Monday.com cut 630 jobs on July 22, 2026, about 20 percent of the company, and took a restructuring charge of 45 million to 55 million dollars. TechCrunch had the numbers straight from the company that same day. Monday.com did not call it a cost cut. It called it a pivot toward what the company named an AI Work Platform, built around a no-code app builder, a customizable AI agent, workflow automation, and a chatbot that writes reports and updates dashboards. Same layoff. Better noun.
We have seen that noun everywhere this year. It is not a fringe explanation anymore. It is the default press release, and it deserves the same scrutiny we would give any other unverifiable claim a company makes about itself.
The running list
- Oracle: about 21,000 jobs cut over the twelve months through mid-2026, taking headcount from roughly 162,000 to 141,000. The company's regulatory filing cited adoption and deployment of AI technologies across its operations, reported by Forbes on June 23, 2026.
- Microsoft: about 4,800 jobs, roughly 2 percent of staff, cut in early July 2026 and concentrated in Xbox and commercial sales. The company told staff AI is changing how work gets done, per TechCrunch coverage that week.
- GitLab: 350 jobs, 14 percent of staff, cut on June 3, 2026, alongside an exit from 22 countries. CEO Bill Staples told TechCrunch that AI agents work at machine scale and are pushing competitors to the brink.
- Amazon: about 16,000 corporate jobs, 9 percent of corporate headcount, cut on January 28, 2026, with generative AI and AI agents named among the reasons, according to CNBC's report that day.
Read those four back to back and Monday.com stops looking like an outlier. It looks like the fifth entry on a list that already had a shape.
What the label buys you
Here is the incentive nobody puts in the press release. A company that says it overhired during the pandemic and is now cleaning house sounds like a management failure. A company that says it is restructuring around AI sounds like it saw the future early. Same spreadsheet. Different story for the earnings call, and a much better one for the stock.
It also happens to be unfalsifiable. No company publishes the headcount model behind an AI attribution: how many support tickets a chatbot now closes, how many engineers a code assistant replaced rather than assisted. We get an adjective and a charge amount. The math stays private.
A test you can actually run
Absent that math, three questions do the work instead.
- Does the announcement name a specific shipped product or workflow change, or does it stop at strategy language? Monday.com named its AI Work Platform down to the chatbot. GitLab named the exact scaling problem: AI agents querying its platform at a volume no human developer team ever produced.
- Does the cut land in a function a current model can plausibly do? Support tickets and first-draft code, yes. Closing enterprise contracts or running a game studio, not yet, no matter what the memo says.
- Does the company's capital spending or hiring elsewhere back up the story? A company pouring billions into the infrastructure that runs its AI while trimming the functions AI displaces is telling a coherent story. A company only cutting cost centers is not.
None of these require inside information. They require reading the release twice and checking the filing instead of stopping at the headline.
Applying it, case by case
Amazon passes the third test cleanly. Its January 28, 2026 cuts hit corporate overhead, and the company has continued to guide toward enormous capital spending on AI infrastructure through 2026. A company spending that much to build the thing it credits is not hiding the ball.
GitLab passes the first two tests. Its cuts landed inside engineering and infrastructure roles, the exact area its CEO described as overwhelmed by agentic traffic. That is not a company hiding a slowdown behind an AI story. That is a company with a capacity problem it can point to.
Microsoft is murkier. The July 2026 cuts concentrated in Xbox and commercial sales, and the stated reason, that AI is changing how work gets done, reads almost identical to Oracle's. But an AI agent does not run a game studio or close an enterprise software deal by itself yet. The label and the function do not obviously match, which is exactly the gap the test is built to catch.
Monday.com sits in between. The product description is specific enough to pass the first question. Which roles actually left the building is not public, so the second question cannot be answered from outside the company, and a 45 to 55 million dollar charge against a company that size fits either story equally well: a real platform pivot, or a standard cost cut that picked a good year to happen.
Where it is real, and where it is not
Some of this is genuine. Software companies with agent-heavy products and matching infrastructure spending, GitLab foremost among them, are describing something that actually happened inside the business. Some of it is a label. A company that will not say which functions it cut, will not connect the charge to a shipped product, and is not visibly spending on the technology it credits, is asking you to take the adjective on faith.
The test does not require faith. It requires the filing, the org chart if you can get one, and a calculator. Most companies would rather you stop at the headline.
Sourcing: TechCrunch on Monday.com, July 22, 2026; Forbes on Oracle's regulatory filing, June 23, 2026; TechCrunch on Microsoft and on GitLab; CNBC on Amazon's January 28, 2026 cuts. Corrections and questions go through our contact page.

