Uber bets its balance sheet on the AV “multitrillion‑dollar” future

An uncommon combination: scale, growth and cash

Uber’s latest numbers read like the financials of a much older company, but the narrative Dara Khosrowshahi chose for Q4 2025 was firmly about the future. The group closed the year with what he called “another great quarter”: platform trips have accelerated to a 15 billion annual run rate, and monthly active users are now more than 200 million, underpinned by a fifth consecutive year of more than 20% annual gross bookings growth.

On the P&L and cash flow statement, the scale is starting to look industrial. Adjusted EBITDA reached $8.7bn in 2025, up 35% year-on-year, while free cash flow was a “phenomenal” $9.8bn, up 42%. That level of cash generation gives Uber room to present itself as both a growth and a capital‑return story.

Incoming CFO Balaji Krishnamurthy was at pains to emphasise that this is not, in his framing, a trade‑off. Uber’s first priority remains reinvestment in its core businesses, but he said the company can simultaneously fund an increasingly ambitious autonomy strategy, evaluate bolt‑on M&A, and continue an “aggressive” share repurchase programme. The implicit benchmark remains the 50% of free cash flow return target discussed over the past year, with management now describing the stock as “really cheap” and buybacks as “pretty awesome” at current levels.

Underneath the headline momentum, management dwelt on the composition of that growth. Audience expansion, rather than mere frequency gains, is still the primary engine. Monthly active platform consumers grew from around 14% year-on-year at the start of 2025 to 18% by year‑end, reaching over 202 million. Annual actives are more than 450 million; Uber is still increasing penetration of that base, with 40% of consumers now using more than one product in Q4.

Krishnamurthy highlighted stronger retention in newer rider and eater cohorts, attributing it to heavier early life‑cycle investments and the flywheel of multiple products and membership. Uber One membership, now about 46 million members and growing more than 50% year-on-year, has become the glue in that ecosystem, accounting for a rapidly growing share of gross bookings and “supercharging” lifetime value.

A quieter revolution in the profit pool

While most investors’ attention is fixed on the race for autonomous vehicles in dense urban cores, Krishnamurthy used his first appearance as incoming CFO to challenge a long‑standing assumption about where Uber actually makes its money. Roughly 30% of bookings come from major U.S. cities, but, he said, “70% of the U.S. is outside of the top markets and nearly 75% of our U.S. profits come from those markets.”

Those non‑top‑20 U.S. cities – and still sparser areas beyond them – are growing faster than the urban flagships and increasingly drive the profit pool. Globally, about 60% of Mobility gross bookings are now outside the U.S., reinforcing the idea of Uber as a geographically diversified transportation utility rather than a purely American urban app.

Pricing dynamics have quietly pivoted in Uber’s favour too. In 2023 and early 2024, U.S. growth was hampered by insurance inflation that was pushed through to riders. Through 2025, Uber kept prices “relatively consistent” while pursuing reforms and product‑driven cost savings, turning insurance from a deleveraging item into a source of operating leverage. The company now expects that price stability, coupled with its “barbell” product strategy, to underpin an unusual prospect for a business at its scale: acceleration in U.S. trips and gross bookings into 2026.

At the low end, products such as Wait & Save, Moto and other discounted options are bringing in more price‑sensitive demographics, particularly in suburbs and less dense geographies. At the higher end, Reserve, SUVs and XXL, airport shuttles and tailored offerings like women‑preferred rides are capturing more affluent or time‑sensitive use cases, often at higher margins and with better earnings for drivers.

The overarching pattern is of a platform broadening both its demographic reach and its geographic footprint, using a proliferating set of use cases – from teen rides to simplified products for older riders – to deepen engagement across cohorts that would never have been Uber’s early adopters.

Delivery and advertising: from bolt‑on to backbone

On the Delivery side, Khosrowshahi laid out five pillars behind the business’ acceleration to multi‑year highs: selection, less dense areas, new verticals, membership and international expansion.

Selection remains a basic but powerful driver. In many countries, Uber’s restaurant and merchant coverage is still only 30–40% of the addressable market. In the U.S., particularly in suburbs and among small and medium‑sized businesses, management concedes its selection is “not where we want to be.” An AI‑enabled salesforce is now pushing to accelerate merchant additions, with the dual aim of drawing in new users and boosting conversion among existing ones.

Beyond restaurants, Uber is steadily expanding into grocery and retail – a “$1 trillion opportunity” in Khosrowshahi’s words. The company now works with five of the top 10 grocers in the U.S., with further U.S. partnerships to be announced, and has struck a multiyear exclusive with Coles, a leading grocer in Australia. Combined with membership discounts and cross‑selling from rides, this is turning Eats into a broader local commerce platform.

Perhaps the most under‑appreciated profit lever is advertising. Early on, Uber guided the market to think of 2% of Delivery gross bookings as a ceiling for ad penetration. That ceiling has already been breached. Small and medium‑sized business (SMB) advertisers now run at well above 2% penetration, while year‑on‑year growth in enterprise ad revenue is “outpacing SMBs by a lot more” as larger brands belatedly ramp spend. On top of that, ads on grocery, retail and Mobility are still “nascent,” offering incremental runway.

Taken together, selection expansion, the push into less dense areas, the layering of new grocery and retail products, the Uber One flywheel and measured international roll‑outs – such as the now dominant organic positions in Japan and a fast‑rising share in Germany and the U.K. – suggest Delivery is graduating from a pandemic boom‑and‑bust adjunct to a structural counterpart to Mobility.

The AV wager: platform, not plant

For all the operational detail, the centrepiece of Uber’s investor‑facing story this quarter was autonomous vehicles. Khosrowshahi returned repeatedly to the idea that AVs will “unlock a multitrillion‑dollar opportunity” and, crucially, that they are more likely to augment than replace Uber’s existing model.

The empirical evidence he offered was partly behavioural. In markets where AVs operate – whether as a competitive service in San Francisco or on Uber’s own platform in Austin and Atlanta – Uber has seen acceleration in gross bookings, faster growth in new riders than the rest of the country and strong trip frequency. AVs, he argued, are expanding the “economic pie” rather than cannibalising it.

More technically, Uber believes its marketplace yields a superior AV product. According to Khosrowshahi, AVs on Uber’s platform currently record trips per vehicle per day roughly 30% higher than on standalone, first‑party AV platforms, with better estimated arrival times. Higher utilisation is the core of Uber’s pitch to AV manufacturers and software providers: in a capital‑intensive category, sweating the asset is the point.

The company is positioning itself firmly as the third‑party orchestrator, not as a vertically integrated AV OEM. Management expects to have AVs operating in 15 cities by the end of this year, via a portfolio of partnerships spanning Waymo, NVIDIA, Waabi, Avride, Nuro, Lucid and others. Many of the deals include explicit economic commitments: Uber has, for instance, invested in Waabi and secured exclusivity on the first 25,000 passenger vehicles produced on its platform.

The deal terms, Khosrowshahi stressed, already deliver “positive economics” at current consumer fares, even as AV products initially come to market at lower margins than mature options such as UberX. The familiar pattern, he suggested, will be to start at thin margins to build liquidity and then improve profitability as volumes scale and routes densify.

On the capital side, Uber’s approach is closer to Marriott than to Tesla. The company will put capital behind AV deployments – from equity investments in software players to some early vehicle purchases and depot infrastructure – but is in active discussions with private equity firms, banks and other financiers to “financialise” the fleet. Just as real‑estate investment trusts own hotel walls and lease them back to operators, Uber expects specialist asset owners to emerge for AV fleets. In practice, it will lean on its existing, vast network of fleet partners, many of whom are already financed by banks to buy EVs and can transition to AVs over time.

The partnership with NVIDIA is central to the technological scaffolding. The two are building what Khosrowshahi called a “real‑world data collection factory,” targeting more than three million hours of highly specific AV training data – pickups, drop‑offs, kerbside complexity – that will be shared with partners. Combined with ever more powerful simulation capabilities, particularly in “AV 2.0” software players such as Waabi, this is meant to erode the data advantage of incumbent AV leaders and prevent the market becoming winner‑take‑all.

Krishnamurthy provided a staged view of AV deployment on Uber’s network. In the first phase, AVs provide a baseload of supply, soaking up trough‑period demand where human drivers are less available. Because demand for AVs does not crater as sharply from Saturday to Monday as it does for the broader network – a roughly 45% drop – AV fleets can run at more even utilisation. Over time, as hardware costs decline and manufacturing scales, Uber expects to expand AVs from trough demand into medium‑load periods, potentially lowering consumer prices and expanding the total addressable market.

The distant endgame, in management’s telling, is a world where AVs constitute a majority of supply in some markets. But they were careful to emphasise how far off that is, given where OEMs are on production ramp curves. For now, the bottleneck is vehicle manufacturing. Commitments for “tens of thousands” of vehicles – and eventually more – are as much about giving carmakers the confidence to build as about signalling scale to investors.

What may ultimately matter more than deployment pace is network utilisation. Khosrowshahi argued that Uber’s combination of Mobility, Delivery and Freight gives it a structural advantage in filling idle capacity. Human drivers already switch between people and parcels; AVs could, in principle, do the same, working peak hours on Mobility and off‑peak on last‑mile delivery or freight, something no pure‑play AV operator can match today.

Management’s baton change

Amid the strategic bravado, there was a more human note as the company choreographed a leadership transition. CFO Prashanth Mahendra‑Rajah, who helped steer Uber to investment‑grade status, launched its first share repurchase programme and shepherded multiple acquisitions, will step down on 16 February to take up an unspecified public service role “to serve America.”

Khosrowshahi’s thanks for a “great partner” were coupled with praise for Balaji Krishnamurthy, described as a “bold thinker and brilliant strategist” who “knows our business inside and out.” Krishnamurthy, for his part, talked about “building from a strong base” of accelerating core businesses, “large and growing cash flows” and a “GO‑GET it culture” that he now has to translate into capital allocation discipline.

For investors, the subtext is that Uber wants to be thought of less as a tech experiment and more as a once‑in‑a‑generation infrastructure company – one that throws off double‑digit billions of free cash flow, returns a large share to shareholders and, at the same time, is placing a carefully structured bet on a new transport layer that could, in time, reshape the very markets that now underpin its numbers.