Mobileye moves toward operating its own robotaxis

Mobileye announced on June 16 that it plans to establish a vertically integrated robotaxi business, with a launch targeted for a U.S. city in 2027. The official announcement says the company wants to move beyond supplying Mobileye Drive, its self-driving system, to automakers and mobility partners. It now plans to own and operate an autonomous ride-hailing service as well. The first deployment is expected to use an initial fleet of about 100 vehicles, phased into a major U.S. metropolitan market and intended to operate under fully driverless conditions.

The shift matters because Mobileye has built its position mainly as a technology supplier. Its stack includes camera-based perception, EyeQ chips, mapping, driver-assistance software, and autonomous-driving systems that other companies can integrate into their own services. In the new plan, Mobileye wants to combine those assets with Moovit, its mobility subsidiary, for consumer apps, multimodal trip planning, mission control, fleet-management tools, and links to teleoperation infrastructure. Teleoperation means remote assistance that can help autonomous vehicles through defined edge cases, without turning the whole service into conventional human driving.

The announcement also includes a concrete scale target. After the initial fleet of about 100 vehicles in 2027, Mobileye says it aims to scale the business substantially, targeting roughly 17,000 vehicles over the following five years if the operating model works. That number is important because robotaxis are not just a software problem. A working service needs autonomous-ready vehicles, depots, maintenance, cleaning, insurance, rider support, city operations, remote monitoring, and incident response. It also needs local approvals, predictable service areas, and enough operational data to show where a driverless fleet can deliver regular transport rather than isolated demonstrations. Those constraints are often where autonomous mobility projects slow down after the technical demo looks convincing.

There are clear limits to the announcement. The service is planned, not live, and Mobileye has not named the U.S. city. The company also says the initiative will not change its commitment to supplying Mobileye Drive to automakers and mobility operators, which leaves a possible tension between being a neutral supplier and being an operator. Still, the signal is useful for mobile robotics: companies working on autonomy increasingly need field learning from full operations. In robotaxis, performance depends not only on onboard perception and driving policy, but also on fleet reliability, supervision, service design, regulation, and public trust. The contest is shifting toward the practical question of who can run, monitor, and maintain useful robots every day.