon accounting, vehicle recall data supervision, and the scenario library used for intelligent connected vehicle safety testing.

China Mobile brings the country's widest 5G-V2X base station network, plus computing network and edge computing assets. The equity link replaces the older model of loose consulting contracts and network procurement with a single entity that carries both industry accreditation and telecom backbone support.

Breaking the data silos

The immediate driver is engineering. Twenty pilot cities are now building out vehicle-road-cloud integration, and each has installed large numbers of roadside lidar units, roadside perception units and smart traffic lights. The data they generate is fragmented.

Formats differ between cities, protocols differ between automakers' onboard terminals, and data does not move across regional borders. Carmakers worry about leaking core vehicle data and hold back from local platforms. Local traffic authorities lack a unified industry-grade data standard.

The new company is meant to act as a trusted third-party relay and clearing house for that data. CATARC's perceived neutrality is intended to ease automakers' concerns, while China Mobile's private 5G networks and cloud infrastructure would handle low-latency, anonymised aggregation of roadside sensor data and vehicle trajectory data.

Compliance as the deeper motive

Underneath sits regulation. As autonomous driving algorithms shift to end-to-end large models, the industry's need for training data — video of rare corner cases and real-road point cloud data — is exploding.

At the same time, rules including the Provisions on the Management of Automotive Data Security set strict limits on exporting mapping data, in-cabin facial and voice data, and sensitive footage from outside the vehicle. Overseas, the EU's carbon border levy and new battery regulation require exported cars to carry a verifiable battery passport and carbon footprint record across the full life cycle.

Housing computing capacity inside a state-backed joint venture is intended to give domestic automakers and suppliers a high-security public data sandbox and anonymised training ground. The entity is positioned to handle cross-region transfer and rights registration for the industry's battery ID and trusted data space.

Two hard tests ahead

Carmakers' grip on their own data

The first obstacle is the resistance of private and foreign automakers to giving up control of core driving data. Huawei, Tesla and Wei Xiaoli (Nio, Xpeng and Li Auto) have already spent billions of yuan building their own supercomputing centres, such as Tesla Dojo and Alibaba Cloud-linked facilities. For companies sitting on large volumes of high-quality closed-loop data, perception and planning-control data is the core of their competitive advantage.

Administrative guidance and standards alone are unlikely to prise valuable real-time anonymised data out of these firms. The venture must design a commercial mechanism that lets automakers pay to connect, protects their trade secrets and returns long-tail scenarios of comparable value.

State R&D routines versus the pace of AI

The second test is organisational. Automotive digitalisation is not like the traditional business of type approval and drafting standards. End-to-end model training, scheduling of heterogeneous computing clusters and dynamic vehicle-road algorithms demand fast response to open-source communities and rapid product cycles.

Both CATARC Data and China Mobile Tianjin are large central SOE structures. Whether the venture can break out of compartmentalised state-owned assessment systems and build market-style incentives for algorithm engineers will determine whether the 120 million yuan produces a globally competitive AI platform for vehicles.

What the filing records

The 120 million yuan of registered capital and the shareholding split recorded on Tianyancha mark an institutional step in the industry's shift from mechanical hardware to software and data. CATARC and China Mobile have used equity to build a new hub for connected-vehicle data. Whether it can make data flow across an industry where automakers still work in their own camps is the question the venture now has to answer.