Uber’s Autonomy Roulette: Portfolio Genius or Strategic Chaos?
Uber walked away from building its own self‑driving cars years ago — and yet it keeps inching back toward the autonomous future. The latest move: a billion‑dollar, long‑term tie‑up with Canadian startup Waabi that is supposed to put tens of thousands of robotaxis on Uber’s platform. For riders and drivers, this isn’t some abstract bet on AI; it’s about who controls urban mobility — and who still has a job — a decade from now. In this piece, we’ll unpack what Uber is really buying with Waabi, why the “bet‑on‑everything” strategy is risky, and what it means for Europe.
The news in brief
According to TechCrunch’s Equity podcast, self‑driving startup Waabi has secured a funding package worth up to $1 billion that significantly deepens its relationship with Uber. The deal reportedly consists of $750 million in immediate funding, plus an additional $250 million from Uber that is tied to deployment milestones.
Waabi, founded by former Uber AI chief Raquel Urtasun, originally focused on autonomous trucking. The new agreement marks a substantial expansion into robotaxis, with a stated plan to deploy more than 25,000 self‑driving cars on Uber’s network over time. TechCrunch notes that Uber already works with more than 20 autonomous vehicle partners globally, and this deal represents another major bet in that portfolio approach. On the podcast, the hosts discuss Uber’s partnership strategy and why Waabi’s “simulation‑first” development model could differentiate it from earlier autonomous efforts.
Why this matters
Uber is quietly reinventing its autonomy strategy from “we will own the stack” to “we will own the demand.” That shift matters more than any single funding headline.
By backing Waabi while already working with more than 20 other AV partners, Uber is effectively turning autonomous driving into a supplier market. It wants to become the Android of urban mobility: the distribution layer that every self‑driving company must plug into if it wants scale. Waabi gets capital, access to riders, and a fast lane to commercialization; Uber gets another option on the future without taking on the brutal capex of building vehicles and safety programs itself.
Beneficiaries are clear:
- Uber gains leverage. If one AV partner stumbles, it can simply route demand to another.
- Waabi gets validation, cash, and real‑world data at a scale few independents can dream of.
- Investors receive a cleaner story: AVs as B2B infrastructure rather than capital‑hungry end‑to‑end mobility brands.
But there are losers too.
Human drivers face a slow‑motion squeeze. Uber will not rip out driver income overnight — politically and operationally that would be suicidal — but every robotaxi mile that becomes cheaper than a human‑driven mile shifts bargaining power. Smaller AV startups without a distribution deal risk becoming science projects. Cities inherit new complexity: they’ll have to regulate not one robotaxi operator, but a constantly shifting mix of fleets operating under Uber’s umbrella.
Uber’s “roulette table” strategy makes financial sense. Strategically, it raises a hard question: if everyone is your partner, who are you really building around — and how do you guarantee safety and consistency across such a fragmented ecosystem?
The bigger picture
This deal sits at the intersection of three long‑running trends in autonomy.
1. From full‑stack dreams to supplier reality.
A decade ago, most serious AV players wanted to be the next Uber, not its vendor. Waymo, Cruise, Zoox, and Tesla all talked about controlling the full experience. That vision ran into the wall of hardware costs, safety setbacks and a stubborn public unwilling to be beta testers at scale. The post‑2020 “AV winter” forced a reset: many companies quietly pivoted to licensing software, selling sensor stacks, or becoming fleet operators for established platforms.
Uber’s Waabi deal formalizes that reality. Instead of competing with Uber, AV companies increasingly treat it as a wholesaler of rides — a marketplace that buys autonomous kilometers and resells them to end users.
2. Simulation eats road‑testing.
Waabi’s pitch, as highlighted by TechCrunch, is “simulation‑first” development: train and validate most of the driving brain in highly realistic virtual worlds, then use real‑world miles mainly for edge cases and validation. This mirrors what has already happened in aviation, chip design and even pharmaceuticals, where high‑fidelity simulation slashes cost and accelerates iteration.
If Waabi’s approach works at scale, it could shorten the painfully long timelines that have plagued AV deployment — and make it easier for Uber to copy‑paste successful stacks into new cities. That would be a genuine shift in the industry’s cost structure.
3. Platform dominance in mobility.
Lyft in the U.S., Didi in China, Grab and Gojek in Southeast Asia, and Bolt in Europe are all watching the same movie: the ride‑hailing app is becoming the gateway to multiple modes of transport, from bikes to scooters to robotaxis. In this world, the brand the rider sees may matter more than the badge on the car.
For Uber, the Waabi move is another brick in that wall. It signals that the company doesn’t need to “win” autonomy technically. It only needs to ensure that when self‑driving finally works reliably, it happens on Uber’s terms and on Uber’s rails.
The European and regional angle
From a European perspective, Uber’s AV portfolio is colliding with a much denser regulatory wall than in the U.S.
Under the EU’s AI Act, high‑risk AI systems in safety‑critical domains — and that clearly includes self‑driving cars — face strict requirements on data quality, transparency, human oversight and post‑market monitoring. Add GDPR on top, and every kilometer driven by a robotaxi — with cameras, lidar, microphones and telemetry — becomes a data‑protection puzzle. Who is the controller: Uber, Waabi, the fleet owner, or all three?
For European cities that already wrestle with Uber over labor classification and local licensing, a future where multiple autonomous fleets operate under a single platform is uncomfortable. Berlin, Paris or Barcelona may demand city‑specific guarantees on safety metrics, data retention, explainability of AI decisions and even local data storage.
There is also a competitiveness angle. Europe has interesting AV tech (Wayve in the UK, major projects at VW, Mercedes, Stellantis, and suppliers like Mobileye), but far fewer scaled consumer ride‑hailing brands. Uber’s multi‑partner autonomy play could become a de‑facto standard that European players are forced to integrate with — especially in markets where Uber and Bolt already dominate the app on people’s phones.
For European riders, the promise is appealing: cheaper, more predictable trips, especially at night or in underserved areas. For unions and policymakers, it’s a coming storm: how to manage the transition without detonating local taxi ecosystems and precarious gig work.
Looking ahead
The headline number — 25,000 robotaxis — invites misinterpretation. This is not a promise that your next Uber in Ljubljana, Berlin or Madrid will be driverless.
Expect a slow, highly localized roll‑out. The first real impact will be in a handful of U.S. and possibly Canadian cities where regulation and weather are friendly, and where Uber can tightly geofence operations. Only after years of safety data, political negotiation and technical hardening will large‑scale deployments in dense European capitals become plausible.
For Uber, the next steps to watch are:
- Metrics: when does it start reporting the share of trips or gross bookings served by AVs, even in a single city?
- Consolidation: does its roster of 20+ AV partners shrink over time as winners emerge, or does it stay intentionally fragmented to preserve bargaining power?
- Labor strategy: how does Uber frame AV expansion to regulators and driver partners — as replacement, or as a way to cover unprofitable times and zones?
For Waabi, the milestone‑based $250 million matters. It implicitly sets a timeline: fail to hit deployment goals, and the money — and perhaps the Uber relationship — weakens.
The biggest unknown is not technical; it is social. One or two high‑profile accidents in Europe could freeze political will overnight, just as we saw in previous AV setbacks. Conversely, a few years of boring, mostly incident‑free robotaxi service could normalize the technology faster than skeptics expect.
The bottom line
Uber’s billion‑dollar deepening with Waabi is less about choosing a winner and more about locking in optionality. As a pure business move, the portfolio approach to autonomy is smart: it keeps Uber asset‑light while positioning it as the default marketplace for self‑driving miles. But the strategy externalizes risk onto cities, regulators and workers, while assuming that safety, trust and governance will somehow be solved by its partners.
The open question for readers: when your next “Uber” arrives without a driver, will you trust the brand on the app — or demand to know much more about the AI actually at the wheel?
