Uber Just Capped Remote at 1%. Source the Relocation Refusers.

On September 2, 2026, Dara Khosrowshahi told Uber it was cutting 3,300 people, roughly 10% of staff, and in the same memo tightened remote work down to about 1% of the workforce. Every sourcer on LinkedIn is already fighting over those 3,300 names. The interesting pool is the one that has no name yet: the engineers who still show "@ Uber" on their profile today and will quietly quit in the next 60 days rather than relocate.

The 3,300 layoff list is the crowded trade

The named layoffs are the worst list to work this week because every recruiter in your city is already working it. The relocate-or-leave cohort created by Uber's remote-work cap is 1.4x to 2.4x larger, invisible on any layoff tracker, and drawing zero competing InMails.

Here is the math. Uber corporate headcount sits near 33,000 after the cut. Pre-policy, remote and hybrid-remote employees at large tech employers typically run 15% to 25% of the workforce. The new policy targets roughly 1%. That implies 4,600 to 8,000 employees now facing a choice: move within commuting distance of an office, or leave.

Compare that to the 3,300 layoff list, which is already:

• Blasted to every recruiter's saved search on LinkedIn
• Being scraped nightly into layoffs.fyi mirrors
• Sitting in the inboxes of every FAANG-adjacent in-house team
• Drawing dozens of InMails per profile in the first week

The relocation refusers are on none of those lists. They still carry a current Uber title. Their "Open to Work" banner is off. And they are the ones sending resumes at 11pm this week.

4,600 - 8,000

Why this pool is high quality, not distressed

Uber's Q2 2026 revenue grew 12% year over year to $14.2 billion, and full-year 2025 revenue hit $52 billion, up 18% from 2024. This is not a distressed company shedding weak performers. It is a growth company reshuffling org design, which means the talent hitting the market did not get cut for being replaceable.

Khosrowshahi's memo did not mention AI once. The framing was management bloat: cut 20% of managers, halve the number of teams with only one or two members, and remove anyone "more than seven layers down from the CEO." Translation for sourcers: these engineers are leaving because they reported to the wrong VP, not because their skills got obsoleted by a model.

That distinction matters when you write the first message. You are not reaching out to a researcher whose lab got zeroed. You are reaching out to a Staff engineer on a merged delivery ops team who is about to lose their weekends to a Bay Area commute they did not sign up for. Different pain, different pitch.

The signal is geography mismatch, not "Open to Work"

Skip the green banner. The high-signal filter is an Uber engineer whose location field lists a city where Uber has no anchor office, while their team's designated anchor days are in San Francisco, New York, or Seattle. That mismatch is the tell.

From Blind chatter on the Uber board, the concentrations of remote employees who will be forced to relocate skew toward Denver and Colorado Springs and Upstate New York, with adjacent clusters in Austin, the Research Triangle, South Florida, and the Mountain West.

The workflow: pull current Uber engineers by title, cross-reference profile location against the office list, and prioritize anyone in a city with no Uber engineering anchor. This is the exact kind of query that traditional Boolean strings fumble because "location" on LinkedIn is self-reported and often stale. Describing the shape of the person in plain English is faster, which is the gap Refolk closes: you ask for "Uber staff engineers in metros without an Uber office" and get back the ranked list with the mismatches surfaced.

Merged teams surface the next domino

Two engineering merges inside the September 2 restructuring produced a specific pool of survivors who are the next to leave: the Core Services Engineering and Science combination, and the three-way merger of the Restaurants, Retail, and Direct Commerce delivery ops groups. Duplicate senior ICs on these teams survived the first cut and are doing two people's jobs by Tuesday.

Prioritize profiles whose current title contains any of:

1. "Staff Engineer, Delivery" or "Staff Engineer, Delivery Ops"
2. "Core Services Engineering"
3. "Restaurants Engineering," "Retail Engineering," or "Grocery"
4. "Direct Commerce"
5. Any Engineering Manager title on a team the merger consolidates - many of these are the "20% manager cut" candidates being told to return as ICs

The seven-layers-deep rule is a second filter worth applying. Uber cut everyone more than seven layers from the CEO. Anyone still standing at Staff or Principal is, by construction, near the top of an engineering ladder at a company that historically overpaid for talent. That is a signal no one else is applying because no one else is reading org-design memos with a sourcing hat on.

The Refolk index: what is actually there right now

Refolk's index shows roughly 3,123 current Uber engineers at Software Engineer, Senior, Staff, and Engineering Manager titles globally. The distribution matters for anyone trying to build a US-focused list, because most of the volume is not in the US.

Two things fall out of this table. First, India is the biggest geographic concentration of Uber engineering, and those hubs are office-anchored already, so the RTO policy barely moves them. The differentiated pool is the US remote engineer. Second, only 19 of 25 sampled profiles were still tagged "Uber" as the current employer. The rest already show partner brands like Uber Freight and Uber AI Solutions. Those sub-orgs may have different remote policies and different attrition timing, and they are worth their own saved search.

The layoff list has dozens of InMails per profile by Friday. The relocation refusers have zero.

The 60-day sourcing window

You have about 60 days from the September 2 memo before this pool is public knowledge and the arbitrage closes. Relocation deadlines at comparable RTO rollouts (Amazon's 5-day mandate in 2024, Dell's return mandate, JPMorgan's 5-day rule that drew a 950-signature petition against a 317,000-person workforce) typically run 60 to 120 days from announcement to enforcement.

Here is a week-by-week outbound plan.

Week 1 to 2: Build the list

• Pull all current Uber engineers by title at Senior and above
• Filter to US locations only
• Cross-reference profile location against Uber office locations
• Flag anyone on Core Services, Delivery Ops, or the merged commerce teams
• Deduplicate against your ATS and CRM

Week 3 to 6: First-touch outbound

• Lead with the specific team merge or geography, not "I saw the news"
• Skip severance and layoff language entirely, these people are still employed
• Offer a 20-minute conversation about "options that don't require moving to SF"
• Reference the actual anchor-day change: three days in-office, with in-person presence required on Tuesday, Wednesday, and Thursday

Week 7 to 9: Second-touch and referrals

• Ask each replier for one name on their team facing the same choice
• Prioritize the Engineering Manager cohort being demoted to IC under the 20% manager cut

Companies already showing up as current employers of former Uber engineers in Refolk's index, and therefore validated destinations you can name in outreach: Rippling, Figma, Datadog, Ashby, Omada Health, and Curai Health. That is a useful reference list when a candidate asks "who else are you placing people into."

What to actually put in the first message

Skip the "hope this finds you well" opener. Skip the "I saw the news about Uber." Every ex-Uber engineer will get that message 30 times this week from recruiters chasing the layoff list.

Write instead to the specific thing that is true about this person and not the layoff cohort:

• "You're in Denver, your team just moved to a Wednesday anchor day in SF."
• "You're on the merged Delivery Ops team, which now has two Staff engineers doing the same roadmap."
• "You went from EM to IC last week, and the IC ladder at [target company] pays better than Uber's L6."

The point of the message is to acknowledge that the person still has a job, is not on any layoff list, and has a decision to make in the next 60 days that Uber's HR team designed to be quiet. Making it not quiet is the whole play.

The sub-brand angle nobody is working

Uber Freight and Uber AI Solutions appear in Refolk's index as separate current employers, and their remote policies are not necessarily the corporate 1% cap. Sub-brand engineers show up in searches for "Uber" but sit under different org charts, different VPs, and often different HR policies.

Two implications:

1. If you are sourcing broadly across "Uber," you are catching sub-brand people who are not actually affected by the September 2 memo. Segment them out.
2. If you are sourcing narrowly for "Uber Freight" or "Uber AI Solutions" specifically, you may find a pool that is affected on a delayed timeline or a different policy. Worth a separate saved search.

This is the kind of surface area that opens up when you stop treating a company as one node and start treating it as an org chart.

FAQ

How is the relocation-refuser pool different from the 3,300 layoffs?

The 3,300 layoffs are named, dated, and already circulating on every recruiter list and layoffs tracker. The relocation-refuser pool is the 4,600 to 8,000 employees created silently by Uber's 1% remote work cap, who still show "@ Uber" on LinkedIn, do not have "Open to Work" turned on, and will quit over the next 60 days rather than move. Every one of them is a warm lead with no competing InMails.

Why should I trust the 4,600 to 8,000 number?

Uber has not disclosed the pre-policy remote share, so this is a derived range, not a company number. It assumes a corporate headcount of about 33,000 after the cut and a pre-policy remote workforce of 15% to 25%, which is the typical band for large tech employers going into a hard RTO. Even the low end of the range (4,600) is 1.4x the named layoff list. The strategic point holds regardless of where the true number lands inside the band.

What tools actually surface geography mismatches at scale?

LinkedIn Recruiter can filter by current employer and location, but it does not natively flag "this person's location is different from where their team anchors." That cross-reference is manual in traditional stacks. Refolk handles it because you describe the person in plain English ("Uber staff engineer in a metro without an Uber office") and the shape of the query includes the mismatch signal directly.

Is this pool really higher quality than the layoff list?

Yes, on average. Uber's cuts targeted management layers and the seventh-plus reporting layer, not skill obsolescence. Revenue is still growing at 12%. The engineers being forced out by the remote cap are the ones Uber wanted to keep but is willing to lose to save on severance. That is a growth-company attrition event, and the talent quality reflects it.