Meta Layoffs 2026: What the 8,000-Engineer Reset Means for Your Open Req

Meta Layoffs 2026: What the 8,000-Engineer Reset Means for Your Open Req

Meta begins the largest round of its 2026 workforce reset on May 20, cutting roughly 8,000 employees or 10% of global headcount across Reality Labs, Facebook, recruiting, sales, and non-AI product teams. Follows a January Reality Labs cut of 1,000 to 1,500 and a March reduction of about 700 across five divisions. Second half cuts are planned but not dated. Revenue is up, so this is a reallocation into AI, not a cost-cutting cycle.

Last updated: April 20, 2026

Mike Carter here. I run a desk inside KORE1’s IT staffing practice. A big share of our inbound this quarter has been ex-Meta. We have moved roughly thirty of these candidates since January, mostly out of Menlo Park, Seattle, New York, and Austin, with a handful from Bellevue and London. The profile on paper is impressive. The usable profile after one intake call is narrower than most hiring managers expect, and I want to explain why.

The national coverage on this story is framed almost entirely for Meta shareholders or for Meta employees. Nobody is writing it for the person staring at an open req with a tight hire date, a comp band that may or may not stretch to where the candidate wants it, and a hiring panel that has probably never interviewed someone from Menlo Park before. I will. Bias up front. We make money when you source ex-Meta engineers through our AI and ML staffing practice or our general IT desk. I will flag the places where that changes the take.

Skimming? Short version. Meta’s 2026 pool is the highest-comp, highest-prestige candidate stream the market has seen since the 2023 wave. Also the most internally reshuffled. Fewer than half the resumes landing in our inbox are from involuntary separations. The rest are people who read the internal tiering memo, saw which quartile they ended up in, and left before the package was handed to them. That context changes how you screen. In ways most intake calls do not cover.

What Meta Actually Cut in 2026, Month by Month

Three actions this year. A fourth pending. Each different in shape.

January opened with a roughly 10% reduction inside Reality Labs, somewhere between 1,000 and 1,500 employees depending on which tracker you trust, covered in detail by TheNextWeb, alongside the quiet shuttering of several Oculus Studios VR game teams that had been building for Quest through 2024. The Reality Labs budget was reportedly cut by 30%. That was itself a signal nobody in the press read properly at the time. Reality Labs had been the division everyone assumed Mark would protect through any downturn. Protecting it would have required a different capex plan. The AI capex plan won.

March 25 was the smaller and quieter event. Roughly 700 positions across at least five divisions, including additional Reality Labs cuts plus targeted trims in Facebook social and the data-center-adjacent teams, per CNBC’s reporting. This round had a specific pattern. Candidates were tagged before the cut. The internal conversation happened weeks ahead of the paperwork. Several of the ones we screened were already interviewing elsewhere by the time the severance envelope arrived.

May 20. The big one. Approximately 8,000 roles, about 10% of Meta’s global workforce, confirmed by Bloomberg and Fox Business. Affected teams span Reality Labs again, Facebook social, recruiting, sales, global operations, and large stretches of middle management. Engineering cuts skew toward product teams that are not adjacent to AI. Heavy impact on infra and platform engineers who sit inside ad-tech and consumer product rather than inside the new Superintelligence Labs organization, which is a distinction that matters when you are reading a resume and trying to figure out which candidates will clear an AI-adjacent bar at your company.

The second-half round has been signaled but not dated. The internal message, as relayed to us by a handful of candidates in second-round conversations at Meta, is that May is a first wave. Not a single event. The HR Digest puts it cleanly. People inside Meta expect a second hit by Q4. If your open req closes in August or September, the market will look different than the market you are screening today.

For broader context on the full year, our tech layoffs 2026 analysis has the Challenger Q1 aggregate data and how Meta fits against Oracle, Amazon, Dell, and the rest of the 2026 wave.

The Performance-Tier Story Nobody in the Press Is Naming

Here is the detail that does not appear in a single news story I read while prepping this piece. Meta redesigned its internal performance review system ahead of the 2026 reset. Four tiers. Top 20%. Middle 70%. Lower 7%. Bottom 3%. The bottom two tiers are the primary target of the performance-based portion of the cuts. Combined, they account for about one in ten employees. Same shape as the May 20 number. Not a coincidence.

Read that twice. Not every laid-off engineer is low-performing. The pool is pre-sorted. Different thing entirely.

Three things follow from this that matter if you are hiring.

First, the “bottom 3%” tag is almost never the reason someone you screen is looking. Those engineers took severance in Q4 2025 and Q1 2026. By the time May hits, the open-market pool is a blend of the “lower 7%” group, voluntary leavers who read the writing on the wall, and the pure Reality Labs separations where whole teams were eliminated regardless of tier. The Reality Labs pool is unmarked. The non-RL pool needs a second layer of screening.

Second, voluntary leavers are the most interesting bucket. A software engineer who saw the internal memo, read their own tier, and left for something better without waiting for the envelope is a fundamentally different candidate profile than someone who was involuntarily separated and blindsided by the paperwork. Our queue is skewing that direction roughly sixty-forty. Ask in the first screen. The answer tells you everything about pattern recognition, risk tolerance, and how they handle internal politics.

Third, Reality Labs candidates are the rarest and best part of this pool in 2026. Not close. These are people who built multi-year C++ systems work on Quest, Horizon, Codec Avatars, and the custom silicon stack, and who spent years in a performance culture that punished shallow demos and rewarded engineers who actually shipped things that worked at sixty frames per second on consumer hardware. Nothing they did was tied to a quarterly OKR for ad revenue. The skill set ports to robotics, automotive, defense, medical imaging, and any AR/VR or sensor-fusion shop that needs real low-level systems talent. Most hiring managers will not instinctively reach for an ex-Reality Labs engineer for a non-VR role. They should.

Where Displaced Meta Engineers Are Actually Landing

The destination map looks nothing like Oracle’s and nothing like Salesforce’s. Meta alumni have unusually strong external gravity. Peer hyperscalers treat them as first-round picks. AI labs recruit them directly. The founder path is active because Meta stock vested aggressively in 2023 and 2024, and plenty of senior engineers have the runway to spend a year angel investing or prototyping a stealth startup without financial stress. That matters.

Two separate maps. One for the consumer-apps engineer population (Facebook, Instagram, WhatsApp, Messenger, Ads). One for the Reality Labs population. I will keep them in one table. Side-by-side is how hiring managers think about it.

One more observation. The best Reality Labs candidates tend to self-select into robotics and defense before they ever reach our queue. If you are hiring for either vertical, move fast. Figure, Anduril, and Tesla’s humanoid group have been in-market aggressively since January, with aggressive sign-on bonuses, written offers on the same day as on-site interviews, and an unusual willingness to stretch levels for candidates with Quest or Codec Avatars experience on the resume. Our AI and ML engineering staffing practice is seeing the last wave of that pool close within two to three weeks of submittal. After May 20 that cadence tightens further.

The Comp Math Changed, Too

This is where hiring managers underestimate what they are up against. Badly.

Meta total compensation is the highest in tech outside of a small slice of elite AI labs. Levels.fyi puts the E5 median at roughly $494K, the E6 median above $690K, and E7 and E8 into seven figures before equity refresh. A mid-senior Meta engineer displaced in May is walking away from a $450K to $550K package. That expectation does not vanish at separation. It resets. Slowly.

The real landing number across our last thirty placements is between 75% and 88% of prior total comp, depending on the destination. Peer hyperscalers land closest to par. Mid-market SaaS lands around 80%. Founder-stage startup equity is a different currency entirely. Usually a real base cut in exchange for a real equity slug. A displaced E5 with three years of unvested stock will often accept that trade, because the mental math on “lottery ticket versus next refresh” runs differently after a layoff than before the layoff, and the downside of staying where you were no longer feels theoretical once the severance paperwork is on the table.

If you are trying to hire an ex-Meta E5 at a $220K base and you think the pedigree plus market pressure will close the candidate, you are going to lose to a $320K-base offer from a Series C that also writes meaningful equity into the package. Happens weekly on our desk. Three of our last five “closed by competing offer” notes in Q1 involved a Meta-alum candidate taking a startup seat over a better-named Fortune 500 offer because the startup paid base more seriously. You can still win this market. The price of admission is higher than it was in 2023. Go in knowing that.

What a lower comp band can still do is hire from the adjacent pools. A strong Instagram or WhatsApp app engineer looking to leave the Meta comp ladder voluntarily will sometimes take a meaningful step down for a product role that fits their values better, especially if the company is working on something they find personally interesting or if the geography matches a life decision they wanted to make anyway. Those candidates exist. They are rarer than the press narrative suggests. Screen for it explicitly. Do not pretend it is the default.

The Q2 and Q3 Playbook for Hiring Managers

Six moves. In order. We have been running this on intake calls since January. It works.

1. Decide early whether you need an AI-adjacent profile or a general systems engineer. The former is the most competitive segment of this entire pool and will quote the highest comp. The latter is far more interesting for most hiring teams and considerably cheaper.
2. Ask in the first screen what tier the candidate was in and why they left. The tier number itself is less important than how fluently they explain it. Hand-waving about “politics” or “it was a reorg thing” is usually the tell. Good candidates will talk about specific project re-scopings or product cancellations and trace the decision back to the Superintelligence Labs reorganization clearly.
3. Treat Reality Labs candidates separately. The skill set is different, the motivation for leaving is usually involuntary, and the destination preferences are different. Your head count plan probably has no Reality Labs slot. Check whether one of the slots you do have can be filled by a Quest platform engineer. Often the answer is yes and you miss it.
4. Lead with mission, not with salary. A Meta-alum who left voluntarily is almost always leaving partly for reasons that are not about comp. A real mission statement, a clear story about what they build, and a plausible path to impact will beat a 10% base bump from a less interesting company.
5. Move fast on the second round. Our average from intake to first interview across tech roles is 17 days, and Meta-profile candidates are running shorter than that right now because top candidates are closing inside three weeks of separation, which means a conventional two-week gap between rounds is the same as dropping out. Second-round interviews booked at the end of week two, not the beginning of week four. We lose placements every quarter to hiring teams that run a tidy but slow process.
6. Close on an offer that respects the prior package. Not at parity. Close to it on base, with real equity and a signing bonus that covers the unvested RSU clawback in a way the candidate can see on paper without needing to do the math themselves. Ex-Meta candidates watched a lot of RSUs evaporate this year. A signing bonus that acknowledges that loss carries more weight than a slightly higher base with no other movement. Every time.

The Broader Pattern Across 2026

Meta is not an isolated event. Five of the biggest companies in enterprise tech have now cut in the same direction inside the same six-month stretch. Oracle’s March 31 cut of up to 30,000 roles was the biggest by headcount. Salesforce’s February action was the smallest and the most surgical. Cisco’s 2024-to-2026 reshuffle was the longest in duration and the clearest example of net-positive headcount hiding behind a bad-looking headline.

Meta is the largest current-quarter event and the most reallocational. The entire 8,000 reduction will land during a year when Meta’s revenue is $201 billion, up 22 percent. Free cash flow is $43.6 billion. A company in that shape does not need to cut for cost. It is cutting to redirect payroll budget into an AI capex plan that has no precedent in the history of this industry. $115 billion to $135 billion in AI infrastructure spend during 2026 alone. A 27-billion-dollar joint venture with Nebius for a gigawatt-scale data-center campus in Louisiana. Chief AI Officer Alexandr Wang running Superintelligence Labs. All of it is being funded partly by the people who just got let go.

For a hiring manager, timing matters more than pattern. The Meta pool is open today. It thickens around May 20 as the main wave lands. It absorbs into AI labs, hyperscalers, and late-stage startups through June and July. By Labor Day the senior end is largely off the market. By Thanksgiving the mid-senior end is tight again. That leaves two usable quarters to make a Meta-profile hire, and the second is measurably harder than the first.

The KORE1 Take

We have placed into tech roles across 30-plus U.S. metros over 15+ years in business. The 92% one-year retention number on our direct hires is real. Audited. It holds up in 2026 across the IT desk the same way it held up in 2019. What changes year to year is which companies generate the displaced-senior profile worth hiring. 2026 is the Meta cycle.

For most hiring managers I have talked to this quarter, the honest recommendation is to widen your slate before May 20 rather than wait for the big round to land. The January and March pools are still fresh. Less competition on any given resume. By the time the May round hits, every recruiter in the Valley will be in your inbox talking about the same candidates. If your req is already open, we can start a submittal pipeline inside a week.

If you are on the other side of this and you were displaced in January, March, or soon in May, the same link applies. We work both sides of the desk. Candidates pay nothing. If you want a second read on the market before accepting the first offer, or you want a recruiter who will tell you plainly whether a $400K mid-market package is the real ceiling for your band this year, start a conversation with our recruiting team.

For a deeper look at how AI itself is changing the hiring process on both sides, our piece on how AI is changing tech recruiting in 2026 has the operational detail.

Before You Call Us

How many people is Meta actually laying off in 2026?

Roughly 9,700 to 10,200 positions across the full year so far, counting the January Reality Labs round of 1,000 to 1,500 employees, the March cut of about 700 positions across five divisions, and the May 20 round of approximately 8,000 roles that will touch Reality Labs again along with Facebook social, recruiting, sales, and global operations. Plus a second-half round that has been signaled in internal communications but not dated or scoped publicly.

Treat the press numbers as ceilings. Voluntary departures, internal transfers, and quiet severance-plus-sixty-day-job-search arrangements usually absorb ten to twenty percent of any announced figure before anyone hits the open market. The actual count of ex-Meta resumes landing in our inbox is lower than the headline, and the quality distribution is pre-sorted by Meta’s own tiering system.

Why is Meta laying off people when the company is profitable?

Because the cuts are a reallocation, not a cost reduction. Meta’s 2026 capex plan calls for $115 billion to $135 billion in AI infrastructure spend during a single calendar year, and the company is moving payroll budget from non-AI functions into the new Superintelligence Labs organization under Chief AI Officer Alexandr Wang so the new pods can staff up fast without blowing past the approved operating expense line.

Revenue is up 22 percent year over year at $201 billion. Free cash flow is $43.6 billion. Nothing about the financials requires layoffs. Everything about the chosen AI spend cadence does.

What is the performance-tier system people keep mentioning?

Meta redesigned its internal review structure ahead of the 2026 reset into four tiers: top 20%, middle 70%, lower 7%, and bottom 3%. The bottom two tiers combined roughly match the 10% workforce reduction planned across the year.

For a hiring manager, the takeaway is that the pool is pre-filtered. Candidates from the lower 7% are on the market for performance-adjacent reasons. Candidates from the middle 70% who left anyway are the voluntary leavers, and they are often the strongest hires in the pool because the decision was theirs. Reality Labs candidates are mostly involuntary and unmarked by tier, because whole teams were cut regardless of individual rating.

Should I target ex-Meta engineers for my open req?

If your comp band can credibly compete in the $300K to $500K total-comp range, yes, and this is a generational hiring window. Move before May 20 if you can. The January and March pools are still fresh and less contested than the May round will be.

If your comp band tops out below $250K, you can still hire from the adjacent pool but your targeting needs to be different. Focus on candidates leaving voluntarily for lifestyle, mission, or geographic reasons, and use equity plus signing bonus structures rather than base alone.

Where are the Reality Labs engineers going?

Mostly to robotics, automotive autonomy, defense tech, and AR/VR startups. Tesla, Figure, Anduril, Skydio, Waymo, and Apple have all hired aggressively out of this pool since January, usually with same-day offers and sign-on bonuses that acknowledge the unvested RSUs the candidate is walking away from at Meta. The skill set is C++ systems, perception and sensor fusion, real-time graphics, and custom silicon adjacency. It ports one-to-one to any industry that needs low-latency systems work at scale.

A small slice has gone to traditional enterprise SaaS for principal or staff engineer roles in platform infrastructure. That path is underused. If you are at a high-throughput enterprise company and need a principal-level systems engineer, a Reality Labs alum is probably the cleanest hire available in 2026.

Is Meta still hiring while all this is happening?

Yes, aggressively, inside Superintelligence Labs and the Applied AI organization. Traditional engineering roles across those teams are being rebranded to “AI builder,” “AI pod lead,” and “AI org lead.” Roughly 1,000 existing employees were moved into these positions before the public cuts, and Meta is actively hiring externally for the same titles.

What salary should I expect to pay a displaced senior Meta engineer?

For an E5 equivalent with five to eight years of Meta tenure, expect a $300K to $400K base and total comp landing between 75 and 88 percent of the candidate’s prior Meta total, including a signing bonus to cover unvested RSU clawback, which has become close to standard on credible competing offers this cycle. For E6 and E7 candidates, base moves to $350K to $475K and total comp crosses $700K.

The candidates who hold out for parity on equity as well as base tend to sit on the market two to three months. Ones who accept 82 to 88 percent of prior total comp in exchange for a better product mandate tend to close inside three weeks.

How fast can you actually fill a role with a Meta-profile candidate?

17 days from intake to first interview is our average across IT roles, and Meta-profile candidates are running shorter than that because inbound volume is elevated, the candidate pool is turning over quickly, and top candidates are closing within three weeks of separation. Best-case, for a well-scoped senior or principal role with a decisive hiring panel, intake to signed offer runs 28 to 35 days. The longest stretches in our recent closings have been where the client’s interview loop had five rounds and the candidate took another offer at round three.