Uber's $52B Platform Bets Everything on Becoming the Android of Autonomous Vehicles

Uber closed FY2025 with $52B in revenue, 200M+ monthly users, and $9.8B in free cash flow — then announced plans to facilitate AV trips in 15 cities by year-end 2026 and 100,000 robotaxis by 2027. The company just swapped its CFO for a robotaxi bull, launched a new AV Labs data division, and partnered with NVIDIA, Waymo, and 20+ AV startups to build the world's largest hybrid human-plus-autonomous mobility network.

UBER · Industrials · February 22, 2026

S&P 500 Position

Classified under Industrials/Passenger Ground Transportation in the S&P 500, which is a somewhat awkward fit for a company that is increasingly a technology platform and AI data business. Uber's closest S&P 500 peers by business model are Booking Holdings and Airbnb (platform marketplaces), though it competes with DoorDash in delivery and faces Tesla and Alphabet/Waymo as emerging AV rivals. Within its formal GICS classification, Uber dwarfs traditional ground transport companies and shares the sector with logistics operators like FedEx and UPS.

Index Weight: ~0.30% | Rank: Approximately #55-65 in S&P 500 by market cap

Company Overview

Uber is executing a dual transformation: squeezing operating leverage out of its now-mature ride-hailing and delivery marketplace while simultaneously building the infrastructure to become the default aggregation layer for autonomous vehicles worldwide. The company surpassed 200 million monthly active platform consumers in Q4 2025, facilitated 40 million trips per day, and generated $193 billion in gross bookings for the full year. Its hybrid network thesis — blending human drivers with autonomous vehicles to maximize utilization and smooth demand variability — is being validated in Austin and Atlanta, where Waymo-powered AV trips on Uber's platform have accelerated overall trip growth rather than cannibalizing it. The strategic posture is clear: Uber does not build self-driving cars. Instead, it is positioning itself as the demand aggregation and fleet management layer that every AV company needs to achieve commercial viability. With partnerships spanning Waymo, Waabi, Avride, May Mobility, Momenta, WeRide, Nuro, and others — more than 20 AV companies total — Uber is assembling a portfolio approach to autonomy. The January 2026 CES unveiling of a co-developed robotaxi with Lucid and Nuro, the NVIDIA partnership targeting 100,000 L4 vehicles starting in 2027, and the launch of Uber AV Labs for driving data collection all signal that Uber is spending real capital to shape this category, not merely riding along. The new CFO, Balaji Krishnamurthy — promoted from VP of Strategic Finance in February 2026 — holds a board seat at Waabi and is a vocal AV advocate. His appointment, paired with a $20 billion share repurchase authorization and a stated intent to invest free cash flow into AV infrastructure, positions Uber to simultaneously return capital to shareholders and fund its autonomous future. The question is whether the company's asset-light playbook can survive contact with a capital-intensive AV build-out.

Products & Revenue

Uber monetizes a three-sided marketplace through take rates on gross bookings across Mobility (ride-hailing, taxis, rentals, micromobility), Delivery (Uber Eats, grocery, retail, Uber Direct white-label logistics), and Freight (digital brokerage for shippers and carriers). Mobility is the cash cow with the highest margins, while Delivery has been closing the profitability gap rapidly — particularly as advertising revenue and grocery/retail grow within the segment. Freight has been a drag through a prolonged cycle downturn but reached breakeven Adjusted EBITDA in Q4 2025 for the first time in three years. Cross-platform users who use both Mobility and Delivery generate 3x the gross bookings of single-product users, which is why the Uber One membership program (46 million members as of December 2025) is a critical retention and engagement tool.

Mobility (~57%): Ride-hailing, car rentals, taxis, micromobility, and transit. Includes financial partnerships products and advertising. Generated ~$8.2B in Q4 2025 revenue alone, up 19% YoY. Highest-margin segment with Adjusted EBITDA margin of ~25.9%.

Delivery (~33%): Uber Eats restaurant delivery, Grocery & Retail (alcohol, convenience, retail), Uber Direct white-label Delivery-as-a-Service for merchants, and advertising. Q4 2025 delivery revenue climbed 30% YoY to ~$4.9B. Grocery & Retail running at $12B+ gross bookings annually.

Freight (~10%): Digital freight brokerage connecting shippers with carriers via the Uber Freight platform. Includes Broker Access capacity-as-a-service. Revenue has been flat due to cyclical freight downturn, but hit breakeven Adjusted EBITDA in Q4 2025.

Advertising (embedded across segments) (Undisclosed but growing): Uber Journey Ads, sponsored listings, and merchant promotions across Mobility and Delivery. Operates as a high-margin revenue overlay across the core marketplace, contributing to Delivery EBITDA margin improvement.

Uber One Membership (Embedded in segment bookings): Subscription program with 46M members across 30+ countries. Drives 35%+ of U.S. Mobility gross bookings. Members get 6% cash back on rides, no delivery fees, and up to 10% off orders. Margin-dilutive initially but improves cohort economics over time.

Based on FY2025 10-K filing (period ending December 31, 2025), Q4 2025 earnings release (February 4, 2026), and FY2024 10-K for historical segment percentages. FY2025 segment revenue percentages estimated from Q3/Q4 2025 quarterly data, as the full annual segment breakdown in the FY2025 10-K was filed February 14, 2026.

Leadership

Dara Khosrowshahi

CEO since 2017. Former CEO of Expedia for 12+ years and CFO of IAC. Brown University BS in Electrical and Electronics Engineering. Took Uber from billions in annual losses to $10B+ in net income and $9.8B in free cash flow. Now staking Uber's next chapter on becoming the world's largest facilitator of autonomous vehicle trips by 2029.

Balaji Krishnamurthy, Chief Financial Officer (effective February 16, 2026): Promoted from VP of Strategic Finance and Investor Relations after six years at Uber. Holds a board seat at AV startup Waabi. Previously spent 8 years at Goldman Sachs. His appointment signals Uber's intent to invest heavily in autonomous infrastructure — he publicly stated the company will invest across AV software partners, vehicle makers, and infrastructure.

Sarfraz Maredia, Global Head of Autonomous Mobility & Delivery: The executive running Uber's AV strategy — responsible for all autonomous marketplace operations, commercial strategy, partnerships, and expansion. Has been at Uber since 2014, previously ran U.S. Mobility and Americas Delivery. Sits on the board of Serve Robotics. Led the Waymo integration in Austin/Atlanta, the Lucid-Nuro robotaxi unveiling, and the Avride/WeRide partnerships.

Sundeep Jain, Chief Product Officer & SVP of Engineering: Oversees product development and engineering operations across the platform. Former VP of PM at Google for Search Ads and VP at Zynga. Responsible for Uber's consumer-facing product experience, including the integration of autonomous ride options into the core app.

Praveen Neppalli Naga, Chief Technology Officer, Mobility & Delivery: Steers engineering and science strategy for Uber's core businesses — focusing on technical execution, platform innovation, and autonomous systems integration. Critical to Uber's evolution in mobility technology.

Albert Greenberg, Vice President, Platform Engineering: Leads data centers, compute, networking, storage, data infrastructure, and developer productivity. Former Microsoft Technical Fellow and Corporate VP for Azure Networking. Serves on Uber's AI Law & Ethics Council. Executive sponsor for Uber's community of senior engineers driving architecture evolution.

The AI Angle

Platform play: selling shovels in the AV gold rush

Uber's AI strategy operates on three distinct layers. The first and most mature is its internal ML platform, Michelangelo, which has powered Uber's operations since 2016. Michelangelo manages approximately 400 active ML projects, runs over 20,000 model training jobs monthly, and serves 5,000+ models in production at 10 million real-time predictions per second at peak. The platform has been extended with full LLMOps capabilities — fine-tuning, prompt engineering, evaluation, and serving — through the GenAI Gateway, a unified internal platform for all LLM use cases that integrates OpenAI, Google Vertex AI, and Uber-hosted open-source models like Llama. Uber has identified 60+ distinct internal LLM use cases spanning process automation, customer support, and content generation. Generative AI now touches the entire platform for demand forecasting, pricing optimization, ETA prediction, fraud detection, and customer support automation. The second layer is Uber AI Solutions, an externally-facing business that monetizes Uber's data infrastructure expertise. Announced in a major expansion in June 2025, Uber AI Solutions offers a data foundry providing ready-to-use and custom datasets (audio, video, image, text) for training large AI models, a global digital task platform available in 30 countries connecting enterprises to expert annotators, agentic AI support tools, and shared annotation/validation infrastructure used internally by Uber. The division is led by Megha Yethadka and competes directly with Scale AI and Labelbox in the AI data labeling market, but with the differentiation of Uber's global operational footprint and real-world data collection capabilities. The third and most strategically consequential layer is AV infrastructure. Uber launched AV Labs in January 2026 — a new division that sends sensor-equipped cars into cities to collect driving data for its 20+ AV partners. This is not a return to building self-driving software; it is a data services play. Uber plans to collect over 3 million hours of robotaxi-specific driving data to fuel L4 model training. The NVIDIA partnership announced at GTC provides GPUs, the Cosmos platform, and simulation tools to build a joint AI data factory. Uber's approach includes running partner software in 'shadow mode' on AV Labs vehicles, flagging discrepancies between human and machine decisions — a technique similar to Tesla's fleet learning but with more targeted data collection. The competitive risk is real. Waymo operates independently in San Francisco and is expanding to 20+ cities by end of 2026. Tesla plans to launch its own ride-hailing network with Cybercab production starting Q2 2026. Uber's bet is that the demand aggregation problem — matching variable rider demand with expensive AV assets across diverse geographies and conditions — is harder to solve than Level 4 autonomy itself. The early data from Austin and Atlanta supports this: Uber claims its hybrid network delivers trips-per-vehicle-per-day roughly 30% higher than standalone AV markets, and more than double the utilization of Los Angeles. If that holds, Uber becomes the indispensable middleware between AV manufacturers and riders.

Financial Snapshot

Revenue (TTM): $52.0B — FY2025 (period ending December 31, 2025) | Net Income: $10.1B net income (includes $5.0B one-time tax valuation release benefit)

Margins: Adjusted EBITDA margin ~4.5% of Gross Bookings (~16.7% of revenue), net margin 19.3% (inflated by tax valuation release)

Uber generated $9.8 billion in free cash flow in FY2025, up 42% YoY — making it one of the most cash-generative tech platforms in the world. The company exhausted its initial $7 billion buyback program and authorized an additional $20 billion repurchase plan in August 2025, repurchasing a record $1.9 billion in Q4 alone, driving a 2% YoY reduction in diluted share count. Balance sheet is clean: $7.6 billion in unrestricted cash, investment-grade credit rating, and 0.50 debt/equity. Capital allocation is shifting from pure shareholder returns toward a blend of buybacks and AV infrastructure investment.

1-Year Performance

UBER trades at $73.86, down ~9.1% YoY and approximately 27.6% below its 52-week high of $101.99 reached September 22, 2025.

Uber rallied through most of 2025 on accelerating earnings and the AV narrative, hitting all-time highs near $102 in September. The stock then sold off on multiple fronts: soft Q1 2026 profit guidance following the Q4 earnings release (shares dropped ~4-6% on February 4), broader concerns about Waymo and Tesla's independent AV ambitions, and a market-wide rotation away from growth names. The Q4 earnings miss on EPS ($0.71 adjusted vs. $0.79 expected) despite a revenue beat underscored that AV investment costs are beginning to pressure near-term profitability. Analyst consensus remains strongly bullish — 34 analysts rate it Strong Buy with a $106 average price target, implying 44% upside.

Recent News

Fun Fact: Uber's internal ML platform, Michelangelo — named after the Renaissance polymath, not any person at the company — has been running since 2016 and now serves over 5,000 models in production at 10 million real-time predictions per second at peak load. The platform was recently extended with a GenAI Gateway that routes all LLM calls across 60+ internal use cases through a unified interface with built-in PII redaction — meaning every time an Uber engineer calls GPT-4 or Llama internally, it first passes through an automated system that strips personally identifiable information before the prompt ever reaches an external model. The PII redactor creates an interesting technical challenge: anonymized queries like 'Who is George Washington?' and 'Who is Abraham Lincoln?' become indistinguishable after redaction, which breaks LLM caching and RAG retrieval pipelines in subtle ways the team has had to engineer around.