Kalanick’s Second Act Comes With a Familiar Cast

The Financial Times reports that Travis Kalanick’s holding company Atoms is building robotaxi technology after acquiring Anthony Levandowski’s Pronto and rehiring Levandowski to lead the effort. Read that again, slowly. The man who left Uber in 2017 is back in autonomous vehicles, and he brought along the engineer at the center of one of the most bitterly litigated trade secret fights in tech history.

My first reaction was not excitement. It was recognition. This is the same casting call as 2016, with better funding and worse memories.

What we actually know

The verified pieces are few, so let me lay them out plainly rather than dress them up:

• Atoms raised $1.7 billion, with reporting pointing to Andreessen Horowitz leading the round
• Uber put in $100 million
• Atoms acquired Pronto, Levandowski’s self-driving trucking outfit
• Hiring is reportedly underway
• The stated ambition is moving autonomous vehicles from pilot programs into actual service

That is the whole factual base. No fleet size, no launch city, no timeline, no vehicle partner. Which tells you something about where this story sits: it is a funding and personnel story dressed as a product story.

Why I review this the way I review tools

I spend most of my time evaluating AI toolkits, and the pattern I see over and over is that money and pedigree get treated as evidence of capability. They are not. They are evidence of belief. A $1.7 billion round means sophisticated investors think the odds justify the check. It does not mean the perception stack works in rain, or that the remote operations layer can handle a construction detour, or that unit economics close at scale.

Every toolkit I have tested that shipped on hype alone had the same tell: impressive demo conditions, vague answers about edge cases. The autonomous vehicle space has been running that playbook for a decade. Waymo eventually earned its way out of it through millions of driverless miles. Most others did not.

So the honest read on Atoms right now is: unrated. Not bad, not promising. Just unmeasured.

The Uber check is the most interesting number

$100 million from Uber is small next to $1.7 billion, but it is the line item that carries the most information. Uber spent years unwinding its own autonomous ambitions and repositioning as the demand layer that any AV operator eventually needs. Buying a stake in Atoms is consistent with that strategy: keep options open across suppliers, pay relatively little for early access, and let someone else absorb the capital burn of building the hardware and software.

For Atoms, having Uber in the cap table hints at a distribution answer to the question that kills most robotaxi efforts. Building a car that drives itself is hard. Filling that car with paying riders every hour of every day is a separate, brutal problem. Uber solved the second one already.

Whether that translates into an actual commercial agreement is not something the reporting establishes. An investment is not a contract.

The Levandowski factor

I am not going to pretend this is a neutral detail. Levandowski is one of the most technically accomplished people to work on self-driving systems, and he is also the person whose departure from Google triggered the Waymo–Uber litigation and a subsequent federal conviction that was later pardoned. Rehiring him is a deliberate signal, and reasonable people will read it in opposite directions.

If you evaluate teams on raw capability, this is a strong hire. If you evaluate them on the institutional discipline required to run a passenger service under regulatory scrutiny, it raises questions worth asking out loud. Robotaxis are not a pure engineering product. They are a permitting, safety-reporting, public-trust product. That has never been the strength of this particular group of people.

What I would watch instead of headlines

If you want to know whether Atoms is real, ignore the funding coverage and track these:

• Whether they file for driverless testing permits in any state, and what disengagement data shows
• Whether Pronto’s trucking-oriented stack gets repurposed or replaced for urban driving, which are meaningfully different problems
• Whether a formal Uber network agreement appears, not just the equity stake
• Who they hire into safety and regulatory roles, because that reveals how seriously they take the unglamorous half

None of those produce viral posts. All of them predict outcomes better than a round size does.

My take

Atoms has assembled real capital, a real technical core, and a plausible distribution partner. That combination is rare enough to take seriously. It also has essentially zero public evidence of working technology, and a leadership pairing that will draw scrutiny from regulators who remember the last decade clearly.

I would not bet against it. I would also not write a word about its capability until someone outside the company gets to measure it. In toolkit reviews, that gap between announced and verified is where most of the disappointment lives. Nothing about this story suggests it will be an exception.

🕒 Published: September 7, 2026