The Aggregator of Autonomy
Uber just committed $1.25B to Rivian’s robotaxis. Plus Zoox, Waymo, Lucid, Nuro, May Mobility — 25+ autonomous vehicle partners. Uber doesn’t build the robots. It owns the network they all depend on.
The announcement was careful to include a quote from Dara Khosrowshahi about Rivian’s technology approach:
“We’re big believers in Rivian’s approach — designing the vehicle, compute platform, and software stack together, while maintaining end-to-end control of scaled manufacturing and supply in the US. That vertical integration, combined with data from their growing consumer vehicle base and experience managing the complexities of commercial fleets, gives us conviction to set these ambitious but achievable targets.”
Dara is praising Rivian’s vertical integration. Every part of that sentence is about how Rivian owns the full stack.
The irony: Uber is the single most horizontally oriented player in the entire autonomous vehicle industry. No hardware. No software. No maps. No fleet. Just a network and a brand.
The CEO of the horizontal aggregator is praising the vertical integrator’s approach. What he doesn’t say out loud: that’s why we’re funding you, not building it ourselves.
The list of Uber’s autonomous partners as of today:
Waymo. Zoox. Rivian. Lucid (with Nuro autonomy). May Mobility. Momenta. Baidu. Wayve. Weride. Motional. Plus a dozen more in various cities and stages.
25+ partners. Approximately $0 of Uber’s own autonomous driving IP.
This is not a failure of ambition or capability. This is a strategy. And if it works, it’s a more defensible strategy than building the technology yourself.
The history of Uber and self-driving is instructive. From 2015 to 2020, Uber ran its own Advanced Technologies Group — the team that would become notorious for the 2018 Tempe fatality, the Anthony Levandowski trade secret lawsuit (a $178M settlement with Alphabet), and the eventual sale of the entire unit to Aurora in 2020 for $4 billion of stock that was worth much less by the time it vested.
Uber spent five years trying to be a technology company in the most capital-intensive segment of transportation technology, generated enormous legal liability, killed someone, and sold the whole thing.
The lesson Dara drew from this isn’t that autonomous vehicles are impossible. It’s that Uber’s competitive advantage has nothing to do with engineering. Their advantage is demand. 150 million monthly active consumers. An app on every phone. A pricing algorithm that maximizes utilization. The distribution network.
The company that wins the technology race may not be the company that profits from the technology wave. Those are different games.
The exclusivity terms in the Rivian deal are worth reading carefully. “The fleet will be exclusively available on Uber’s network.” 50,000 Rivian robotaxis, available only through Uber.
Last week, Zoox: also deploying exclusively through Uber.
This is the key asset. Not the investment dollars. Not the technology roadmap. The exclusivity agreements. If every robotaxi that achieves meaningful scale signs an exclusivity deal with Uber, then Uber becomes the unavoidable distribution layer for autonomous transportation. Not because they built better software, but because they locked up supply.
It’s a land grab disguised as a portfolio investment strategy.
The counterargument is Waymo.
Waymo has been building L4 autonomous vehicles for fifteen years. They have the most miles driven, the best safety record, the most refined hardware-software stack. Their cars work in Phoenix and San Francisco without human operators. They are not waiting for milestones.
Waymo does not need Uber. They have their own app. They have their own fleet. They have Google’s balance sheet and data infrastructure. When I think about who doesn’t need Uber’s distribution layer, Waymo is at the top of the list.
And indeed, Waymo operates on Uber — but it’s a partnership of equals (or close to it), not the dependency relationship Uber has with, say, Rivian or Zoox. Rivian is years away from deploying a robotaxi. Uber’s funding is essential to their timeline. That’s leverage.
The Waymo relationship is different. If Waymo decided to go direct-to-consumer in San Francisco, they could. The fact that they choose to partner with Uber is a business decision, not a dependency.
Uber’s strategy works perfectly if the technology continues to be fragmented — if there’s no single dominant robotaxi technology that can afford to walk away from the distribution network. The moment one player achieves overwhelming technological superiority and a brand-name that consumers trust, the dynamic flips.
There’s a historical parallel that makes Uber nervous.
In the early streaming wars, Netflix needed content studios to provide the movies and shows. The studios needed Netflix to reach streaming audiences. It looked like a healthy interdependence.
Then Disney+ launched. Disney had the brand, the IP, the direct consumer relationship, and forty years of emotional lock-in. They didn’t need Netflix. And they proved it by pulling their content and building their own platform.
For Uber, the Disney+ scenario is: one robotaxi company achieves such brand recognition and technological superiority that consumers start asking for it by name. They launch their own app. They price their rides slightly lower. And they use the direct consumer relationship to bypass the aggregator entirely.
Waymo is the most likely candidate. Apple Maps data shows people in Waymo-operating cities already specifically seek out Waymo rides rather than generic Uber-dispatched robotaxis. The brand is emerging.
Dara’s bet is actually quite clear: technology will stay fragmented, no single player will own the market, and everyone will need Uber’s network to achieve meaningful scale. The 25-partner portfolio is a hedge against any one company going Disney+.
If that fragmentation bet is right, Uber doesn’t need to win the technology race. They just need to own the demand side while everyone else wins the technology race on their behalf.
If it’s wrong — if Waymo or some successor achieves decisive dominance — Uber’s portfolio of milstone-contingent investments in the second tier becomes a lot less valuable.
50,000 Rivian robotaxis by 2031, exclusively on Uber. That’s the headline. The subtext is a different question:
In the autonomous transportation future, who owns the relationship with the passenger?
Uber’s bet: the aggregator does. The network effect is the moat, not the robot.
The counter-bet: the brand that passengers trust with their physical safety owns the relationship. And in safety-critical mobility, brands matter more than apps.
Both could be right, in different segments, for different kinds of trips. The boring short-trip market might aggregate. The premium long-distance market might go direct.
In the meantime, Uber has 25+ partners racing to build the technology that will eventually test whether its aggregator strategy holds. They’ve turned the competitive landscape of autonomous vehicles into a laboratory for their distribution theory.
It’s the most leverage-efficient position in the industry. Whether it’s the winning one is what the next decade will determine.