The next chapter of the AI layoff story is not a bigger round of cuts. It is companies quietly rebuilding the teams they just eliminated, and discovering that reversing a workforce decision costs more than making it did.
The cutting has not slowed down
Cloudflare gave the clearest read yet on how far companies are willing to go. In its first quarter 2026 results, the company said it would cut about 1,100 people, roughly 20 percent of its workforce, and take $140 million to $150 million in restructuring charges as it moves to what CEO Matthew Prince called “an agentic AI first operating model.” Prince told investors the company does not just build and sell AI tools, it is “our own most demanding customer,” and that AI agents now handle work that used to require headcount across HR, marketing, finance and engineering. The cuts came alongside strong results, with revenue up 34 percent year over year to $639.8 million, underscoring that this was a strategic bet on AI rather than a response to financial distress.
Cloudflare is not alone. Trade coverage compiled from workforce-tracking services this month counted Oracle cutting 25,254 roles even as its net income surged 95 percent, Amazon eliminating more than 17,000 jobs while committing up to $200 billion to AI infrastructure, and Block cutting 4,000 roles after leadership said AI could handle “a meaningful share” of the affected work. The pattern is consistent: profitable companies, redirecting savings into AI capacity, framing the reduction as a capability shift rather than a cost cut.
What the win looks like from the inside
Some employers are choosing the opposite path and getting a different outcome. Ingka Group, the largest IKEA franchisee, reskilled roughly 8,500 call center agents into remote interior design consultants after its Billie chatbot began resolving 47 percent of inbound customer inquiries on its own. Rather than eliminate the roles the chatbot displaced, the company retrained the people into a function AI could not do, and says the reskilled team now generates about 1.3 billion euros, or roughly 3.3 percent of revenue, with a target of pushing that figure to 10 percent. It is the counterfactual to the Cloudflare model: same triggering event, a chatbot absorbing routine work, a completely different HR response.
The correction is already showing up in the data
At Ford, the fix looked like a rehire. The automaker brought back roughly 350 veteran engineers over three years after its automated inspection systems missed design and quality defects the more experienced staff would have caught. Charles Poon, Ford’s vice president of vehicle hardware engineering, told Bloomberg that the company had let go of some of its most experienced people before their knowledge could be used to train the AI systems meant to replace them, a sequencing problem more than a technology one.
Ford is not an outlier. A survey of hiring managers by talent firm Robert Half found roughly three in ten employers had eliminated a position after adopting AI, then rehired for it later, with the rate reaching 35 percent in HR functions and 44 percent in finance, against an all industry average near 32 percent. Separate workforce data from Visier, drawn from 2.4 million employees across 142 companies, put the broader boomerang rehire rate at 5.3 percent, but found those rehires cost about 5 percent more on average than employees who had simply stayed, versus a 2 percent premium industry wide. In finance alone, the premium on boomerang hires was estimated at close to $19 million in a single year.
The people making the cuts are starting to say so publicly. Daniela Seabrook, chief human resources officer of Adecco Group, said the biggest risk facing organizations today “is not moving too slowly on AI, it is moving too quickly without a people strategy in place.” Peter D. Banko, president and CEO of Baystate Health and author of a book on organizational transitions, put it more bluntly: cutting roles is treated as an “easy button” for cost control, when it is, in his words, “the most important decision a leader will make.”
What this means for the HR leader
Three practical shifts follow from the data. First, model the knowledge transfer cost before the cut, not after the rehire. Ford’s problem was not that AI failed, it was that the people who understood the AI’s blind spots left before anyone captured what they knew. Second, treat a rehire rate as a KPI, not a footnote. If boomerang hiring costs five percent more than retention, a pattern of AI driven cuts followed by rehires is a budget line HR can quantify and take into the next restructuring conversation. Third, watch the legal exposure stacking up around the automated side of these decisions. HR Tech has covered how AI driven layoff scoring is now the subject of a federal discrimination lawsuit, which raises the bar for any employer treating an algorithm’s output as sufficient justification for a headcount decision on its own.
None of this means AI restructuring is a mistake. Cloudflare’s bet may well pay off exactly as planned. But the emerging data from Ford, Robert Half and Visier suggests that a meaningful share of this year’s AI headcount decisions were made faster than the organizations could actually absorb, and the bill for that speed is now landing on HR’s desk in the form of rehiring costs, knowledge gaps and, increasingly, litigation.
Source: Cloudflare
