Uber: How Deep Should the Platform Go in Autonomous Mobility?
Download case PDF (5 pages)Request teacher’s pack
Decision setting: September 4, 2026
Information cutoff: September 4, 2026, 23:59 UTC
Case convention: This is an author-created strategy exercise based on public information. It does not represent private deliberations or an actual decision by Uber management.
The question
Uber had spent years building a marketplace that connected riders with independent drivers. Autonomous vehicles (AVs) introduced a different kind of supply: vehicles and software supplied by partners, often with their own capital needs, operating requirements and customer relationships. Uber could provide demand, dispatch, payments and local operations without owning the technology. But if a dependable ride depended on coordinating the vehicle, autonomy system, fleet, insurance, charging, customer support and local approvals, how much of that system should Uber integrate itself?
As of September 2026, the company had announced a broad set of partner and operating capabilities, several deeper partnerships and a small supervised service in London. Management also described substantial partner commitments and a large expected capital envelope. Its public statements included signs of existing partner activity and rider interest, but did not disclose AV-specific profitability, utilization or customer retention.
Uber’s strategic choice was how to commercialize AV rides as deployments expanded. Should it remain a broad marketplace and commercialization layer; make deeper, sometimes exclusive commitments to selected partners; or combine broad access with selective investment through explicit readiness gates? Which activities should Uber control, and what evidence should determine the pace?
A marketplace with room to grow
Uber entered the question with a large existing platform. In the second quarter of 2026, it reported 208 million monthly active platform consumers, 3.867 billion trips and $58.022 billion in gross bookings. Mobility gross bookings grew 20% year over year on a constant-currency basis. These figures describe the full platform and its human-driver and delivery businesses; Uber did not report AV trips or AV revenue separately. They cannot establish robotaxi demand, utilization, or economics. [1]
Uber’s February 2026 launch of Uber Autonomous Solutions described three groups of services: infrastructure, user experience and fleet operations. The company listed mapping and data, regulatory support, financing, in-car experience, customer support, fleet management, field assistance and insurance among the capabilities. This positioned Uber to help partners enter markets and serve riders without requiring Uber to develop every vehicle or autonomy system. The announcement described offered capabilities, not independently measured cost savings or margins. [2]
At the second-quarter earnings discussion, management said AVs were live on Uber in seven cities and targeted presence in “as many as 15” by year-end. It also said partners had committed approximately 120,000 vehicles over coming years. These are three different measures: a management-reported current footprint, an expansion target and partner commitments. The target was not a completed rollout; committed vehicles were not delivered or productive fleet capacity. [3]
Management also said partners were completing millions of autonomous trips each year across Uber’s Mobility, Delivery and Freight businesses. It said Uber’s category share in Los Angeles, San Francisco and Phoenix—both citywide and within AV operating zones—was higher than a year earlier. The trip figure spans several services and does not isolate robotaxis, any one city, or vehicle utilization. The remarks did not specify the share metric or methodology, and the comparison does not establish that AVs caused a change in share. [3]
The same prepared remarks said Uber expected to commit more than $10 billion over coming years across equity investments, infrastructure and vehicle offtake. Management was evaluating external capital to preserve balance-sheet flexibility. This was a forward-looking, multi-year expectation, not cash already deployed or an AV investment budget allocated by city or partner. The remarks did not specify the timing or composition of the expected amount. [3]
Uber’s reported financial capacity was significant but did not resolve the choice. In the first half of 2026, operating cash flow was $5.213 billion and company-defined free cash flow was $5.078 billion. The company also reported $4.870 billion in cash and cash equivalents at June 30. These consolidated measures include all of Uber’s businesses and obligations. Second-quarter net income included $1.6 billion of pre-tax unrealized investment gains, underscoring why accounting income should not be mistaken for recurring AV cash generation. [1]
Two partnerships, different depths
Uber’s March agreement with Rivian illustrated a deeper and more exclusive approach. Uber described up to $1.25 billion of milestone-contingent investment through 2031, an expected initial deployment of 10,000 R2 robotaxis, and an option for up to 40,000 additional vehicles. The planned launches were to begin in 2028, and the vehicles were to be available exclusively on Uber. The investment was contingent; the vehicles and rollout were future plans, not a delivered fleet or realized return. [4]
The September 3 London launch with Wayve looked different. Uber and Wayve began offering autonomous rides with a small fleet. A TfL-licensed private-hire driver remained onboard, and riders could choose to switch before the vehicle arrived. The launch was supervised and deliberately gradual. Uber said more than 140,000 London riders had opted in to improve their chance of being matched with an autonomous vehicle. That figure reflected stated preference, not completed rides, repeat use or retention. The announcement reported no trip volume, wait times, cost, safety record or unit economics. Wayve and Uber said they aimed to expand to additional markets, but an aim was not a completed deployment. [5]
Together the examples raised a portfolio question. A broad set of partnerships could give Uber access to developers with different technologies and operating domains, while the app and operations layer could help connect supply to rider demand. But coordination could become harder as vehicles, software, fleet readiness and local rules differed. A partner might use other channels or build a direct customer relationship; the public record did not establish whether any partner could or would multihome. Exclusivity could support tighter coordination and a differentiated service, while concentrating capital and execution exposure in fewer partners.
The work behind one autonomous ride
An AV ride involved more than matching one rider with one vehicle. The vehicle and autonomy system had to operate in a permitted domain; the fleet had to be financed, maintained and positioned; dispatch had to meet demand; and the service needed insurance, incident response, customer support and regulator engagement. A delay or failure in one activity could constrain the rider proposition even if the app or vehicle performed as intended.
Research on platform boundaries helps frame the choice. Hagiu and Wright describe a trade-off: integration can coordinate activities with important spillovers, while platform participation can preserve partners’ incentives and their ability to adapt using information Uber may not possess. Their model is stylized and concerns professional services; it does not estimate the best AV ownership or contracting choice for Uber. Adner’s ecosystem framework directs attention to whether the partners and activities needed for a proposition are aligned. Baron’s historical analysis of Uber highlights links between market design and institutional engagement, but it describes 2016–17 and cannot establish current AV regulation or labor conditions. [6]
Uber itself warned investors that AV development, competitors, partner performance, crashes and regulation could affect the business. It identified risks that a partner could fail to meet expectations or withdraw and that rivals might develop superior or cheaper technology. These disclosures described possible risks, not their likelihood. [1]
Three paths to consider
Uber could organize its next phase around one of three broad approaches:
- Broad commercialization layer. Enable a wide range of AV partners through the app, marketplace and operating services. Keep investment and exclusivity limited, while seeking to make Uber a useful route to riders and local execution.
- Deeper anchor integrations. Concentrate capital and operating resources on selected partners, with tighter integration and, where justified, exclusive supply. Seek greater coordination over readiness, quality and rider experience.
- Stage-gated mix. Keep broad access while deepening only selected relationships as they meet defined operating, safety, financing and service-readiness conditions. Decide which activities Uber should control and which it can leave to partners.
The alternatives can vary by city, vehicle and operating domain. The public evidence does not establish that one boundary will fit every market. Nor does it disclose the partner-level economics required to calculate an optimal investment or rollout pace.
Discussion
- What is Uber’s distinctive contribution to an AV ride, and which activities are critical to delivering it reliably?
- Compare broad access with deeper integration. How do coordination needs, partner incentives, differentiation, capital exposure and counterparty risk change across the alternatives?
- Recommend a commercialization approach. Which decisions should Uber centralize, and which should remain with partners or local operators?
- What evidence would you require before expanding a partnership, adding exclusivity or committing more capital? Identify measures Uber should track without treating platform-wide results as AV economics.
Sources
All facts are limited to information public by the stated cutoff. Management statements, partner commitments, targets and vendor descriptions are attributed; none is presented as independently verified operating performance.
- Uber, Q2 2026 Form 10-Q, filed August 5, 2026: SEC filing. Q2 platform highlights, printed pp. 34–35; Mobility and Delivery segment discussion, pp. 40–41; cash flow statement, p. 9 (six-month consolidated operating cash flow and capital expenditures); June 30 consolidated balance sheet, p. 4 (cash and cash equivalents); “Free Cash Flow” reconciliation and “Liquidity and Capital Resources,” pp. 43–44 (company-defined FCF); Q2 highlights, p. 35, “Other Income (Expense), Net,” pp. 39–40, and Note 2 (Q2 pre-tax unrealized investment gains); autonomous-vehicle risk factor, pp. 54–56. Cash flow, cash and FCF are consolidated company measures; FCF is defined by Uber and is not an AV measure.
- Uber, “Uber Unveils Uber Autonomous Solutions to Accelerate Autonomous Mobility & Delivery Worldwide,” February 23, 2026: company release.
- Uber Q2 2026 prepared remarks, August 5, 2026, “Autonomous mobility,” PDF pp. 4–5: PDF.
- Uber and Rivian, “Uber and Rivian Partner to Deploy up to 50,000 Fully Autonomous Robotaxis,” March 19, 2026: company release.
- Wayve and Uber, “Wayve and Uber Launch First Ever Autonomous Rides in the UK,” September 3, 2026: company release.
- Academic sources: Hagiu and Wright, “Multi-Sided Platforms,” International Journal of Industrial Organization 43 (2015), DOI; Adner, “Ecosystem as Structure,” Journal of Management 43(1) (2017), DOI; Baron, “Disruptive Entrepreneurship and Dual Purpose Strategies: The Case of Uber,” Strategy Science 3(2) (2018), DOI.