Restructurings Fell by More Than Half. The Share That Mention AI Quadrupled.

When Angi cut about 350 jobs on January 7, 2026, its filing gave a modern reason. The cut was “in light of AI-driven efficiency improvements.” Sentences like that now turn up every week, and they feed a fear every leader has heard in a board meeting: AI has started taking jobs, at scale.

We wanted to know what the filings say, not the headlines. So we went through every restructuring filing US companies made with the SEC from January 2023 to September 2026. That is every Form 8-K reporting Item 2.05, the item a company files when it commits to exit costs such as severance. 15 quarters, every filer, every word of the filing and its press release.

We also checked the other side of the ledger. Filings only cover public companies, so we set them against the Bureau of Labor Statistics' monthly count of layoffs across every employer, and against the cost of money, which drove the last big wave of cuts. If AI were reshaping employment at scale, at least one of those three records should show it.

The answer is two numbers moving in opposite directions, and the gap between them is the story.

Restructuring filings are getting rarer. From January to September, companies filed 344 of them in 2023 and 153 in 2026: fewer than half. Among large public companies, the ones our Decision Ledger tracks, the count for those months was 124 in 2023 and 81 in 2026, back to the 85 of 2019.

At the same time, AI moved into the filings. In 2023, 7% of restructuring filings mentioned AI anywhere. In 2026, 29% do, about 3 in 10. The share has quadrupled while the number of filings fell by more than half.

A mention is not a reason. So we read each of the 45 filings this year that mention AI and sorted them by what they say. 16 tie the cut to AI or automation doing the work. 10 say the cut pays for AI, or reorganizes the company around it. 17 mention AI only in passing: in product pitches, company descriptions and boilerplate. 2 say outright that the cut was not about AI.

The first kind sounds like Snap, which said “rapid advancements in artificial intelligence enable our teams to reduce repetitive work.” The second sounds like Cisco, which cut “in order to allow it to invest in key growth opportunities including silicon, optics, security and artificial intelligence (AI),” or Pinterest, “reallocating resources to AI-focused roles and teams.” That is AI as the destination of the savings, not the thing that did the work.

And some companies went out of their way to say no. Autodesk wrote that its changes were not “an effort to replace people with AI.” Etsy: “these decisions weren't driven by AI.”

The second kind deserves a closer look, because it is the one most often misread. A cut that pays for AI is a capital decision. Money moves from salaries to computing power, new products and a smaller number of new hires with different skills. It tells you where the company wants to spend. It does not tell you that AI did the work of the people who left. Pinterest said as much in its own filing: it is reducing staff in the near term and plans to reinvest in key development areas. In a headline, both kinds read as “AI layoffs.” In a budget, they are opposites.

Put it together: 16 of 153 restructuring filings this year say AI is doing the work. About 1 in 10. That is real, and it is growing. It is not a wave.

Filings only cover public companies, and only cuts big enough to report. The Bureau of Labor Statistics counts layoffs and discharges across the whole economy every month. From January to August 2026, the layoff rate averaged 1.08% of jobs a month, against 1.19% for the same months of 2019. This year's rate is 91% of the pre-pandemic level.

There is one exception, and it matters. In the information sector, which includes software publishers, internet companies and media, the layoff rate averaged 1.74% this year against 1.38% in 2019: 26% higher. That is also where most of the AI-mentioning filers come from. In professional and business services, another sector often called exposed, the rate was 2.03% against 1.94%.

If AI is not driving the count, what is? The filings cannot prove a cause, and neither can we. But the timing is worth seeing. The federal funds rate went from 0.08% in January 2022 to 4.65% by March 2023, the fastest rise in decades. The burst of restructuring filings came right behind it: 161 in the first quarter of 2023 alone. As rates came down from their 5.33% peak to 3.75% by September 2026, the filings fell too.

That does not make rates the only cause. Companies that hired fast when money was free cut when it stopped being free, and slowing growth does the rest. What the timing shows is that the restructuring cycle of the past three years has an older, duller explanation available, and it fits the numbers better than AI does.

Three objections deserve an answer.

First, a filing is not a layoff. Many cuts are never filed, and one filing can cover a few dozen jobs or many thousands. True. That is why this call rests on a second, independent count: the BLS layoff rate covers every employer, filed or not, and it is below 2019.

Second, the information sector is different. Its layoff rate is 26% above 2019, and the companies saying AI does the work are mostly software and internet firms. Also true. Inside that sector, AI may already be part of the story. This call is about the economy, and about what leaders outside that sector are told.

Third, sorting what a filing says is a judgment. Someone else might take Coinbase's “AI era” or Groupon's “AI-native” differently. That is why every one of the 45 filings, with how we sorted it and a link, is published with this call. Move any of them from one kind to another and the conclusion holds: even counting every filing that gives AI as a reason of either kind, AI is a reason in 26 of 153.