Uber CEO Dara Khosrowshahi on self-driving's future, changing business model, job displacement

A platform betting on other people’s wheels

When Dara Khosrowshahi talks about the future of transportation, he speaks from a vantage point few companies enjoy: a global marketplace that sits where demand meets movement. Uber, under his stewardship, is no longer a simple app that matches drivers to riders. It is an orchestration layer—an enormous, liquid funnel of demand that can pour customers into increasingly varied modalities, from human drivers to Waymo robo-taxis, from sidewalk robots to eVTOL air taxis. That position changes the stakes of technological progress. It also reframes the central question for cities, fleet owners and investors: who will own the vehicles, and who will own the customer?

Partnerships, not exclusivity

Rather than trying to be the sole builder of autonomy, Uber is assembling partners. Khosrowshahi describes more than twenty collaborations spanning mobility and delivery, including established players in the U.S. and fast-moving firms in China such as Baidu and Pony. These alliances are pragmatic: some firms bring mature no-driver operations, others offer promising regional capabilities. The result is a mosaic approach—multiple technical strategies deployed across diverse urban contexts, with Uber as the distribution hub that turns sporadic supply into consistent, monetizable trips.

Why partnerships matter

There is a clear economic logic. Fleets that feed into Uber’s network benefit from higher utilization because demand already sits on the platform. A car integrated into Uber’s marketplace will likely generate more revenue per day than a directly operated, standalone fleet, simply because requests come from much closer and more frequently. That flows through to improved unit economics for owners and investors, and creates a strong incentive for independent robotaxi makers to collaborate.

Technical divides, and the contours of safety

The conversation revealed a tension at the heart of autonomy: divergent engineering philosophies. One approach leans on camera-only stacks that emphasize software ingenuity and tight compute integration. Another favors redundancy—camera, radar and lidar—paired with high-definition maps that make permanence in the environment explicit and simplify perception. Khosrowshahi framed these as different trade-offs between up-front hardware cost and the complexity of the software problem. For Uber, safety is the non-negotiable filter: partners must meet a safety bar that is multiple times better than human driving, and that assessment blends technical architecture with empirical performance data.

How fleets will be financed and structured

Uber’s forecast for fleet ownership is less romantic than revolutionary: expect financialization. Vehicles will likely be owned by specialized financing entities—think fleet REITs or institutional owners—rather than the technology companies themselves. In the interim Uber is willing to take balance-sheet risk to prove economics and accelerate supply, but the end-state Khosrowshahi envisions is a market of capital owners monetizing assets through platforms like Uber. This mirrors other industries where brands and operators are distinct from asset owners, creating a market for investors to buy exposure to mobility revenue streams.

Delivery’s verticals: robots, drones, humans

On the delivery front, the company is pursuing a plural strategy. Sidewalk robots—slow-moving, low-cost machines that handle tight, short-distance drops—are already operating in test markets and suit dense urban pockets. Drones fill a complementary role in lower-density, suburban geographies where vertical separation is safer and more efficient. Together, these approaches could address a substantial share of last-mile delivery demand, but there remains a stubborn challenge: the first and last few meters of a delivery into an apartment or building, where human hands or new micro-robotics will still be necessary.

• Sidewalk robots capture short, under-one-mile urban deliveries.
• Drones address spread-out suburban drops and speed-sensitive use cases.
• Human couriers remain crucial for complex, multi-step handoffs.

The social cost and the slow pivot

Perhaps the most sobering thread in the discussion is the labor question. Autonomous vehicles, like many forms of automation, pose a long-term displacement risk for drivers and couriers who form the backbone of gig platforms. Khosrowshahi emphasized a practical timeline: in the next five to seven years, growth in demand means more human drivers will be needed, not fewer. Beyond that horizon, the displacement problem becomes real and intractable at scale. The company’s current mitigation strategy is incremental—retraining, redirecting labor to other platform roles, and creating new on-demand opportunities—but the conversation acknowledged there is no tidy policy fix awaiting implementation.

Airspace, the third dimension of mobility

Uber’s interest in eVTOL and urban air mobility signals a broader belief: cities will ultimately use vertical space to relieve lateral constraints. Investing in Joby and exploring partnerships for aerial taxis are consistent with a future where urban transportation is a multi-layered system; ground-based congestion creates an economic case for higher-cost, faster vertical options for certain trips. The implication is that urban planning, charging infrastructure and regulatory frameworks will need to evolve together to make that layered system coherent.

Balancing buybacks and bets

Running a platform with robust cash generation changes strategic possibilities. With substantial free cash flow, Khosrowshahi argues Uber can repurchase stock while also funding aggressive investments in autonomy and new mobility modes. The dual approach—returning capital to shareholders and seeding future-facing infrastructure—reflects confidence in both near-term profitability and long-term market expansion driven by autonomous technology and new vehicle forms.

Final thought: platforms as civic actors

What emerges from the conversation is a portrait of a company that sees itself as both a market operator and a civic actor. By offering planning data to cities, helping to site charging and drop-off zones, and shaping how fleets move through neighborhoods, platforms will be active participants in urban design. That role brings responsibilities that go beyond optimization and into questions of equity, labor policy, and the character of public space. If the promise of autonomy is fewer crashes and broader access to mobility, the real test will be whether the benefits are distributed or concentrated, and whether institutional incentives align with the public interest.

Insights: A network that delivers demand can steer the economics of autonomous fleets, but the transition will require new financing models, municipal cooperation, and deliberate attention to displaced workers.