Uber’s product chief on hotels, robotaxis, and why the company doesn’t want to be “everything for everyone”

Uber's product chief Sachin Kansal will not call his company a super-app. He says it plain: "We don't want to be everything for everyone." The line lands like a correction. For years the press has measured Uber against Grab and Gojek, checking off boxes — food, rides, payments, shipping — as if feature parity were the same thing as strategy. Kansal refuses the frame. His product roadmap does not chase breadth for breadth's sake. It chases leverage.

The most revealing move Uber made this year never hit a consumer screen. AV Labs, a six-month-old business unit, is building its own fleet of sensor-laden vehicles to hoover driving data at scale. This is not R&D for Uber's own robotaxi — the company sold that unit to Aurora in 2020. AV Labs exists to give Uber a data layer it owns outright, independent of Waymo, Cruise, or any other autonomy partner whose cars already ply Uber's network. The leverage is naked. Uber holds equity in several of these partners. It also competes with them. Waymo's own app now hauls riders in Phoenix and San Francisco. Every mile Waymo drives on Uber's platform is a mile Waymo learns without Uber. AV Labs flips that asymmetry. Uber collects the sensor feed. Uber decides who sees it. Kansal calls it "strengthening relationships." The rest of us can call it a hedge.

The consumer-facing launches — hotels via Expedia, boat rentals in Europe, a "shop for me" concierge that scrapes local stores not on Uber Eats — look like super-app sprinkles. Kansal insists they are not. He anchors them to a single number: 1.5 billion trips a year that originate outside a user's home city. That is the travel wedge. Riders already summon Uber from airports to hotels. They already order dinner via Uber Eats when room service disappoints. The new features stitch those existing behaviors into a revenue stream Uber currently lets Expedia or DoorDash capture. The strategy is parasitic in the best sense: attach to habits the network already produces, then monetize the seam.

Financial services follow the same logic — but inverted. The "Uber Pro Card" debit product targets drivers and couriers, not riders. Kansal ticks off the constituencies: consumers, drivers, merchants. Only the second group gets a dedicated banking stack today. That choice is deliberate. Drivers are captive. They earn on Uber's rails. They need instant payouts, expense tracking, credit access. Uber owns the income data; underwriting becomes trivial. The consumer side stays thin — no digital wallet, no buy-now-pay-later checkout. Kansal won't rule it out, but he won't lead with it. The super-app playbook says own the consumer's financial identity. Uber's playbook says own the worker's financial plumbing first.

AI shows up where it compounds network effects, not where it makes demo videos. Kansal describes a routing system that now predicts which driver will accept a trip before the request even dispatches — cutting cancellations by double-digit percentages. He mentions a support bot that resolves driver deactivations in minutes instead of days. These are dull sentences that move money. The flashy generative features — travel itineraries, meal planning — stay in experiment lanes. Uber's AI budget buys throughput, not press releases.

The Waymo relationship sits at the center of every strategic calculation. Uber needs Waymo's robotaxis to supply capacity it cannot hire. Waymo needs Uber's demand density to keep utilization high. Both need the other less each quarter. Waymo's app grows. Uber's AV Labs fleet grows. The equity stakes Uber holds in autonomy startups look less like partnerships and more like option contracts on a future where only one platform survives the transition. Kansal dances around this. He stresses "multi-modal" supply, "partner diversity," "consumer choice." The subtext is a knife fight.

Kansal's clearest signal is what he refuses to build. No social feed. No content studio. No marketplace for freelance services beyond driving and delivery. No consumer credit card. Each omission is a bet that focus compounds faster than breadth. The Asian super-apps grew in markets where banking, logistics, and communications infrastructure were absent — so they built all three. The U.S. has Venmo, Zelle, Amazon, iMessage. Uber gains nothing by replicating them badly. It gains everything by becoming the logistic layer they all plug into.

The hotel partnership with Expedia illustrates the discipline. Uber did not build a booking engine. It did not acquire a hotel chain. It embedded a white-label catalog, took a referral cut, and called it travel. The margin is thin. The operational drag is near zero. The data — where riders sleep, when they arrive, what they eat nearby — feeds the same routing and demand models that power the core rides business. That is the whole game: every new vertical must sharpen the horizontal.

Investors should read the omissions as loudly as the launches. No AV Labs data shared with partners yet. No consumer wallet. No timeline for either. Kansal's "we don't want to be everything for everyone" is not modesty. It is a capital allocation statement. Every engineer hired for a consumer finance stack is an engineer not hardening the driver payout rails. Every PM assigned to a social feature is a PM not integrating the next autonomy partner's API. Uber's moat is not its brand. It is the density of its two-sided network and the data exhaust that density produces. AV Labs protects that exhaust. The travel features harvest it. The driver debit card locks the supply side down.

The "everything app" question bores Kansal because it mistakes Uber's asset. Grab became a bank because Indonesian drivers needed banking. Uber becomes a bank for drivers because American drivers need instant liquidity — and Uber already moves their money. The product is the same. The reasoning is not. One chases total addressable market. The other chases structural indispensability. Kansal has chosen the latter. The market will decide if he chose well.