TechCrunch Mobility: The Nvidia backup plan for China’s self-driving

Chinese autonomous vehicle developer Pony.ai and peer mobility firms are executing domestic backup hardware roadmaps as leadership turmoil at Nvidia raises new concerns over global silicon supply, according to TechCrunch's report. The shift follows the sudden resignation of Nvidia founder and chief executive Jensen Huang after admitting to an affair with an alleged spy, an event that reverberated through the automotive ecosystem during Web Summit in Portugal.

This dispatch breaks down how Chinese autonomous vehicle companies are adjusting training and inference computing stacks to withstand tighter export controls, while cataloging fresh vehicle partnerships, venture financings, and regulatory probes across the autonomous transportation market. It is written for automotive systems engineers, robotics developers, and commercial fleet operators tracking computing supply chains and autonomous platform rollouts.

01Silicon supply constraints and domestic chip adoption

Pony.ai chief financial officer Leo Wang confirmed that computing demand inside China's autonomous vehicle sector currently outstrips available semiconductor supply. In response to potential tightening of cross-border trade restrictions following the executive shakeup at Nvidia, the company is directing software engineers to optimize deployment efficiency across existing clusters. Development teams are actively routing AI training jobs through midtier compute resources rather than relying exclusively on high-end processing hardware.

Chinese autonomous vehicle developers have spent several years preparing contingency computing frameworks to insulate their pipelines from hardware cutoffs. Pony.ai is increasingly turning to semiconductors manufactured inside China to help close the compute deficit for AI model training. Wang noted that while domestic silicon alternatives may not yet achieve the peak cost-effectiveness of incumbent architectures, the maturity of domestic chips has progressed significantly compared to prior years.

02Full stack automotive processors and compute reliance

Nvidia functions as a foundational technology provider across the autonomous driving ecosystem, though operational reliance differs among major fleet developers. Several companies deploy Nvidia automotive processors inside vehicle chassis to run real-time perception and automated driving stacks, whereas others consume Nvidia data center GPUs strictly for offline model training. Full-stack commercial partnerships incorporating Nvidia technology include programs with Mercedes-Benz, Zoox, and Wayve.

The sudden departure of Jensen Huang removes a high-profile proponent of open-source artificial intelligence from American corporate technology leadership. The broader industry implications of his resignation arrived as Web Summit co-founder Paddy Cosgrave highlighted rising competition from Chinese founders in opening conference remarks. Autonomous vehicle teams across Asia are assessing whether subsequent regulatory or export enforcement will further limit GPU shipments to international research facilities.

03Global platform deployments and mobility funding rounds

Commercial autonomous fleet operators announced several production vehicle and capital agreements across Europe, Japan, and the United States. May Mobility entered an agreement to deploy Volkswagen all-electric ID.

Buzz vehicles for its international robotaxi service, targeting more than 10,000 automated vehicles through 2031. Volkswagen indicated that its MOIA unit continues working with Mobileye while concurrently offering automated driving-ready base vehicles, a framework the automaker showcased at Nvidia GTC in Berlin.

International mobility startups secured substantial funding across municipal transit and logistics markets. Japanese robotaxi startup Newmo closed a 20.8 billion yen financing round, roughly equivalent to 135 million dollars, led by telecommunications firm KDDI. In urban logistics, delivery startup Port raised 14.5 million euros in an investment round backed by the European Bank for Reconstruction and Development and Balnord, while Dutch parking technology provider ParkBee completed the acquisition of Mobian city and parking assets.

04Electric vehicle finances and regulatory safety investigations

Hardware manufacturers and autonomous operators faced independent financial constraints and federal regulatory scrutiny over automated vehicle software. Electric vehicle manufacturer Lucid reported third-quarter revenue of 350 million dollars, missing Wall Street estimates, while narrowing its quarterly net loss to 830.9 million dollars. To support ongoing manufacturing operations, Lucid increased its delayed draw term loan from Saudi Arabia's Public Investment Fund to 1.75 billion dollars.

The National Highway Traffic Safety Administration opened an inquiry into a year-old Waymo software recall involving autonomous vehicles improperly passing stopped school buses. Regulators noted that a majority of reported violations occurred after Waymo deployed an over-the-air update intended to correct the driving behavior. Meanwhile, Tesla CEO Elon Musk confirmed Cybertruck production will continue through next summer before ending, as Tesla expands robotaxi testing into San Antonio, Texas.

05Enterprise ride hailing updates and agent access bans

Consumer transportation networks rolled out service modifications alongside security disclosures and travel distribution restrictions. Uber introduced an in-app service enabling passengers to reserve and pay for parking across major United States cities, arriving nine months after purchasing parking reservation platform SpotHero. Simultaneously, Uber began notifying thousands of drivers that their personal records and driver's license numbers were compromised in an October breach at an external technology vendor.

Commercial airlines are establishing technical firewalls to prevent unauthorized autonomous artificial intelligence software from accessing booking infrastructure. United Airlines announced policy changes that restrict artificial intelligence agents from directly scheduling passenger flights or managing travel reservations on behalf of users. The shift establishes explicit operator boundaries against automated transactional bots across commercial ticketing interfaces.