Inside Uber's Autonomous Delivery And Robotaxi Machine

Uber posted its first year of GAAP profitability in 2023. Net income of $1.9 billion. It followed with $9.9 billion in 2024 and $10.1 billion in 2025. Revenue went from $17.5 billion in 2021 to $52 billion in 2025. The business scaled through the pandemic, through the driver pay adjustments, through the regulatory pressure in the UK and California, through inflation in fuel costs. It kept compounding.

The mix shift matters. Mobility is 55 to 60 percent of revenue, Delivery is 35 to 40 percent, Freight and advertising make up the rest. Advertising is the fastest growing line at high incremental margin. Delivery profitability has improved meaningfully since 2022 as the delivery fee structure rationalised and grocery began contributing at scale.

Autonomy sits inside this as a cost-side intervention. Every ride and every delivery currently pays a driver or courier. That is the largest cost line. Any reduction in human labour per trip flows directly to take rate.

Coco Robotics is a last-mile sidewalk delivery company using small autonomous ground vehicles. Uber Eats has been running pilots in Los Angeles and Miami. The San Jose expansion on 22 April 2026 is the first time the deployment scaled into a top-twenty metro with a density profile that makes the unit economics interesting. Coco vehicles at current operating cost run roughly $2 to $3 per delivery at scale. Human couriers on Uber Eats average somewhere in the $5 to $7 per delivery range when factoring tip pass-through and platform fees. The gap is the business case.

Lucid on the robotaxi side is more complex. Uber has explicitly positioned itself as the marketplace layer above whichever autonomous hardware vendors achieve Level 4 scale in which markets. The Lucid stake is a signal that Uber wants commercial optionality across the Tesla, Waymo, Lucid, Aurora, and smaller-player universe. Owning a stake in one robotaxi hardware vendor is not a bet on that vendor winning. It is a bet on having multiple routes to onboard autonomous supply when each geography opens.

The unit economics on robotaxi are more aggressive than delivery. Human-driven Uber rides pay out somewhere in the 60 to 70 percent take rate range to drivers. A robotaxi configuration keeps 90 to 95 percent of the fare inside the platform, with the remainder paying down the vehicle cost amortisation and the vehicle operator partner. The per-ride margin profile of an autonomous ride is structurally two to three times the margin of a driver-delivered ride. At scale, a 10 percent mix of robotaxi in a single market can lift mobility segment EBITDA by 15 to 20 percent.

The execution risk is real. Regulatory approvals vary by city. Safety incidents at any single autonomous operator can set back the whole industry. Tesla's FSD strategy is a direct competitive threat. But the Uber hedge is structural. Uber does not have to pick the winning autonomous hardware platform. It has to be the preferred route to the consumer for whoever wins.

At 23x forward earnings and a $159 billion market cap, Uber is priced closer to a software compounder than a transportation marketplace. The multiple is defensible if you believe either mobility TAM grows into the $300 billion range by 2030 or autonomy lifts platform-level EBITDA margins from the current 14 to 15 percent range toward 20 percent plus within five years.

Free cash flow at $9.8 billion in 2025, with capex at $336 million, is the profile of a business that does not need outside capital to fund the autonomy pivot. That is the structural advantage over any pure robotaxi hardware play. Tesla's robotaxi ambitions require hundreds of billions of capex. Waymo's deployment requires continued Alphabet subsidy. Uber's autonomy strategy costs roughly $500 million to $1 billion a year in partnership equity plus software investment. That is 5 to 10 percent of current FCF.

Operating margin at 12.4 percent is below the long-run target. Advertising mix shift alone should push that toward 14-15 percent inside two years. Autonomy layered on top pushes it toward 17-18 percent by 2028.

Five autonomous archetypes are competing for mindshare. Tesla's vertically integrated approach with FSD. Waymo's closed ecosystem rolling out city by city. Lucid, Zeekr and new-entrant vehicle OEMs supplying hardware to marketplace operators. Aurora and similar pure-play software-plus-vehicle technology stacks. Uber-style marketplaces aggregating demand.

Only the marketplace approach scales immediately with existing consumer behaviour. Tesla has the technology but not the rider distribution. Waymo has the technology and limited rider distribution. Uber has the rider distribution and a growing portfolio of technology partnerships. In a scenario where multiple autonomous technologies reach Level 4 inside five years, Uber is the highest-probability aggregator across them.

The sector pattern is revealing. Every marketplace that layered a technology shift on top of existing liquidity has outperformed over a five-year window. Booking did it with mobile. Amazon did it with cloud. The list is short but consistent. Uber is attempting the same move with autonomy.

Advertising revenue is the fastest and cleanest growth line. It is already running at a multi-billion dollar annual run rate and scales with take rate rather than gross bookings. Incremental EBITDA flow-through is 70 percent plus.

Grocery delivery remains underpenetrated relative to Instacart and continues to scale. Every grocery order carries higher basket size and stickier frequency than a restaurant order.

Freight has been a drag segment but is positioned to inflect in 2026-2027 as freight pricing normalises. It is the optionality inside the portfolio that most analysts ignore.

Autonomy, as discussed, is a margin story rather than a revenue story. The Coco and Lucid announcements are the first concrete revenue touchpoints that show autonomy is contributing rather than costing.