Bookings up 24%, and a $10bn down payment on autonomy

Summary

Gross bookings grew 24% to $58.0bn, above the top of guidance and a fourth consecutive quarter above 20%. Revenue grew only 12% to $14.2bn, but that gap is accounting rather than demand: business model changes cost eight points of growth, and Mobility revenue rose 1% against Mobility bookings up 22%. Profitability did the real work — adjusted EBITDA up 33% to $2.8bn at a 4.9% margin, non-GAAP operating income up 40%, non-GAAP EPS up 35% to $0.81. Trailing twelve-month free cash flow passed $10bn for the first time. Treat the $1.17 GAAP EPS with care: $1.6bn of that net income is equity revaluation, not operations.

Metric	Reported	YoY
Gross bookings	$58.0bn	+24%
Revenue	$14.2bn	+12%
Mobility gross bookings	$29.0bn	+22%
Mobility revenue	$7.4bn	+1%
Delivery gross bookings	$27.5bn	+26%
Delivery revenue	$5.2bn	+28%
Adjusted EBITDA	$2.8bn	+33%
Non-GAAP operating income	$2.1bn	+40%
Non-GAAP EPS	$0.81	+35%
Free cash flow	$2.8bn	+13%
Monthly active consumers	208m	+16%
Trips	3.9bn	+18%

Guidance for Q3 is $58.25–60.25bn of gross bookings, 18–22% growth in constant currency, with non-GAAP EPS of $0.84–0.88 and adjusted EBITDA of $2.86–2.96bn. Street revenue consensus sits at $14.69bn for Q3 and $57.85bn for the full year, rising to $67.01bn in 2027.

One warning on the EPS "beat" you will see quoted elsewhere. Uber's GAAP earnings carry revaluations of its equity stakes, which are large and unforecastable — GAAP EPS across the last four quarters printed $3.11, $0.14, $0.13 and now $1.17. Consensus is struck on an adjusted number, so screens comparing the two are not measuring the same thing: this reads as a 40% beat against a $0.83 estimate when the comparable figure is $0.81, which is marginally below. Q3 estimates spanning $0.79 to $1.66 across 29 analysts make the point. Bookings, adjusted EBITDA and free cash flow are the numbers carrying information.

Autonomous vehicles

This was the quarter Uber put a number on it: a $10bn multi-year commitment, split between equity stakes in software partners and balance-sheet support for the physical layer — fleet operations, real estate and OEM commitments covering roughly 120,000 vehicles. The equity piece does more work than its size suggests. Stakes carry milestones that buy roadmap visibility and, in the CFO's words, a place "front of the line for commercialization", and every dollar Uber has committed has catalysed a further $2.50 from other investors, with third-party sponsors now being lined up so the fleet does not sit entirely on Uber's balance sheet.

The strategic bet is that autonomy resembles AI rather than search: Khosrowshahi's case is that AVs are "essentially physical AI", and that as with foundation models, no single stack wins outright. If that holds, the scarce asset is not the driver but the demand aggregation and market-by-market commercialization layer, which Uber already owns. The roadmap has widened accordingly — seven live markets today, 15 by year end, with Nuro and Lucid in the Bay Area, Zoox in Las Vegas, Wayve in London and Tokyo, Baidu in London and Pony and Verne in Zagreb, followed by NVIDIA in 28 cities globally by 2028 and Rivian in San Francisco and Miami the same year. Waymo stays in Austin and Atlanta, though management was pointed about not depending on a single partner.

Two pieces of scale context matter. AVs run hundreds of thousands of trips a week against Uber's 300 million, under 0.5% of volume, and only 30% of US gross bookings come from the top 20 cities where they operate at all. The story is running a long way ahead of the numbers, in both directions: the revenue is immaterial today, and so is the threat.

Positives and negatives

Conclusion

With the broader market at highs, Uber remains one of the few large names trading well below its own. That gap is the setup. Growth is strong and the key metrics continue to improve, but the stock has been held back almost entirely by the AV competition story, Waymo above all.

I think that is overblown. The market is discounting a threat that currently amounts to under half a percent of trips, in a business where 70% of US bookings sit outside the twenty cities where AVs operate at all. Regulation is a genuine brake and will remain one. Meanwhile the business being marked down is compounding: bookings up 24%, adjusted EBITDA up 33%, margins widening, and trailing free cash flow through $10bn for the first time. Set against how much else has already re-rated, that combination makes Uber a more compelling entry than most of what is currently on offer, and a more comfortable position to hold at these prices.

The roadmap is the part the bear case tends to omit. None of it depends on Waymo's goodwill: seven live markets going to fifteen by year end, with NVIDIA, Rivian, Nuro, Lucid, Zoox, Wayve, Baidu, Pony and Verne all arriving between now and 2028. Uber is positioning itself as the software and commercialization layer on a platform that already owns the demand, which is a considerably better position than competing to win the driver-stack outright.

The risks deserve naming. Regulation remains a brake. Insurance is running as a tailwind this year and Uber is reinvesting the savings rather than banking them, so that support is not permanent. Competitive intensity in AV is the likeliest route to margin pressure. And this is discretionary consumer spending, so it will trade with the macro cycle. But at a price this far below the highs, with the operating business performing this well, I would rather carry that risk than the alternative.