Uber Cuts 3,300 Jobs, About 10% of Staff, and Its Memo Never Mentions AI

Uber will cut about 10 percent of its workforce, roughly 3,300 roles, chief executive Dara Khosrowshahi told staff in a memo published on Uber's newsroom on Wednesday 2 September 2026. It is the company's largest reduction since the pandemic, according to Al Jazeera, and it targets management layers rather than drivers. The memo does not mention artificial intelligence. It arrived four weeks after Uber reported second-quarter revenue of $14.2 billion, up 12 percent, and $2.4 billion of net income, and three months after Bloomberg reported that Uber had capped spending on some AI coding tools after using up its annual AI budget. Whether these are "AI layoffs" depends on which document you read.

What happened

Khosrowshahi's memo, titled "Building a simpler, faster Uber", argues that growth brought "more layers, more coordination, more fragmented ownership". It says the company will reduce team size by about 10 percent, to "make Uber simpler and faster, and create more capacity to invest in our future". The structural changes it lists include:

• Reducing the number of employees seven or more layers below the CEO by about 20 percent.
• Cutting "micro-teams", where a manager has one or two direct reports, by nearly half.
• Merging three delivery operations teams (restaurants, retail and direct) into single-threaded teams at global, regional and country level.
• Requiring most remote staff to move to an office, leaving about 1 percent of employees remote.

Bloomberg, as published by The National, reported that manager roles fall by 20 percent, with some managers moved into individual contributor roles, and that the cuts also reach non-managers. Uber also confirmed to Al Jazeera that it was leaving Nigeria and Uganda. The memo says the savings will be reinvested, names drivers, couriers and merchants among the areas to invest in, and says Uber has the chance to "build the autonomous future".

The numbers

Reports disagree on Uber's total headcount before and after the cut, so I have left it out.

Why it matters

The AI context is all around this announcement, even if the memo avoids it. Gizmodo pointed to three facts. Bloomberg reported in June that Uber put spending caps on some AI tools after exhausting its annual AI budget in "just a few months"; follow-up coverage named Claude Code and Cursor among the capped tools. Khosrowshahi had said earlier in the year that about 10 percent of Uber's code was being built by AI agents. And Uber had said it would slow hiring because of AI. Al Jazeera added that the July customer service cuts came as Uber "embraced artificial intelligence", per Bloomberg.

The cut is also a clear case of "flattening": removing managers with very few direct reports. That pattern has run through 2026 at Amazon, Meta and Block, whose chief executive Jack Dorsey said his company's tools enabled "smaller and flatter teams", as Gizmodo noted. The argument underneath is that AI tools increase what an individual contributor can do and reduce the coordination a manager provides. Nobody has shown that with data, but it is now a standard line in layoff memos.

For the Challenger layoff tracker, this kind of announcement is ambiguous. Challenger counts the reason employers give, and Uber gave organisational complexity. I covered how that has changed the year's AI-attribution numbers in AI-Cited Job Cuts Fell to 1,265 in October.

What sceptics say

Taken at face value, the memo describes a familiar mid-size company problem: too many layers and too many tiny teams after years of growth. Neither needs AI as an explanation. Uber is also funding a large autonomous vehicle push and faces direct competition from Waymo, which Gizmodo says runs fully driverless rides in 14 US cities. Cutting operating cost to fund capital-intensive bets is ordinary corporate finance.

There is a fair point the other way too. If AI coding tools were so obviously replacing engineering labour, Uber would not have been rationing them by June. A company that runs out of its AI tooling budget is a company where AI is adding cost, at least before any saving shows up in headcount.

What it means for engineering and DevOps teams

My take, as opinion: the most useful detail here for engineers is the budget cap, not the layoff. It shows that AI coding agents are now a line item large enough for finance to notice and ration, at a company with a strong engineering culture.

Practical points:

• Treat agent spend like cloud spend. Per-team budgets, tagging and alerts before the money is gone, not after. The same FinOps for AI discipline applies to coding tools as to inference.
• Flatter orgs push work down. Fewer managers means more planning, prioritisation and incident coordination done by senior individual contributors. On-call rotations and ownership maps need updating when the reporting lines change, or incidents bounce between teams that no longer exist.
• Measure what the agents actually ship. "10 percent of code written by agents" is an input metric. Lead time, change failure rate and review load are the outputs. If you are being asked to justify agent seats, measure DORA metrics from GitHub before and after.
• Return-to-office changes on-call. A 1 percent remote cap will shift who can carry a pager overnight and from where. That is worth raising before it causes problems.

FAQ

How many people did Uber lay off in September 2026?

About 3,300, or roughly 10 percent of its workforce, announced in a memo on 2 September 2026. The cuts focus on management layers, with manager roles reduced by about 20 percent and micro-teams by nearly half.

Did Uber blame AI for the layoffs?

No. The memo cites organisational complexity and reinvestment in drivers, couriers, merchants and autonomous vehicles. AI enters through context: earlier AI-linked customer service cuts, a slowdown in hiring attributed to AI, and caps on AI tool spending reported by Bloomberg.

Why did Uber cap AI coding tools?

Bloomberg reported in June 2026 that Uber exhausted its annual AI budget within a few months and put spending caps on some tools, including Claude Code. Uber's reasoning has not been published in full.

Related: Coding Agents in 2026, FinOps for AI in 2026 and CI/CD in 2026: Agents Opened 1M PRs in 5 Months.