Lyft is getting ready to take a back seat, literally.

The company announced plans to buy and operate hundreds of self-driving “Robocars” through a new partnership with Silicon Valley startup Tensor Auto, signaling a major turn in its business model. The fleet is expected to hit streets across North America and Europe in 2027, pending regulatory approval.

For a company built on other people’s cars, it’s a huge move. Lyft has spent a decade acting as the middleman between riders and independent drivers. Now, it’s gearing up to own the hardware itself… an approach that could completely change how ride-hailing makes money once autonomous driving scales up.

Tensor’s vehicles, known as Robocars, look more like rolling data centers than sedans. Each one is equipped with over 100 sensors (cameras, radars, lidars) and powered by NVIDIA’s high-end AI processors capable of crunching 53 gigabits of data every second. They’ll come off the line with Lyft’s software already built in, ready to accept fares the moment regulators give the green light.

“Lyft has created earning opportunities for millions of people through its platform,” said Jeremy Bird, Lyft’s Executive Vice President of Driver Experience. “This partnership expands that opportunity by allowing vehicles themselves to generate income when owners aren’t behind the wheel.”

For Tensor, the deal is validation years in the making. Founded in 2016 and spun out of Chinese AV firm AutoX, the San Jose-based startup has focused on creating autonomous cars for private ownership… not only corporate fleets. The vehicles can drive, park, charge, and even clean themselves, which could make downtime almost obsolete.

Production begins in late 2026 through VinFast Auto in Vietnam. If the rollout stays on track, Tensor’s partnership with Lyft could make it one of the first large-scale commercial deployments of consumer-ready, Level 4 autonomous vehicles.

The Robocars’ brains run on eight NVIDIA DRIVE AGX Thor chips, giving each vehicle the processing muscle of a data center on wheels. Tensor trains its AI models on NVIDIA’s DGX platform, running millions of simulated drives to teach the cars how to handle everything from highway merges to jaywalking pedestrians.

For Lyft, not only does the deal upgrade its tech stack… it puts the company in closer competition with Uber, which already has partnerships with Waymo, Lucid, and Nuro. But Lyft’s strategy leans toward control. Instead of licensing space in someone else’s autonomous fleet, it’s positioning itself to own the fleet and the platform.

Tensor’s Chief Business Officer Hugo Fozzati called the partnership “a major step toward mainstream autonomous mobility,” saying it gives owners a way to turn their cars into revenue-generating assets while accelerating public adoption of AV technology.

Analysts at Morgan Stanley and Bernstein say Lyft’s move is the next evolution in ride-hailing… turning the business from a people network into an asset network. Instead of drivers logging in to make money, cars will. The hybrid model, where individuals own vehicles that can work on-demand for the platform, could increase margins and stabilize earnings over time.

If everything falls into place, Lyft could go from renting labor to owning infrastructure… an evolution Wall Street’s been waiting years to see. For a company once known for pink mustaches and promo codes, it’s a serious bid to prove it still belongs in the future of transportation.

At the time of publishing this article, Stocks.News holds positions in Uber and Google as mentioned in the article.