Uber Cuts 3,300 Jobs to Fund a Robotaxi Future It Has No Intention of Building

On 2 September 2026, Uber announced it was cutting about 3,300 roles — roughly 10% of its global workforce. Management layers take the heaviest hit, with the number of managers falling by around 20%. CEO Dara Khosrowshahi delivered the news by internal email.

It is Uber's largest reduction since the pandemic. And unlike 2020, the company is not in trouble.

Cutting Into Growth

This is the part that makes the announcement worth reading twice: Uber is growing at double digits while doing this. Rides and delivery are not collapsing. There is no crisis being managed.

That reframes the whole exercise. These are not survival cuts — they are reallocation cuts. Money and headcount are being moved out of one version of the company and into another, while the current version is still working perfectly well.

Which is, historically, the hardest kind of decision for a large company to make and the one most often deferred until it is too late.

Where the Money Is Going

Uber has committed more than $10 billion over the coming years to autonomous vehicles — covering robotaxi fleets and equity stakes in self-driving developers. Partners have committed roughly 120,000 vehicles.

Note what is absent from that sentence: Uber is not building the self-driving technology. It sold its in-house autonomy unit years ago and has not gone back.

The Actual Strategy: Own the Boring Part

Look at what Uber already does for Waymo vehicles running through the Uber app in Atlanta and Austin. Uber handles:

Charging
Cleaning
Inspections
Demand, dispatch and the customer relationship

Waymo supplies the intelligence. Uber supplies everything a fleet needs in order to be a business. That is a deliberate division of labour, and it is worth being precise about who is exposed to what:

Layer	Who owns it	Risk profile
Autonomy software	Waymo and other developers	Enormous R&D cost, regulatory exposure, winner-takes-most
Vehicles and capital	Partners, plus Uber equity stakes	Heavy capex, depreciation
Fleet operations	Uber	Unglamorous, low margin, hard to replicate at scale
Demand and dispatch	Uber	Existing asset, already profitable

Uber is betting it does not need to win the self-driving race. It needs to be the place the winners come to sell rides — and the operator they cannot economically replace.

If autonomy commoditises, that is a very good position. Several developers competing on software, all of them needing depots, cleaning, charging and a demand engine, and only one company already running that at national scale.

Why Managers Specifically

The 20% cut to management is not incidental. A platform coordinating third-party fleets needs fewer coordination layers than a company running hundreds of thousands of human drivers through regional operations teams.

Human-driver logistics is a people-management problem. Robotaxi logistics is largely a software and depot problem. The org chart is being rebuilt for the second one before it fully arrives.

The Risk Nobody Should Skip

Three things could break this:

An autonomy winner goes direct. If one developer achieves a decisive lead, it may prefer its own app to Uber's aggregation — the fleet operations moat is real but not infinite.
Timelines slip. Robotaxis have been consistently later than promised. Uber has now cut capacity against a schedule it does not control.
Regulation bites unevenly. Autonomous deployment is being decided city by city and, as this week's federal audit of Tesla's Cybercab showed, the rules are still being written in public.

The Bottom Line

The headline is 3,300 jobs. The story is a profitable company deliberately shrinking the org built for its current business in order to pay for the one it expects next — and choosing to own the depots, the demand and the drudgery rather than the algorithm.

It is a bet that in a robotaxi market, the least glamorous layer is the most defensible one. History suggests that is usually right, and always uncomfortable to be early on.