Uber's $10B gamble: When aggregators stop being just 'connectors'

Last week in Abu Dhabi, I opened Uber to hail a ride and noticed a new option called "Autonomous." When I tapped on it, the license plate showed up as a night-blue WeRide GXR. In that moment, it hit me: Uber is no longer just an asset-light ride-hailing aggregator.

Over the past two years, Uber has signed partnership agreements with over 30 autonomous driving companies, committing to invest more than $10 billion cumulatively. From Nuro to Lucid, from Rivian to Baidu, this list keeps growing. Uber isn't building cars; it's building an "App Store for Autonomous Driving"—but the problem is, it's also slapping its own label on some of the products in that store.

From "Connection" to "Control": Uber's Logic Has Changed

Previously, Uber's model was simple: you drive your car, I take my orders, take a cut, and leave. This asset-light mode allowed for rapid expansion but planted seeds of danger—when robotaxis truly land, automakers could bypass Uber to operate fleets directly. Where does Uber's value lie then?

Now, Uber's strategy is to "bet on both sides." On one hand, it continues to act as an aggregation platform, integrating fleets from Waymo, WeRide, Nuro, etc., into its app. On the other hand, it has started investing directly in vehicles and operations: partnering with Nuro to deploy 35,000 robotaxis, committing nearly $500 million to Nuro, with funds released based on milestones.

This move is smart. The investment isn't a lump-sum payment but is disbursed in installments based on technical delivery, operational data, and user satisfaction—a pure product manager mindset. You park the money there, but the partner must prove they can actually run before getting paid.

The "Multiple Choice" of Tech Routes: One Platform, Three Solutions

Hidden within Uber's partnership list are three completely different technological routes:

• Nuro Solution: Focuses on urban delivery and low-speed scenarios. The tech route is relatively conservative but lands quickly. Partnering with Lucid to use the Gravity SUV means Uber is testing the feasibility of "high-end autonomous driving."
• WeRide Solution: Already operational in Abu Dhabi and is Uber's first global dedicated "Autonomous" option. This company's characteristic is "cross-regional deployment"—with layouts in the Middle East, Europe, and Southeast Asia.
• Waymo Solution: Most mature technology but highest cost. While using Waymo's fleet in Austin and Atlanta, Uber publicly criticizes its tech route and deployment strategy.

This contradiction is interesting. As a product manager, I understand Uber's anxiety. Waymo is the industry benchmark, but its tech route is too "heavy"—LiDAR, HD maps, customized vehicles—these costs eventually reflect on the user's bill. Meanwhile, Nuro and WeRide follow a "lightweight" route, using cheaper sensors and smarter algorithms to keep costs within Uber's acceptable range.

The "Old Problem" of Vertical Integration

When I worked on Xiaomi's IoT platform, we discussed how "vertical integration might bring brand risks." Now Uber is walking a similar path—it wants to be both a platform and an operator.

The risks lie in three points:

1. Trust Crisis with Partners: You integrate my fleet while investing in my competitors. Why should I give you core data? Uber's public criticism of Waymo is hard not to see as paving the way for "taking control itself."

2. Uncertainty of Tech Routes: The ultimate winner in autonomous driving hasn't emerged yet. Uber is betting on Nuro, WeRide, and Waymo simultaneously, but if one route proves to be a dead end, billions could go down the drain.

3. Fragmented User Experience: Suppose you hail a ride in Hangzhou; it might be a Lucid equipped with Nuro tech or a WeRide GXR. If the riding experience, pickup logic, and emergency handling differ completely between the two types of cars, users will be confused.

The Real Challenge Isn't Tech, It's "Trust"

Back to the product manager perspective. For Uber doing this, the core KPI isn't "how many cities are onboarded," but "do users dare to sleep in driverless cars."

The barrier to autonomous driving isn't in the algorithm, but in the user's psychological account. Users are willing to pay 20% more for "a visible driver" but refuse to pay even 1% more for "an invisible algorithm." What Uber is essentially doing now is helping users undergo a "trust transition"—using the reputation of the aggregation platform to endorse the reliability of autonomous fleets.

But here's the issue: If a partner's car gets into an accident, do users blame the tech company or Uber? Based on C-end user habits, they'll likely blame Uber. So Uber must establish a "screening + elimination" mechanism to ensure all vehicles integrated into the platform meet unified standards, at least in terms of "safety fallback."

Advice for Action-Takers

If you're following the implementation of autonomous driving, don't stare at Uber's stock price. Look at its partners' roadmaps. Uber's $10 billion investment is essentially helping the entire industry "trial and error":

• Which tech routes work? See who Uber doubles down on.
• Which regions land first? See where Uber's "Autonomous" category goes live.
• Which business models are profitable? See the "milestone payment" criteria Uber gives to partners.

Investors can watch Uber, but practitioners should watch how its partners "die." Because Uber's $10 billion will ultimately become a "survivor list"—those eliminated are precisely those Uber doesn't want to waste money on anymore.

---

📌 This article is compiled from TechCrunch. Original text: https://techcrunch.com/2026/08/01/ubers-autonomous-vehicle-deal-tracker/

Copyright belongs to the original author. This is a compilation and independent analysis based on public reports.