H-1B Sponsors With Layoffs: Introducing the WARN Cross-Reference

Which H-1B sponsors have announced layoffs? Until now, that question meant opening two unrelated stacks of paperwork. This site now cross-references state WARN Act layoff disclosures with the same employers' LCA (Labor Condition Application) filings — the paperwork that underpins H-1B sponsorship. Two public records, placed side by side. This is the launch of our new layoffs section, and it is Part 1 of a five-part "Layoffs × H-1B" research series.

Before anything else, one hard rule that governs the entire project: placing these two records next to each other implies no cause and effect in either direction. A company that files WARN notices and also sponsors H-1B workers is doing two legal, routine, and completely separate things. A WARN notice never identifies which employees are affected, so nothing here says — or could say — that H-1B workers specifically were laid off.

What a WARN notice actually is

The Worker Adjustment and Retraining Notification (WARN) Act requires larger employers — generally those with 100 or more employees — to give 60 days' advance notice of a mass layoff or plant closure. That notice goes to the state's workforce agency, which publishes it.

Three things follow from that definition, and each one matters when you read the data:

• A notice is an announcement, not a completed event. It is filed ahead of the action. Some notices are later reduced, rescinded, or overtaken by events. Treat a WARN filing as a plan on the record, not a body count.
• Headcounts are employer estimates at filing time. The number of workers listed is what the employer projected when it filed. It is not an audited final figure.
• It says nothing about visa status. WARN reporting is about the size and timing of a workforce action. It does not break workers down by immigration status, job title, or team.

For the formal definitions we use throughout the site, see the glossary, and for how these records are joined, the methodology page.

What we cross-referenced, and how carefully

As of the August 2026 data snapshot, we track 6 state sources — California, Texas, New Jersey, Illinois, Massachusetts, and Virginia — holding 7,656 WARN notices in total. Of those, 1,038 notices were confidently linked to 560 H-1B sponsors, together naming roughly 107,350 workers across all the years and states covered.

The word confidently is doing real work there. Employer names are messy: subsidiaries, abbreviations, "Inc." versus "LLC," and near-duplicates make automated matching error-prone. Our standard is deliberately conservative — only exact-verified or human-reviewed matches are ever published. When a name is ambiguous, it goes to manual review rather than being guessed. That means the linked set is an undercount by design. We would rather miss a real match than publish a wrong one.

The employers doing the most H-1B sponsoring are a useful backdrop for reading these matches, since the largest sponsors are also among the largest employers overall:

Coverage windows differ by state — read them first

This is the single most important caveat for anyone browsing the layoffs section. Each state began publishing (or archiving) WARN data at a different time, so the historical depth is wildly uneven:

Why this matters: California, one of the largest H-1B states, appears here only through a rolling current-year window at snapshot time. A California employer with a big layoff in, say, 2022 simply would not show up — not because it didn't happen, but because that history isn't in the feed we can currently access. Comparing raw notice counts across states without adjusting for these windows will mislead you every time.

And the bigger point: the absence of an employer says nothing. We track 6 states. Forty-four states are not tracked at all. A sponsor with no WARN notice in our data may have filed one in a state we don't cover, may be below the WARN size threshold, or may simply have not had a qualifying event. "Not listed" is not a clean bill of health, and it is not evidence of anything.

Reading a sponsor page without jumping to conclusions

When you land on an employer that has both LCA filings and linked WARN notices, you are looking at two timelines that happen to belong to the same company. Useful ways to read them:

• Scale check. A firm with tens of thousands of employees will naturally generate more of both filings. Volume alone is not a story.
• Timing is context, not causation. A notice dated near a hiring push tells you the company was doing both at once — which is common at large, multi-division employers — not that one caused the other.
• Mind the units. LCA figures describe offered wages and positions requested; they are not headcount, not take-home pay, and not visas granted. WARN figures are estimated affected workers. They don't add up to a single number, and they shouldn't be treated as if they do.

A few recurring guardrails carry over from the rest of the site and are worth repeating because they are so easy to trip over. An LCA is not a visa — it's a prerequisite step. A certification rate is not a USCIS approval rate. Withdrawn is not denied. An offered wage is not take-home pay. And FY2026 LCA data is partial — roughly two quarters reported — so any current-year totals are incomplete. Our full list of caveats lives on the disclaimer page.

What's coming in the series

This launch piece is the map; the next four parts explore the terrain:

• Part 2 — Pharma's dominance. Which industry shows up most in the linked notices, and why the health-and-science side of the data behaves differently from tech.
• Part 3 — The biggest sponsors' notices. What the WARN records of the largest H-1B filers actually look like when you line them up.
• Part 4 — The 2020 vs. 2023–2026 pattern. How the shape of layoff announcements changed between the pandemic shock and the more recent stretch.
• Part 5 — What this data cannot tell you. A full accounting of the limits, so no one over-reads a chart.

The bottom line

We built the WARN cross-reference because two important public records had never been readable in one place, and readers deserve to see them together. What we did not build is a layoff-causation engine. Every design choice — the conservative name matching, the published coverage windows, the refusal to guess — points the same way: show the records, state the limits, and let readers draw careful conclusions. Start at the layoffs hub, and read each employer page with the coverage windows firmly in mind.