Uber's layoffs weren't a performance problem. They were an intake problem nobody revisited.

Uber cut 3,300 corporate jobs this week — about 10% of its non-driving workforce, its largest reduction since the pandemic. But the headcount number isn't the interesting part. The interesting part is what CEO Dara Khosrowshahi said was actually being removed: not underperformers, but structure.

In his memo to staff, Khosrowshahi described years of growth that had produced "more complexity: more layers, more coordination, more fragmented ownership, and in some cases structures that made sense when businesses were smaller but no longer serve us well at our current scale." The specifics back that up. Uber cut its "micro-teams" — units with only one or two direct reports — by nearly 50%. It cut staff sitting seven or more layers below the CEO by 20%. This wasn't a P&L-driven purge of a bad division. It was a correction for years of org-chart accretion that nobody had gone back to check on.

Nothing in an org chart ever comes up for renewal

Here's the asymmetry that produced Uber's memo: approving a new team, a new layer of management, a new sub-function takes one meeting and one yes. Un-approving it takes nothing, by design — there's no calendar reminder, no recurring gate, no moment where someone is required to ask "does this still deserve five people and a manager?" So org structure only ever grows. Every micro-team Uber just eliminated started as somebody's reasonable answer to a real problem, at a specific moment, at a specific scale. The problem changed. The approval never got revisited. Multiply that by nine years of hypergrowth and you get exactly what Khosrowshahi described — and exactly what a "Great Flattening" now rippling across tech, as Fast Company has called it, looks like from the inside.

This is the same failure mode that shows up in project backlogs, just with a longer fuse. A request gets approved, funded, staffed. It ships, or it stalls, or the market shifts underneath it — and nobody circles back, because the intake process only has one gate, and it's at the front. The backlog doesn't shrink on its own any more than an org chart does. It just accumulates commitments that made sense once, until someone finally audits the whole thing under duress and calls it a "restructuring."

The fix isn't a better purge. It's a recurring gate.

The lesson companies tend to take from a layoff cycle is "we should cut harder next time." The actual lesson is that the cutting should never have been saved up for one traumatic event in the first place. If teams, initiatives, and projects had to clear a scoring bar not just at creation but on a recurring cadence — does this still deliver relative to what it costs, given what else is competing for the same budget — the correction happens continuously, in small increments, instead of arriving as a memo to 3,300 people nine years later.

That's the part of intake most organizations get only half right. They'll happily build a process for deciding what gets approved. Almost none of them build the matching process for deciding what stops being approved. Admitura treats both as the same problem: every request gets scored against the same weighted criteria going in, and the record of that score is what makes it possible to ask, later, whether the thing it justified is still earning its keep — instead of finding out nine years and one layoff memo at a time.

Sources: Uber is laying off 10% of staff, or 3,300 people, TechCrunch, September 2, 2026; The 'Great Flattening' rolls on as Uber lays off middle managers, Fast Company, September 2026.