Uber Finally Turned a Profit, Just as the Cars Learned to Drive Themselves

---

Begin with the genuine triumph, because Uber's transformation is real and the question here is about the future, not the present. For most of its existence, Uber was the emblem of growth-at-all-costs unprofitability — a company that moved millions of people but lost money doing it, subsidizing rides and drivers to buy market share. That era is over. In the first quarter of 2026, Uber reported gross bookings of $53.7 billion, up 21%; revenue of $13.2 billion, up 14%; free cash flow of $2.3 billion; non-GAAP net income of $1.5 billion, up 39%; and fifty million Uber One subscribers. Mobility and delivery both grew briskly. This is a genuinely profitable, cash-generative, dominant platform, and the market's decision to re-rate it from a speculative growth story into a quality compounder is justified by the operating results. Uber won the ride-hailing war, achieved scale, and turned that scale into profit. On its own terms, the present is a success.

The question this essay examines is not whether Uber is profitable today — it is — but whether the foundation of that profitability survives the transition to autonomous vehicles, which is now underway. Because Uber's profit comes from a specific structure: it is a two-sided marketplace, owning neither the cars nor the labor, that takes a cut for matching demand (riders) with supply (drivers). That structure has been enormously valuable in a world of human drivers. Whether it remains valuable in a world of robot ones is the trillion-dollar question, and the answer is genuinely unknown — which is exactly why pricing the stock as though the answer were settled is the risk.

What Uber actually owns, and what it doesn't

To understand the autonomous question, you have to be precise about what Uber's moat actually is — because the moat is not what casual observers think. Uber does not own cars; the drivers own the cars. Uber does not employ the drivers; they are contractors. What Uber owns is the marketplace — the network of riders and the network of drivers, and the sophisticated technology that matches them in real time, at scale, across the world. Its moat is the two-sided network effect: riders use Uber because there are always drivers available, and drivers use Uber because there are always riders, and this self-reinforcing loop, built over more than a decade across hundreds of cities, is extraordinarily hard for a competitor to replicate, because you cannot bootstrap both sides of a marketplace at once.

This is a real and powerful moat — in a world of human drivers. The drivers are an effectively infinite, self-supplying, capital-free pool of supply: anyone with a car can become an Uber driver, the drivers bear the cost of the vehicles, and Uber simply skims a percentage of the transaction. Uber's genius was building the demand network and the matching technology while letting millions of independent drivers provide the supply at their own expense. The company captured the value of coordination without bearing the cost of the assets or the labor — one of the great business models of the platform era.

Now ask what happens to that structure when the supply side — the drivers — is replaced by autonomous vehicles. The infinite, self-supplying, capital-free pool of human drivers is replaced by fleets of expensive robot cars that someone has to build, own, and operate. And that someone — Waymo, owned by Google; Tesla; or another AV developer — is a very different kind of supplier than a gig driver with a Honda. Unlike the millions of atomized, powerless individual drivers, the AV operators are enormous, sophisticated, well-capitalized companies that own their supply and can make their own strategic choices about how to distribute it. The supply side, in the autonomous world, consolidates from millions of replaceable individuals into a handful of powerful corporations — and that changes everything about Uber's leverage.

The disintermediation question

Here is the existential crux. When the supply side is a handful of powerful AV operators rather than millions of gig drivers, those operators face a choice that no individual driver ever had the power to make: distribute their rides through Uber's demand network and pay Uber a cut, or build their own app and sell rides directly to riders, keeping all the economics. This is the disintermediation question, and it is the whole game.

The case for disintermediation is straightforward and already visible. Tesla is explicitly building its own robotaxi network and its own app, designed to connect Tesla-owned autonomous vehicles directly with riders, bypassing Uber entirely — a fully vertically integrated model where the same company makes the car, runs the autonomy, and owns the customer relationship. Waymo, while it partners with Uber in some cities, also operates its own direct-to-consumer app, Waymo One, in others — meaning even Uber's flagship AV partner is hedging, distributing through Uber where convenient while building its own demand channel where it can. The logic is obvious: if you have spent billions building autonomous vehicles, why hand a cut of every ride to a marketplace when you could own the customer yourself? The most valuable AV operators have both the incentive and the capability to go direct, and the more dominant autonomy becomes, the more the supply that Uber's marketplace depends on belongs to companies that would rather not share it.

The clearest sign that Uber itself takes the disintermediation risk seriously is where it is putting its money. Even as it insists autonomy is a tailwind, Uber is reportedly pouring more than $10 billion into building and owning robotaxi alternatives — funding partners like Rivian and others to create AV supply that Uber can control rather than merely rent from Waymo or be bypassed by Tesla. A company truly confident that the AV operators will happily distribute through its app would not need to spend $10 billion securing its own captive supply. The investment is a hedge, and a hedge is an admission that the partnership model is fragile — that Uber knows the supply owners may go direct, and is trying to ensure it owns enough of the supply itself that it cannot be cut out entirely. That is prudent, but it is also expensive, capital-intensive in exactly the asset-light-marketplace way Uber spent a decade avoiding, and a tacit concession that the demand-layer moat the bulls celebrate may not hold on its own.

If that disintermediation plays out, Uber's network-effect moat erodes from the supply side. A marketplace that owns neither the cars nor an exclusive lock on supply, in a world where the supply owners can and do sell directly, is a far weaker business than the one the current valuation prices. The riders Uber aggregates are valuable only as long as the AV operators need Uber to reach them — and a vertically integrated operator with its own app does not.

Uber's counter — and why it might be right

In fairness, Uber's rebuttal is serious and may well prove correct, and the company makes it with conviction. The argument runs like this: autonomy does not shrink Uber's opportunity; it amplifies it. Building and operating autonomous fleets is enormously capital-intensive and geographically constrained — an AV operator can realistically serve only a handful of cities at a time, and even within those cities, its fleet will have idle capacity at off-peak hours and underserved demand at peak. Uber, by contrast, has global demand density everywhere, fifty million subscribers, and a decade of experience matching supply to demand in real time. So Uber's pitch to the AV operators is: you focus on building great autonomous cars; let us fill your empty seats with our enormous, always-on demand, maximizing your fleet utilization and your returns. In this telling, the AV operators need Uber's demand network to make their expensive fleets economical, and Uber becomes the indispensable demand layer that every AV fleet plugs into — its marketplace more valuable in the autonomous world, not less, because matching becomes harder and demand density becomes more precious.

There is real evidence for this view. Uber's autonomous trips grew more than tenfold year over year; it has more than two dozen AV partners and live deployments in multiple cities; and Waymo, now delivering on the order of 250,000 paid rides a week and rising, routes a meaningful and growing share of them through Uber's app — a genuine, scaling, revenue-generating proof that at least one major AV operator finds value in Uber's demand network. The Uber-Rivian partnership to deploy tens of thousands of robotaxis points the same way. The chief executive, Dara Khosrowshahi, calls autonomy a "net positive" and points to data showing stronger rider behavior in the cities where AVs are already live. If the total mobility market expands dramatically as autonomy makes rides cheaper and more available, and Uber remains the platform connecting riders to the cheapest available vehicle — human or robot — then Uber benefits from the expansion regardless of who builds the cars. That is a coherent, evidence-backed, genuinely plausible bull case.

The honest synthesis is that both stories are credible and the outcome is genuinely uncertain. Uber may become the indispensable demand layer for a vastly expanded autonomous mobility market — or it may be disintermediated by vertically integrated AV operators who own their supply and their customers. The evidence today supports the optimistic view at small scale (Waymo through Uber works), but the scale is tiny relative to Uber's billions of trips, and the most important potential competitor — Tesla — is building explicitly to bypass Uber. The question is not settled; it is the opposite of settled. And that is precisely the problem with the price.

The market has priced the answer it doesn't have

The investment risk is not that the bear case is certain — it isn't — but that the market has re-rated Uber as a clean, durable profit machine, a quality compounder deserving a quality compounder's multiple, as though the autonomous question had been resolved in Uber's favor. It has not been resolved. The re-rating treats Uber's demand-network moat as durable through the autonomous transition, when the durability of that moat is the single largest open question about the company, with credible, well-funded forces — Tesla above all — working to prove it isn't.

This is the classic error of pricing a binary, unresolved, existential question as though it were already answered. Uber today is genuinely profitable on the strength of a human-driver business model whose long-term future is uncertain, and the stock's quality-compounder valuation embeds the assumption that the model survives and thrives through autonomy. If the optimistic case is right, the valuation is fine and possibly cheap, because the market expands and Uber rides it. If the disintermediation case is right, Uber's moat erodes over years as the supply consolidates into operators who go direct, and the quality-compounder multiple was a serious overpayment. The market has chosen the first scenario as its base case — but a base case for an unresolved existential question is not analysis; it is a bet on optimism, and the price reflects optimism rather than the genuine, unresolved uncertainty that an honest assessment requires.

The transition is slow, which cuts both ways

There is a temporal dimension that matters for how this resolves, and it genuinely supports a measure of calm: the autonomous transition is going to take many years, probably more than a decade, to reach meaningful scale. Put the numbers in proportion. Waymo's roughly 250,000 paid rides a week — only a portion of which run through Uber — impressive as the growth rate is, is a rounding error against Uber's roughly 3.6 billion trips a quarter — autonomous trips are a fraction of a fraction of total volume. Self-driving cars remain geographically limited to a handful of favorable cities, constrained by regulation, weather, cost, and the sheer difficulty of the technology, and scaling a robot fleet to cover the world the way millions of human drivers already do is a multi-decade, capital-devouring undertaking. Human drivers will be the overwhelming majority of Uber's supply for years to come.

This slowness cuts both ways, and honesty requires naming both edges. For the bulls, it means Uber has abundant time and abundant cash flow to adapt, to deepen its AV partnerships, and to entrench itself as the demand layer before disintermediation could ever become material — the threat, even if real, is distant and gradual, not a cliff. For the bears, it means the market is paying a quality-compounder multiple today for a profit stream whose long-term foundation is under a slow but unmistakable erosion, and "slow erosion" is exactly the kind of risk a stock priced for permanence is least prepared for, because it arrives quarter by quarter rather than in a single shock that forces a reckoning. The transition's slowness is why Uber's profits are safe this year and next; it is also why the market feels free to ignore a question whose answer will not be forced for a decade — and decade-out questions, left unpriced, are precisely how richly valued stocks set themselves up for a long, grinding disappointment when the future finally arrives.

The delivery business is the real diversification

It is worth crediting the part of Uber that the autonomous debate ignores entirely: Uber Eats and the broader delivery business, which grew gross bookings 23% in the quarter and represents a genuine second engine less directly exposed to the robotaxi question. Delivery is its own large, growing marketplace — matching restaurants and merchants with couriers and customers — and while autonomous delivery (sidewalk robots, drones) is a long-term factor, the near-term disruption risk is far lower than in ride-hailing, and the same demand network and logistics technology that power mobility extend naturally to delivery. A meaningful and growing share of Uber's value sits in this business, which diversifies the company beyond the pure ride-hailing-versus-autonomy question and gives it a second foundation if the mobility moat is tested. The bear case on autonomy is a case about Uber's mobility business; it is not the whole company, and the delivery engine is a real and underappreciated hedge that a single-issue focus on robotaxis tends to overlook.

What the bulls genuinely get right

To be fair to the present, several things genuinely favor Uber and should weigh heavily. The profitability and cash generation are real and substantial — $2.3 billion of quarterly free cash flow is not an accounting illusion, and it gives Uber the resources and the time to navigate the autonomous transition from a position of strength rather than desperation. The network effect, in the human-driver world, remains a real and powerful moat that competitors have repeatedly failed to breach. The delivery business (Uber Eats) is a genuine second engine, diversifying the company beyond pure ride-hailing. The Uber One subscription, at fifty million members, deepens customer loyalty and recurring engagement. And the bull case on autonomy — that Uber becomes the demand layer for an expanded market — is not a fantasy; it is a serious thesis with early supporting evidence and a credible champion in management. Uber is a strong company with real options, and it is far better positioned to win the autonomous future than the cash-burning Uber of the past would have been.

The point is not that Uber loses; it is that whether it wins or loses is genuinely unknown, hinges on choices its most powerful potential competitors have not yet made, and is being priced as though the favorable outcome were already secured. A company facing a real, binary, existential question deserves a valuation that reflects the uncertainty — and Uber's, after the re-rating, reflects the resolution.

The kicker

Uber's story is one of the great corporate turnarounds of the era: a company that the market once doubted could ever make money now prints billions in cash and dominates its industry. That achievement is real and earned. But it arrives at a hinge moment, because the entire business was built on a structure — a marketplace skimming a cut from millions of independent human drivers — that autonomy directly disrupts, replacing the infinite, powerless, self-supplying driver pool with a handful of powerful companies that own their robot fleets and can choose whether Uber lives in their future or not. Uber insists it will be the indispensable demand layer; Tesla is building to prove it won't; Waymo is hedging both ways; and the honest answer is that the question is wide open. The market, dazzled by the profitability it waited fifteen years to see, has decided the autonomous question is settled in Uber's favor and priced it accordingly. It is not settled. Uber won the war it was fighting, exactly as the war it will actually have to fight began — and the stock is priced for the first victory as though it guaranteed the second.

Uber spent a decade and a fortune proving a marketplace of human drivers could make money, and the quarter it finally did, the drivers started disappearing — replaced by robot fleets owned by the very companies that might prefer to keep the riders for themselves. The profit is real. The moat is real. Whether the moat survives the moment the supply learns to drive itself is the only question that matters, and it is the one the price has quietly decided not to ask.