Half the May 2026 Layoffs Are Coming Back Offshore. Don't Plan a Surplus.
May 2026's layoffs look like a US talent flood, but Forrester says half get rehired offshore. Here's how to adjust sourcing for a repricing event.
If you run sourcing, the May 2026 headlines read like Christmas. Meta cut 8,000 on May 20. PayPal 4,760. Cloudflare 1,100. Coinbase 700. Snap around 1,000. BILL up to 30%, Upwork around 25%, Microsoft offering buyouts to roughly 8,750 US workers. The temptation is to staff up your inbound funnel and wait for the resumes. Don't. The numbers underneath the press releases say this is a repricing event, not a surplus, and the talent you actually want to hire is not the talent landing on your careers page.
The Forrester number that breaks the surplus narrative
Forrester's Predictions 2026 report is blunt: roughly half of AI-attributed layoffs will be quietly rehired offshore or at significantly lower salaries. Invezz, citing Bloomberg coverage on May 4, framed it as "a labour repricing story, not purely a labour reduction story." Companies cut on the promise of AI capability, the AI underdelivers, and the seats refill in Bengaluru, Hyderabad, Ho Chi Minh City, or Kraków instead of Sunnyvale.
That reframes everything a US recruiter is supposed to do with a layoff list. The Cloudflare 1,100 are not all becoming your candidates. A meaningful share of those reqs are reopening, just not in a zip code your ATS understands.
The data that should have already changed your plan
Q1 2026 saw 81,747 tech layoffs per Challenger, Gray & Christmas. Tech YTD through April hit 85,411, the sector's highest since 2023. AI was cited as the #1 layoff reason in April, accounting for 26% of cuts. By every aggregate measure, 2026 should be a buyer's market for US engineering talent.
It isn't. Median Bay Area senior engineer time-to-hire stretched from 38 days in Q3 2025 to 67 days in Q1 2026. During the largest tech layoff quarter since 2023, hiring senior engineers in the densest engineering market in the world got 76% slower. If supply were actually flooding US pipelines, that line goes the other way.
Why the pools don't overlap
Layoff totals aggregate roles. The 8,000 at Meta on May 20 absorbed recruiting and HR at 35 to 40% of cuts. PayPal's 4,760 included support, operations, and legacy product. Oracle's 10,000 to 30,000 multi-wave restructuring is heavy on Cerner integration headcount and legacy DBAs. The seats opening on US req lists are something else entirely: AI/ML engineers, distributed systems infra, applied research, security, and the new "founding AI engineer" role that didn't exist 18 months ago.
Startup Fortune's reporting puts 275,000+ AI-related US job postings open even as cuts mount. Big Tech's combined $725B 2026 capex (up 77% YoY across Amazon, Microsoft, Alphabet, Meta) is partly financed by payroll cuts, but the specialized AI hiring inside those same firms hasn't paused. This is bifurcation, not contraction.
Which means the layoff list strategy, source ex-Coinbase engineers in the 72 hours after the announcement, only works if the people you actually need are on the list. Mostly, they aren't. The applied ML engineer Coinbase kept. The growth PM and the L4 backend they didn't.
The AI-washing problem
Marc Andreessen has called AI a "silver bullet excuse" for cost cutting. Wharton's Peter Cappelli says there's "very little evidence" AI eliminates jobs at the scale claimed. 9i Capital's Kevin Thompson reads this as a delayed correction of pandemic bloat. Whichever frame you prefer, the operational consequence for sourcing is the same: many of the laid-off workers are not obsolete, they're priced out. They reappear as contractors, at smaller competitors, or at offshore captives, not as flooding direct applicants to your $180k mid-level req.
Glassdoor tech-sector confidence dropped 6.8 points YoY, the largest drop of any industry, and tech wages outside specialized AI roles are flat vs 2025. The mood of the available pool is bad. Forrester projects 28% of workers will be disengaged in 2026, watching colleagues cut for AI that never shipped. The passive candidates you most want to poach are also the most cynical about employer pitches right now. Generic outreach is going to underperform Q2 like it has never underperformed before.
The offshore captive is the actual competitor
US recruiters keep modeling the post-layoff market as a competition between US employers for laid-off US workers. The honest model is a competition between the US req and the same employer's offshore captive.
India's GCC ecosystem now numbers 1,580 to 1,700 centers employing roughly 2 million professionals, with FY26 GCC revenue at $98.4 billion (Nasscom/Zinnov). North American firms drive two-thirds of new GCC setups. JPMorgan Chase, Nvidia, even McDonald's run the playbook. Bengaluru alone hosts 880+ GCCs. Hyderabad and Pune fill out the tier one. Ho Chi Minh City and Manila are the tier two plays. Kraków, Warsaw, and Bucharest serve EU-aligned work.
The cost gap explains the gravity. Senior full-stack developers in Vietnam and Malaysia at $3,000 to $5,000/month deliver comparable output to US peers earning three to four times that. Offshore developer rates run $18 to $65/hr vs $80 to $150+/hr in the US. Then layer on the new $100,000 H-1B fee for the 2026 cycle, which Forrester's Biswajeet Mahapatra explicitly cites as reinforcing offshore over onshore staffing. Your employer's CFO is doing this math whether your TA leadership wants to or not.
A laid-off generalist in Austin and a fresh GCC hire in Pune are both candidates for the same req. Pretending otherwise loses the req.
What changes in your sourcing plan
1. Stop pipelining off layoff lists alone
Layoffs.fyi and TrueUp (128,270 impacted in 2026 YTD) are useful signals, not sourcing strategies. Match by skill cluster, geography, and comp band, because the cuts are concentrated in roles that don't reopen at US rates. If your req is "Staff infra engineer, Bay Area, $320k base," the relevant signal is not who was on PayPal's list. It is who at PayPal worked on platform infra and is now bored at their landing spot. That's a search problem, not a list problem.
This is exactly the kind of search where natural-language sourcing earns its keep, and why we built Refolk around plain-English queries that compose skills, employer history, geography, and comp band into one ranked shortlist instead of forcing you to stitch a Boolean string across LinkedIn Recruiter, GitHub, and a spreadsheet.
2. Build for two pools, not one
A thin, expensive US AI/infra pool. A deep, cheaper offshore pool absorbing the rehired generalist work. If your hiring plan only sources one, you'll lose half your reqs to your own org's GCC and the other half to a 67-day time-to-hire. The senior IC pool isn't tiny in absolute terms. Refolk's index shows roughly 165,000 matching senior SWE and Staff Engineer profiles in the US, concentrated in SF, NYC, and Austin. The bottleneck is not headcount. It's matching specific skill profiles to specific reqs faster than the offshore alternative reprices the seat.
3. Treat passive outreach like the cynical pool it is
The "coaster tax" is real. 28% disengagement, flat wages outside AI, confidence collapsing. The reply-rate-killer in Q2 2026 is generic mission language and "exciting opportunity" framing. The senior engineers you want have watched two friends get cut while their CEO talked about "AI-first." They are not in a mood to read a LinkedIn InMail that opens with your funding round.
What works: specificity. Cite the candidate's actual project. Name the technical problem. Be honest about scope, comp band, and whether the role is being created because someone got cut. The more transparently you address the moment, the higher your reply rate.
4. Watch Q2 Challenger data and the next earnings cycle
The first real test of whether the AI-attribution language gets walked back under shareholder and regulator pressure comes with Meta, Microsoft, and Amazon Q2 earnings, plus the early-June Challenger report. If "AI" stops being a layoff justification and starts becoming a "we over-hired" admission, the market re-prices again. Pharma layoffs already jumped roughly 500% YTD with AI attribution. The pattern is bleeding out of tech. Recruiters in adjacent verticals shouldn't assume tech-only dynamics.
5. Source against GCC affiliation, not just company name
If you're a US employer who actually wants to compete with the offshore captive instead of losing reqs to it, you need visibility into who is currently inside a GCC and might consider a remote-friendly North American role. That's a public-data problem (LinkedIn employer history, GitHub commits to repos owned by the parent, conference talks). Most ATS-bolted sourcing tools can't see it cleanly. A natural-language query like "senior platform engineers in Hyderabad who joined a US-headquartered GCC in 2024 and have public OSS contributions to Kubernetes" is the right shape of question, and the right shape of answer is a ranked, contactable list, not a 1,000-result LinkedIn cap.
The honest take for talent leaders
The May 2026 wave is not a clean talent surplus. It is a labor repricing dressed in AI language, financed by $725B of capex, accelerated by a $100k H-1B fee, and absorbed by 1,580+ GCCs in India alone. Aggregate US tech layoffs went up. Aggregate US senior engineer time-to-hire also went up. That's not a contradiction. That's the bifurcation working.
If your post-layoff talent pipeline plan assumes the resumes are coming, replan now. The pool you actually want is smaller, more cynical, and harder to reach than it was 90 days ago, and your employer's CFO has a cheaper option in Bengaluru that doesn't need your help.
FAQ
Are the May 2026 layoffs going to flood the US senior engineer market?
No. Aggregate cuts are real (81,747 in Q1 2026 alone), but Forrester estimates roughly half of AI-attributed layoffs reopen offshore or at lower salaries, and the roles being cut (recruiting, support, generalist, legacy) are not the roles US req lists are trying to fill. Bay Area senior time-to-hire stretched from 38 to 67 days during the largest layoff quarter since 2023, which is the strongest signal that the surplus narrative is wrong.
Should I still source aggressively off layoff lists like layoffs.fyi?
Yes, but as one signal among many, not as your strategy. Match laid-off candidates against your req by skill cluster, geography, and comp band rather than by employer name. The high-value laid-off engineer is rarely the obvious one on the list, and the obvious ones are getting outreach from 40 other recruiters in the first 72 hours.
How does the new $100,000 H-1B fee change US tech sourcing?
It pushes employers toward offshore captives over onshore visa-dependent hires. Forrester's Biswajeet Mahapatra has called it out explicitly. Practically, that means US recruiters competing for laid-off mid-level talent are also competing against their own employer's GCC, where the same work gets done at one-third to one-quarter of the loaded cost. Build that into your req-prioritization and your TA pitch upward.
What's the right sourcing tool stack for a bifurcated market?
You need something that can search across LinkedIn, GitHub, and the open web in one query, handle natural-language inputs (skill cluster + employer history + geography + comp band), and surface candidates inside GCCs as cleanly as candidates in SF. That's the gap Refolk was built for. Boolean-on-LinkedIn-Recruiter alone is the wrong tool for a market split between a thin US AI pool and a deep offshore pool, especially with LinkedIn Recruiter's 1,000-result cap clipping the top of every search you'd actually want to run.