Uber's WARN Names 483. The 1% Remote Rule Names 2,817 More.

If you are sourcing off the Sep 17 California WARN filing for Uber, you are looking at 390 names and calling it a layoff list. It is 14.6% of a 3,300-person cut, and the other piece of the Sep 2 memo, the one that limits fully remote work to roughly 1% of corporate staff, is doing more sourcing work for you than any WARN scraper ever will.

What Uber actually announced on Sep 2, 2026

Uber announced 3,300 layoffs (about 10% of global staff) plus a return-to-hub policy that caps fully remote employees at around 1% and concentrates surviving teams in San Francisco and New York. CEO Dara Khosrowshahi's memo framed the cut as an org-design problem (layers, coordination, fragmented ownership), not an AI substitution story, which matters for how you pitch the survivors.

The concrete pieces:

• 3,300 cuts globally, roughly 10% of the ~34,000 headcount Uber reported at end of 2025 in its 10-K.
• ~1% fully remote cap, with everyone else expected in an approved hub.
• 20% cut to managers, with "almost half" of micro-teams (1 to 2 direct reports) dissolved and some EMs converted to individual contributors.
• No AI justification in the memo, unlike Meta's 8,000-person cut earlier in the cycle.

The AI framing matters because it tells you what the survivors think about themselves. Meta's cuts made surviving researchers feel replaceable. Uber's make surviving engineers feel reorged. Different pitch, different outreach, different close.

Why WARN only sees 483 of the 3,300

WARN captures roughly 14.6% of Uber's September cut because the federal Worker Adjustment and Retraining Notification Act only triggers on US layoffs above specific site thresholds, and most of Uber's cuts fall outside those triggers. The named US filings so far:

The California filings break down by address: 151 at 1655 3rd St. SF, 102 at 1725 3rd St. SF, 94 at 200 Mathilda in Sunnyvale, and 43 at 190 Mathilda. All effective Nov 2, 2026. The Seattle notice, signed by VP HR Julie Viray, covers 93 permanent terminations at 1191 Second Avenue, same effective date.

Both California sites are Uber engineering centers, per finalroundai.com's teardown. If you assumed these were "corporate/ops" cuts, you are targeting the wrong resumes. The named 390 skew IC engineer and engineering manager, which is the same cohort you are competing for at DoorDash and Rippling right now.

14.6%

The 2,817 you cannot scrape

The residual sits in three buckets. First, sub-threshold US sites (Chicago, Dallas, Miami, Raleigh) where fewer than 50 people got cut at any single address. Second, non-WARN states with different rules or no rules. Third, and largest, the non-US footprint: Amsterdam, Mumbai, Bengaluru, Gurugram, São Paulo. None of those geographies have a WARN equivalent, so the residual cuts are effectively unsourceable through public filings.

This is where the sourcing problem changes shape. You cannot chase a list that does not exist. You have to reverse-engineer the population it was drawn from.

The 1% remote rule is the actual sourcing signal

The return-to-hub mandate is a leading indicator that publicly labels every Uber employee outside SF and NY as forced to relocate, commute farther, or quit, which is a larger and higher-quality pool than the WARN list will ever produce. WARN is a lagging, incomplete signal. RTO is a live one, self-updating every time an Uber employee thinks about their kids' school district.

The mechanism is simple. Uber's non-SF/NY US footprint (Seattle, Chicago, Dallas, Miami, Raleigh, Austin) contains thousands of employees who accepted their offers under a remote or hybrid understanding. The Sep 2 memo revoked that understanding for 99% of them. Some will relocate. Most will not. The ones who will not are not on any layoff list. They are still employed, still credentialed, and now actively open to a first message.

That is the poach cohort. Not the WARN 483. The RTO-refuseniks.

WARN is a list of the fired. The RTO memo is a list of the pushed. The second list is bigger, fresher, and still employed.

How to build the filter

Concretely, the sourcing spec looks like this:

1. Current employer contains Uber, joined before Jan 2024 (post-pandemic remote-friendly hires are the most exposed).
2. Location is not San Francisco, not New York, not Sunnyvale.
3. Role family matches your open req (backend, ML, data, PM).
4. Bonus signal: a manager title with a small team, given the 20% manager cut and the micro-team dissolution.

Every one of those filters is a plain-English clause, which is the exact gap Refolk closes. You describe the person you want in a sentence and get a ranked shortlist back across GitHub, LinkedIn, and the open web, rather than stitching four Boolean strings together and hoping the recruiter seat license covers the geography.

Where the 2,817 actually sit

In Refolk's index, roughly 26,038 people currently list Uber across India, the Netherlands, and Brazil combined, concentrated in Amsterdam, Mumbai, Bengaluru, and Gurugram. That population is where the non-US portion of the 2,817 unnamed cuts is landing, and it is also the population most exposed to any future hub consolidation.

26,038

A few practical observations from the same index:

• Alumni destinations for ex-Uber ICs are led by Meta (13), Google (3), Rippling (2), DoorDash, Snowflake, Salesforce, and LinkedIn. DoorDash is the obvious rival poacher; if you are recruiting for a marketplace or logistics team, you are competing with a company that already knows how to close this profile.
• US-based Uber engineering/PM/DS profiles matching a specific sample query returned 128 in the index, which is small because the query was keyword-filtered. The point is precision, not volume; if you need 8 senior Go engineers who worked on Rider growth, you do not want 5,000 profiles.
• Manager-to-IC demotions never show up in any layoff dataset. These are the highest-quality poaches in the entire cohort because they kept their salary but lost title and scope, which is the exact resentment pattern that closes an offer.

The international arbitrage window

Nov 2 is the US effective date. International severance and consultation timelines (particularly in the Netherlands) lag by weeks or months. That gives non-US recruiters a roughly 60-day window where EU and India-based Uber employees are actively interviewing while US-focused sourcers are still scraping California WARN. Amsterdam is the single most valuable geography here: high engineer density, English-language interviews, EU work authorization already sorted.

Comparable events, and why Uber is different

Uber is not Meta and it is not Oracle, which changes both your outreach and your close rate. Meta's 8,000 cuts were framed as AI substitution, meaning survivors feel their skills are obsolete. Oracle's 30,000-role restructuring against a $2.8B budget was a cost play, meaning survivors feel like line items. Uber's cut is an org-design story, meaning survivors feel reorged, not replaced.

That matters because the pitch to a surviving Uber engineer is not "we still value humans" (Meta) or "we are growing" (Oracle). It is "you can stop reporting through three layers of new leadership and just build the thing." That message closes ex-Uber ICs faster than a compensation delta will, because the compensation deltas in this market are compressed.

For volume context, Layoffs.fyi has logged more than 128,000 tech layoffs across ~290 companies in 2026, and Silicon Valley alone laid off 7,295 people in the first half. Bay Area tech cuts are already 36.6% above all of 2025. The market for laid-off Uber talent is saturated. The market for RTO-refusenik Uber talent (still employed, still credentialed, quietly interviewing) is not.

A 30-day plan for sourcing the 2,817

Here is what a working recruiter should actually do between now and Nov 2, in order:

1. Skip the WARN scrape. 390 California names, 93 Seattle names. Everyone on your competitor list already has them. Diminishing returns start on day one.
2. Build one saved search for RTO exposure: current Uber employees, hired 2020 to 2023, based outside SF/NY/Sunnyvale. This is the pool.
3. Layer role filters (backend, ML, marketplace PM, ops research). One sentence beats four Boolean strings; iterate faster.
4. Prioritize former managers now working as ICs. Look at LinkedIn tenure gaps in title, not headcount.
5. Open a parallel Amsterdam and Bengaluru track for the 60-day international arbitrage window before local severance closes.
6. Message with the org-design pitch, not the AI pitch. "Fewer layers, direct ownership" beats "we still value humans" for this cohort.

If you are running this on a team, the useful move is to split the WARN 483 to a junior sourcer (fast, low differentiation) and the RTO cohort to your best closer. The RTO pool is where the offers accepted this quarter will actually come from.

FAQ

How many of Uber's 3,300 September 2026 layoffs are named in public filings?

Roughly 483 across all US WARN filings so far: 390 in California (253 SF plus 137 Sunnyvale, effective Nov 2, 2026) and 93 in Washington State at 1191 Second Avenue in Seattle. That is 14.6% of the announced total. The remaining ~2,817 sit in sub-threshold US sites, non-WARN states, or the non-US footprint (primarily India, Netherlands, and Brazil), and will not appear in any public filing.

What does Uber's "1% remote" rule mean for sourcing?

It means every current Uber employee based outside San Francisco, New York, and Sunnyvale is effectively labeled as forced to relocate, commute, or quit, which turns the return-to-hub policy into a live flight-risk filter. This pool is far larger than the WARN 483, still employed and credentialed, and self-updating every time an affected employee weighs relocation. Filter by current employer Uber, hire date 2020 to 2023, and location outside the approved hubs.

Why are ex-Uber engineering managers a better poach target than the WARN list?

Because Uber cut managers by 20% and converted some surviving EMs to individual contributor roles, meaning a whole cohort kept salary but lost title and scope, and none of them appear in any layoff dataset. They are still on payroll, still credentialed, and quietly resentful, which is the exact profile that closes on an offer with more direct ownership. Look for LinkedIn profiles where the title changed but the company and tenure did not.

Which non-US geographies matter most for Uber sourcing right now?

Amsterdam, Mumbai, Bengaluru, and Gurugram carry the densest ex-US Uber engineering population, and none of those countries have a WARN equivalent, so the residual cuts there are only discoverable through professional profile indexes rather than public filings. International severance timelines lag the Nov 2 US date by weeks, giving non-US recruiters a roughly 60-day arbitrage window before local competitors start their outreach in earnest.