Layoffs to fund AI: the bad math hundreds of companies are making

Real competition is not on cost. It is on structure.

Many leadership teams are reaching the same conclusion right now: cut staff to fund AI, protect margins, stay in the race. That is understandable. It is also probably a mistake—not for ethical reasons, but for purely competitive ones.

---

Three companies, same sector

Take a concrete picture. Three insurers. Call them A, B, and C.

Company A is an incumbent—more than a thousand employees, processes honed over decades. Facing the AI wave, it acts: it adopts tools, cuts headcount in half, improves margins. On paper, that sounds reasonable.

Company B, of similar size, prefers to wait. It keeps teams as they are, bets on experience, watches. It believes it has time.

Company C does not have that kind of question—it was built around AI from day one. Ten people. An architecture designed so humans orchestrate automated systems, not the other way around. That is not a footnote: it is the whole difference.

---

What company A did not see coming

Company A cut, integrated tools, improved margins. But something did not change: its structure. It still has the same cascade of approval meetings, silos between teams, six layers of decision-making inherited from twenty years of traditional organisation. AI was grafted onto an organism that was not built for it.

The result: company C, with 10 people, can deliver value comparable to a team of 100—because its workflows were designed for AI, not patched in a hurry. Its burn? Around €700,000 a year. Its break-even? One million. Company A must generate €30 million just to cover payroll.

This is not a question of individual productivity. It is a question of organisational design. And on that front, company A starts with a structural handicap that layoffs did not remove.

---

The option company B did not take

If a 10-person startup can compete with 100, it means every well-equipped employee is worth ten times more than before—not by doing the same work faster, but by doing things no one could afford without AI. New offers, new markets, new capabilities.

Company B could have seized that. Train teams. Rethink processes end to end. Turn 1,000 employees into 1,000 augmented operators. It would then have had productive capacity that a halved company A simply cannot match.

That is exactly the bet IBM made—against the grain. In February 2026, while dozens of tech firms announced mass layoffs, IBM tripled junior hiring. Its CHRO, Nickle LaMoreaux, put it plainly: “The highest-performing companies three to five years from now will be those who bet on early-career hiring in this environment.”

The logic behind that choice deserves attention. IBM does not deny that AI can now do much of what juniors did three years ago. But it separates tasks from skills. AI can write code—it does not grasp why a client needs that code to behave differently than expected. It can answer an HR question—it does not notice that the question signals something deeper in the organisation. Those capabilities take years to build. They always start with a first job.

Cutting junior hiring today empties tomorrow’s pipeline of managers. In five years, firms that made that choice will have to recruit them elsewhere—at far higher cost than if they had grown them internally.

---

The numbers—and what they leave out

Q1 2026: more than 78,000 jobs cut in tech, nearly half of them tied directly to AI or automation. Block—Square, Cash App—went from 10,000 to under 6,000 in one move. CEO Jack Dorsey was unusually direct: this was not about financial distress; it reflected what AI can now cover.

Yet a Gartner prediction from the same year casts a shadow: they project that 50% of companies that cut customer-service headcount because of AI will rehire into similar roles by 2027—often under new titles (“trusted advisor,” “solution consultant”) but doing substantially the same work.

There is even a name for it in English-speaking HR circles: AI replacement regret. The mechanism is simple. An organisation cuts roles based on what AI promises to absorb. Months later, it realises automatable tasks were only part—often the visible part—of what those teams actually delivered. The rest—client relationships, contextual judgement, memory of complex cases—never shows up in a spreadsheet. Until it is missing.

---

What changes, in practice

AI benefits companies that use it to do things they could not afford before—not those that use it to do the same thing with fewer people.

That is a small distinction, but it changes everything. Those who cut without transforming will face competitors who did the opposite—and they will have done it while shedding the very teams that could have helped them change course.

---

Sources: TwinLadder Research (IBM case study, Feb. 2026), Fortune (IBM CHRO interview, Feb. 2026), Tom's Hardware / Nikkei Asia (Q1 2026 tech layoffs), Gartner (2026 AI workforce predictions)