Uber's Real Numbers Keep Getting Better. The Stock Keeps Getting Sold Anyway

Uber just posted the strongest quarter of its life — trailing-twelve-month free cash flow crossed $10 billion for the first time in company history, gross bookings grew 24%, and non-GAAP earnings per share rose 35% year over year. The stock, meanwhile, is down roughly 24% over the past year and about 12% in 2026, hovering near $72.5, far below its 52-week high near $102.

That gap is the whole story. It is also the reason the market may be pricing Uber's future wrong.

What the market is actually selling

The selloff has a clear pattern: every time a robotaxi headline hits — Tesla, Waymo, anyone — Uber shares get marked down. The fear is straightforward. If autonomous vehicles make human drivers obsolete, the argument goes, Uber loses the thing that makes it a business, because the robotaxi owners will cut out the middleman and run their own apps.

There is real substance to that fear, and it should not be waved away. Waymo, Uber's biggest autonomous partner, is exploring ending Uber's exclusive right to distribute its robotaxi rides in Austin and Atlanta by early 2028, and has said passengers could instead book directly through Waymo's own app. If a robotaxi operator owns the demand as well as the fleet, Uber's take rate is squeezed out of the middle. That is the bear case, and it is specific, not hypothetical.

But it is one partner in one strategy. The rest of the evidence points the other way.

The membership number that changes the economics

The clearest counterweight is Uber One. As of the first quarter, the subscription program had crossed 50 million paying members, up from roughly 30 million in mid-2025 — 20 million net adds in 15 months. More important than the headcount is what those members do: they now account for more than half of all Mobility and Delivery gross bookings, and they spend about three times as much as non-members.

This matters because it quietly changes what kind of company Uber is. A large base of subscribers who routinely order rides and food is recurring demand attached to a fee-bearing platform, not a series of one-off trips that could vanish if the supply changes. Uber One members are sticky users of the demand side of the network. Gross bookings are what Uber takes its cut from, and the cut compounds as engagement rises — monthly active platform consumers hit 208 million in the second quarter, up 16%, with more trips per user than a year earlier.

The other growth engine: sparse markets

The membership milestone sits alongside a second, less-noticed driver. About 20% of Uber's trips now originate in what the company calls "sparser markets" — the lower-density cities and suburbs outside the saturated urban cores — and those markets are growing faster than the biggest cities. Uber has said it is launching hundreds of new cities to capture the next wave. These are expansion markets with thin competition and the same fee economics per trip, which is why management describes them as an extension of the same runway rather than a bet on a new business.

Put the two pieces together and the picture is of a company widening the pool of demand it monetizes — more geographies, more recurring members — at a time when the market keeps pricing it as though the pool is about to empty.

The AV strategy is asset-light by design

The third piece is how Uber actually plans to compete with autonomous vehicles. Rather than build its own fleet, it is spreading bets across partners. The $10-billion-plus commitment Uber announced with second-quarter results funds an "aggregator" strategy across more than three dozen AV partners — Waymo, Wayve, Nuro, Avride, Volkswagen, WeRide and others — each deploying its vehicles on Uber's demand network in different cities, from Los Angeles to Riyadh.

The logic is that Uber's earnings power comes from the fee it takes on demand and operational services, not from owning vehicles or the software stack. So it is trying to be the platform that every robotaxi plugs into, collecting on rides whether the driver is human or autonomous. That is coherent — and it is the direct counterargument to the "robotaxis eat Uber" narrative — but it only holds if those partners keep routing their rides through the app. The Waymo exclusivity expiration is the concrete test of whether they will.

The math: a growth multiple, not a value trap

Here is where the valuation comes in. Uber trades at roughly 24x forward earnings even after the slide. That is not a fallen-angel price on a broken company — it is a reasonable multiple on a company compounding non-GAAP EPS at about 35% a year, which puts the price-to-earnings-growth ratio under one. Free cash flow is now running above $10 billion a year against a roughly $148 billion market cap, and capital expenditure is light because Uber does not own the cars. The second-quarter guidance that knocked the stock down 5% — non-GAAP EPS of $0.84 to $0.88, a touch below consensus — is a slowdown in the pace of beat-and-raise, not a deterioration in the business.

What would break the case

The honest frame is that the bull case rests on three facts all holding at once: members keep feeding more of each order's value to the platform, sparse-market expansion keeps adding fee-bearing demand, and AV partners keep using Uber's app even as some flirt with going direct. The first two are delivering today. The third is the genuine open question, and the Waymo exit from Austin and Atlanta by early 2028 is the date to watch.

Uber is not a cheap stock buying back yesterday's loss. It is a growing platform whose core metrics — membership, geographic breadth, and the take-rate engine on top of both — improved through a period when the market treated every robotaxi headline as a reason to sell. If the partnership model holds, the market has been pricing the wrong scenario. If it does not, the multiple will not protect you. Either way, the divergence between the improving math and the falling stock is the only number that matters.