Mobileye puts self-driving tech in 230 million cars. Now it wants to run its own robotaxis.

For 25 years, Mobileye has been the arms dealer of self-driving, selling the cameras, chips, and software that now sit inside more than 230 million cars, while pointedly leaving the driving business to everyone else. On 16 June, it said it would start doing the driving itself.

The Jerusalem company announced plans to launch its own robotaxi service in a US city in 2027, moving from supplying autonomous-driving technology to owning and operating an autonomous ride-hailing business outright. Investors liked it: the shares rose about 6 per cent.

Mobileye plans to start small, with a fleet of around 100 vehicles in a major US metro area, phased through 2027 and running fully driverless. If that works, it wants to scale to roughly 17,000 vehicles over the following five years.

From supplier to operator

The shift matters because of what Mobileye has always been. Its Mobileye Drive system is sold as a standalone product that carmakers and mobility operators plug into their own vehicles, the neutral-supplier position that made it a partner to almost everyone.

The 💜 of EU tech

The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!

Running its own service breaks that neutrality.

It puts Mobileye in direct competition with some of the customers that license Mobileye Drive, a tension the company is clearly aware of. It frames the move as a “complementary path to market” rather than a replacement, and insists its commitment to supplying partners “does not alter”.

To pull the full stack together, Mobileye is leaning on Moovit, the trip-planning app it owns, for the consumer-facing side: booking, multimodal routing, rider engagement, and fleet operations. Moovit says it reaches 1.7 billion users across more than 3,500 cities, which is the kind of demand-side reach a new operator would otherwise take years to build.

A swipe at Waymo, and a late start

Chief executive Amnon Shashua framed the launch as a corrective to a market he says has narrowed. “The industry has become increasingly dependent on a small number of technology providers and business models,” he said, a thinly veiled nod to Waymo’s commanding lead and Tesla’s push.

The problem is that 2027 is late to make that argument. Waymo is already running hundreds of thousands of paid rides a week across a growing list of US cities, while Mobileye’s own supplier deals have had a habit of slipping, including the Verne robotaxi launch in Zagreb that switched from Mobileye to Pony.ai before going live.

The vertical-integration logic is sound, and increasingly fashionable. XPeng is making the same bet, that owning the chip, the software, and the operation produces a better robotaxi than stitching the pieces together. Mobileye owns more of that stack than most.

The hard part is the business, not the driving

The deeper question is financial. Selling chips is a high-margin, asset-light business; running a fleet of robotaxis is the opposite, a capital-hungry grind of vehicles, depots, cleaning, insurance, and teleoperators that even Waymo is still working to make pay, hence its cheaper, purpose-built Ojai vehicle.

Mobileye, still majority-owned by Intel and fresh from a roughly $900m deal for humanoid-robot startup Mentee Robotics, is taking on that burden just as it pushes into physical AI. It says it will share commercialisation and operational detail at a Capital Markets Day before the end of 2026.

The targets are bold; whether Mobileye can run a robotaxi business as well as it builds the technology behind one is the thing those numbers have to prove.