At the Beijing auto show in late April 2026, DeepRoute.ai CEO Maxwell Zhou stepped in front of reporters and offered a number meant to signal that his company had hit a new gear: more than 300,000 vehicles on Chinese roads now carry DeepRoute.ai’s advanced driver-assistance technology. If accurate, that figure would represent a threefold jump from the 100,000 cars the Shenzhen-based startup claimed just seven months earlier.
The statement, first reported by Reuters, landed during Auto China 2026, the biennial showcase where Chinese automakers and their technology partners compete for attention and investment. But the milestone rests almost entirely on company disclosures. No independent registration data, third-party audit, or OEM production tally has surfaced to confirm it.
The numbers DeepRoute.ai has put on the record
DeepRoute.ai’s growth narrative follows a clear arc through its own press releases. When the company unveiled its DeepRoute IO 2.0 platform at IAA 2025 in September, it said more than 100,000 passenger cars were already running its autonomous driving stack, a system built around a vision-language-action model and marketed to global automakers as a turnkey city-driving solution.
By November 2025, the company raised its public target. A separate announcement pegged DeepRoute.ai as on track to deliver 200,000 equipped vehicles by year-end, with Zhou claiming roughly 40% of China’s third-party city navigation-on-autopilot (NOA) market for October 2025. Chinese-language reporting from the Guangzhou auto show echoed that figure, attributing the market-share estimate to an unnamed “authoritative institution.” No specific research firm or industry body has been publicly identified as the source, which makes the percentage difficult to verify.
Taken together, these disclosures create a paper trail: 100,000 in September 2025, a 200,000 target for December 2025, and now 300,000-plus announced at Auto China 2026 in April. The trajectory is steep but not implausible for a supplier embedded in the production lines of multiple Chinese automakers during a period when city NOA adoption has accelerated across the industry.
Where the claims get harder to pin down
The jump from 200,000 to 300,000 in roughly four months has not been corroborated by any source outside DeepRoute.ai’s leadership. Reuters relayed Zhou’s statement as a claim made at the show, not as a figure the wire service independently checked against vehicle registration databases or automaker records. Gasgoo, a Chinese automotive news service covering the same event, used softer language, describing “nearly 300,000 production vehicles” with city NOA. The gap between “over” and “nearly” is small, but it hints at the imprecision that can creep in when reporters transcribe live remarks, sometimes through translation.
Zhou’s forward-looking projection adds another wrinkle. Sina Finance reported that he expects one million additional vehicles to be equipped in 2026, which would push the total fleet past 1.3 million by year-end. DeepRoute.ai’s own late-2025 press release said deployments were “expected to reach one million in 2026,” phrasing that could mean one million total or one million new additions. No OEM partner has publicly confirmed production volumes large enough to support either reading.
Safety statistics introduced at the show also lack transparency. Gasgoo reported that Zhou said DeepRoute.ai’s system “prevented over 180,000 potential collisions in the past year.” That is a striking number, but the company has not disclosed how it defines a “potential collision,” what sensor thresholds or driver-intervention events trigger the count, or whether any external body reviewed the methodology. Without those details, the figure functions as a marketing highlight rather than a verifiable safety metric.
The competitive picture Zhou did not dwell on
DeepRoute.ai is not operating in a vacuum. China’s third-party ADAS market has become one of the most contested spaces in the global auto industry. Huawei’s Advanced Driving System (ADS) powers city NOA features in vehicles from Chery, BAIC, JAC, and Seres, and Huawei has claimed its own rapidly growing deployment numbers. Momenta, backed by investors including Mercedes-Benz and Toyota, supplies highway and city NOA to multiple Chinese brands. Horizon Robotics, which went public on the Hong Kong Stock Exchange in late 2024, provides the underlying computing chips and software for a wide range of automakers.
Against that backdrop, DeepRoute.ai’s 40% market-share claim for third-party city NOA deserves scrutiny. The figure depends heavily on how “third-party” and “city NOA” are defined. Automakers like Xpeng and Li Auto develop much of their autonomous-driving software in-house, so they may fall outside the comparison set entirely. If the addressable market is narrowly drawn, a 40% share is easier to achieve but less meaningful as a measure of industry influence.
Geely Auto Group’s presence at Auto China 2026, which included the debut of what it calls China’s first purpose-built robotaxi, underscores how crowded the field has become. The Beijing show served as a stage not just for DeepRoute.ai but for an entire ecosystem of suppliers, chipmakers, and automakers racing to define the next generation of assisted and autonomous driving in China.
The robotaxi question
DeepRoute.ai outlined plans in late 2025 to launch robotaxi operations using consumer-grade production vehicles by year-end. That timeline has now passed, and no public evidence of a commercial robotaxi service has emerged in the company’s press materials or in independent reporting from the Auto China 2026 event. Zhou’s remarks at the Beijing show focused on fleet size and collision-prevention statistics, not on ride-hailing revenue or regulatory approvals for driverless operations.
The distinction matters. A fleet of 300,000 cars with driver-assist features embedded in OEM models is fundamentally different from a network of paid robotaxis operating without human backup drivers in dense urban traffic. The former requires automaker partnerships and consumer adoption; the latter demands regulatory permits, insurance frameworks, and a far higher bar for safety validation. If DeepRoute.ai’s robotaxi ambitions have been delayed or scaled back, that context would reshape how investors and analysts interpret the company’s growth story.
What the 300,000 number is really worth
For readers trying to gauge DeepRoute.ai’s position, the honest answer is that the company appears to have real momentum, but the data trail remains opaque. It is reasonable to conclude that the installed base has grown substantially since the 100,000 mark cited at IAA 2025 and that partnerships with major Chinese automakers are driving higher volumes. Whether the true count is closer to 250,000, 300,000, or somewhere else is not something any outside observer can confirm today.
More granular disclosures would change that. Breakdowns by automaker, vehicle model, and city of deployment, or anonymized registration data verified by a neutral auditor, would give the company’s narrative the kind of backbone that investors and regulators can actually lean on. Until then, the 300,000 figure and the 180,000 collision-prevention statistic are best understood as directional signals from a company that wants the market to know it is growing fast, offered at a venue designed to maximize attention, during a period when every autonomous-driving supplier in China is fighting to prove it has scale.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
