Travis Kalanick's Atoms Eyes Robotaxis — With Uber's $100 Million and His Old AV Team

Travis Kalanick's physical AI startup Atoms may be getting into the robotaxi business, TechCrunch reported this weekend — setting up one of tech's stranger full-circle stories: the man who bet Uber's future on autonomy a decade ago, lost the company before it paid off, and now returns to the same fight with $1.7 billion in fresh capital.

The pieces on the board

Atoms, which Kalanick launched in March to apply AI, sensors and robotics to industries that "make, move and store" materials — food, mining, transportation — raised $1.7 billion earlier this year in a round led by a16z, with Ben Horowitz joining the board. Notably, Uber itself invested $100 million.

According to the report, Atoms is now recruiting veterans of Uber's old Advanced Technologies Group and preparing a hiring spree for autonomous-driving talent. The company has held early talks with Uber about deploying its technology for robotaxis on the ride-hailing network — though sources stress robotaxis are only part of Atoms' plans.

Why now

The robotaxi market Kalanick left is unrecognizable. Waymo operates commercially across US metros; Tesla's Cybercab launched in Austin (and promptly drew a federal probe); Baidu's Apollo Go and Pony.ai are scaling across China; Zoox is working through recalls. What no Western player has solved is Uber's old dream: a demand network that doesn't own the cars, married to autonomy it doesn't have to build alone.

That's the seam Atoms could fill. Uber has assembled a coalition of AV partners but lacks a captive technology bet of its own — and its $100 million check into its estranged co-founder's startup suddenly looks less sentimental and more strategic.

The irony, priced in

Kalanick has said Uber's decision to sell its AV unit was the mistake that ceded the decade to Waymo. Rebuilding that capability outside Uber, with Uber's money, and selling it back to Uber's network may be the most Kalanick maneuver imaginable. Investors at a16z are effectively underwriting the thesis that the robotaxi endgame still needs a network — and that the person who built the network once can do the harder half too.