Uber's AV strategy: own the demand, rent the driver

TechCrunch's running tracker of Uber's autonomous vehicle deals — now numbering more than 30 partnerships and investments struck over two years — reads less like a partnership list than a hedging strategy made visible. Aurora, Avride, Baidu, Cartken, Coco, Lucid, May Mobility, Mercedes-Benz, Momenta, Motional, Nissan via Wayve, WeRide, Autobrains: the roster spans continents, business models, and levels of technical maturity, and Uber has placed itself at the centre of nearly all of it, either as investor, distribution channel, fleet operations partner, or all three at once.

This is a deliberate departure from the company's first attempt at autonomy. Uber Advanced Technologies Group, founded in 2014 under Travis Kalanick with researchers poached from Carnegie Mellon, was meant to build the self-driving stack in-house. That ambition ended in a sequence of events TechCrunch recounts plainly: a trade-secrets lawsuit from Waymo, Kalanick's resignation, and the fatal 2018 Tempe crash involving an Uber test vehicle and an inattentive safety operator. Dara Khosrowshahi's subsequent reset saw ATG sold to Aurora, with Uber retaining equity rather than walking away entirely — a decision that, in hindsight, set the template for everything that followed. Uber did not abandon autonomy so much as abandon the cost and liability of building it.

The current approach inverts the ATG model. Rather than developing a driving system, Uber is assembling a marketplace in which multiple AV developers compete to fill the app with rides, while Uber collects equity stakes, order commitments, and — critically — control of the demand layer through which all these vehicles reach customers. The Lucid arrangement is instructive: a $500 million combined investment, an order scaling from 20,000 to 35,000 vehicles, and an 11% equity stake, all wrapped around Nuro's self-driving system rather than anything Uber built itself. Hertz's role in that same deal, managing charging, cleaning, and depot staffing for Lucid robotaxis, underlines how much of the actual operational burden Uber is farming out to specialists in fleet logistics — Avomo and New Horizon perform similar functions elsewhere in the portfolio.

For automotive leaders, the more interesting signal is what this implies about where value is accruing in the robotaxi stack. Uber's bet is that owning the customer relationship and the dispatch layer matters more than owning the sensors and software underneath, a position that puts real pressure on OEMs and AV developers to differentiate on cost, safety record, and operational reliability rather than brand alone. Mercedes-Benz's S-Class robotaxi collaboration with Nvidia, Hyundai's Motional partnership, and Nissan's involvement via Wayve all suggest legacy manufacturers are willing to supply hardware into someone else's marketplace rather than build their own ride-hailing front end — a reversal of the vertical integration instinct that dominated much of the industry's EV-era strategic thinking.

The list also exposes how uneven the underlying technology remains. Motional's Las Vegas vehicles and Momenta's planned European rollout still carry human safety monitors. Avride is under NHTSA investigation following more than a dozen crashes reported by TechCrunch. Baidu's promised London testing had not begun as of June despite an announcement the previous autumn. And the Cruise entry stands as a cautionary footnote: a strategic partnership announced in August 2024 that ceased to exist within months when GM dissolved Cruise's robotaxi business entirely, a reminder that Uber's diversified approach is itself a hedge against exactly this kind of collapse among individual partners.

What deserves close attention going forward is how Uber manages the inevitable consolidation among these 30-odd partners, and whether its neutral-platform positioning survives contact with an OEM or AV developer strong enough to demand exclusivity. The Munich robotaxi launch with Autobrains, explicitly pitched as OEM-agnostic, may be the clearest statement yet of Uber's long-term intent: to remain the marketplace through which autonomy is sold, regardless of whose vehicle, whose stack, or whose safety record sits underneath it.