Uber is building a data operation that could give it leverage over the same self-driving companies it partners with. That’s the standout signal from a new TechCrunch AI interview with Uber Chief Product Officer Sachin Kansal, who walked through the company’s expansion into hotels, financial services, and autonomous vehicles. The piece, as detailed in TechCrunch AI, shows a company quietly reshaping what it wants to be.
The headline most people missed: a six-month-old unit called AV Labs. It’s a fleet of sensor-equipped vehicles, separate from Uber’s regular driver network, built to gather driving data at scale.
Why the data play matters
Uber frames AV Labs as a way to strengthen ties with autonomous vehicle partners, several of which it holds equity in. But look closer and it reads like a hedge. Uber competes directly with some of those partners, Waymo chief among them. Owning the data layer gives Uber two things it badly wants: leverage and optionality.
This is significant because the robotaxi race is turning into a question of who controls the inputs. Models need driving data. Lots of it. If Uber becomes a neutral data source that every AV player depends on, it stays valuable no matter which company wins the car itself. That’s a smart position to hold when your biggest partner is also your biggest threat.
The ‘everything app’ question, answered carefully
Asia has super-apps like Grab that do rides, payments, food, and more inside one shell. Is Uber headed there? Kansal’s answer is a deliberate no.
“We’re not trying to be everything to everyone,” he told TechCrunch AI. The strategy is more disciplined than that:
- Travel as the third leg. Rides, then Eats, now travel. Uber says 1.5 billion trips a year happen outside a user’s home city, so hotels (via an Expedia partnership), airport rides, and food fit a real pattern.
- Partner first, integrate later. Boat rentals in Europe hand users off to a partner’s booking flow. Hotels got a deep, custom-built integration. The rule: test with a light handoff, integrate deeply only when traction shows up.
- Let experts be experts. No plans for an in-house buy now, pay later product. Uber would rather plug in partners at checkout.
What stands out is the restraint. Uber is expanding fast, but it’s picking where to own the experience and where to rent it.
Financial services, but for earners first
The money angle is real, just not aimed where you’d expect. Uber’s financial products lead with drivers and couriers, not consumers. The Uber Pro debit card lets earners route their pay into one account. Merchant products are in early testing in some regions. For consumers, the currency is Uber credits tied to the membership program, not a bank account.
That membership program is doing heavy lifting. Uber One now has 51 million members and accounts for roughly half of bookings. On delivery, it takes two to three orders to cover the monthly fee, and members are cross-buying: mobility-only users start ordering food, delivery-only users start taking rides. Uber Eats, long one of the hardest businesses in tech to make profitable, is now independently profitable, according to Kansal.
What to watch
For anyone building or investing in AI, the AV Labs move is the tell. The value in autonomous driving is shifting toward whoever controls the data pipeline, not just the vehicle. Uber is positioning to be that layer.
A few practical takeaways:
- If you partner with a company that could become a rival, own something they can’t easily replace. For Uber, that’s data and demand.
- Expansion doesn’t require building everything. The partner-first, integrate-later model lets Uber test new markets cheaply before committing engineering.
- Membership is the cross-sell engine. Habit inside one product pulls users into the others. That flywheel is where the profit compounds.
The open question is how the Waymo relationship holds up as Uber’s data ambitions grow. Partner and competitor at once is a hard balance to keep. You can find the full interview at the original source.