On February 26, 2026, Jack Dorsey cut 4,000 employees from Block — roughly 40 percent of the company's 10,000-person workforce. The stated reason: "Intelligence tools have changed what it means to build and run a company." The stock surged 24 percent in a single session, its best day in four years.
Block was not a company in crisis. Its Q4 gross profit came in at $2.87 billion, up 24 percent year-over-year. Cash App generated $1.38 billion in gross profit alone. Revenue, margins, user growth — all trending in the right direction. Five months earlier, in September 2025, Block hosted a company-wide event that reportedly cost $68 million. The people who attended that event — who flew in, networked, celebrated the quarter — were now being shown the door.
The market's reaction was not subtle. Wall Street did not price this as a neutral restructuring. It priced it as a premium. Block's shareholders earned a 24 percent return in a single day because the company removed 4,000 people from its payroll and said the word "AI" while doing it.
Then Dorsey went further. He did not just explain his own decision. He prescribed it for the entire industry: "I think most companies are late. Within the next year, I believe the majority of companies will reach the same conclusion."
That is not a CEO managing his company. That is a CEO writing a playbook.
The 25 Percent Premium
When a company announces mass layoffs, the stock usually drops. Goldman Sachs research on restructuring announcements found that layoff stocks underperform their sector by roughly 2 percent in the weeks following the announcement — the market reads headcount cuts as a signal of distress, not strength.
But there is an exception. When the cuts are framed as AI-driven restructuring and paired with higher forward guidance, the penalty inverts. The stock goes up. Goldman's data shows this pattern clearly: the market treats "we are cutting costs because we are struggling" differently from "we are cutting costs because AI made these roles unnecessary and our margins will expand." Same action. Different narrative. Opposite market reaction.
Block executed the second version flawlessly. eBay did the same thing on the same day — cutting 800 roles with identical AI framing. The market rewarded both.
The pattern is not isolated. Challenger, Gray & Christmas tracked 108,435 job cuts in January 2026 alone — the highest monthly total since 2009, when the economy was in freefall. The firm counted approximately 55,000 AI-related layoffs across 2025, a twelve-fold increase from the prior year. In January and February 2026, the rate accelerated to roughly 850 tech workers per day.
"AI restructuring" has become a pricing signal. It tells the market that these costs are not coming back — that the cuts are permanent, structural, and efficiency-driven, not cyclical. The narrative transforms what would normally be a distress indicator into a margin expansion story. As I wrote in The AI Bubble Question, the gap between AI spending and AI returns is enormous. But the gap between AI narrative and AI reality is where the money moves.
The Confession Nobody Noticed
Buried inside Dorsey's announcement was a sentence that deserved more attention than it received. He admitted that Block had overhired during COVID — that the company had built two redundant organizational structures, one for Square and one for Cash App, with overlapping functions across engineering, product, and operations.
This is an honest admission. And it changes the story.
If Block genuinely had 10,000 employees doing the work that 6,000 could handle, then a significant portion of these 4,000 cuts are not about AI at all. They are about unwinding the hiring excess of the zero-interest-rate era, when capital was free and headcount was a vanity metric. Every major tech company did the same thing: hire aggressively from 2020 to 2022, discover in 2023 that interest rates had normalized and the growth assumptions had not, then cut.
The pattern is consistent across the industry. Meta cut 21,000 people in 2022-2023 and called it a "year of efficiency." Google cut 12,000. Amazon, 27,000. The stated reasons shifted over time — first it was "macroeconomic uncertainty," then "operational efficiency," and by 2025-2026, it was "AI." The structural cause remained the same: companies that hired for a world of zero percent interest rates adjusting to a world of five percent interest rates.
Sam Altman himself acknowledged this. In a January 2025 interview, he noted that companies are "AI washing" their layoffs — using the AI narrative to justify cuts that would have happened regardless of any technological shift. The technology provides the cover story. The balance sheet provides the motive.
The Permission Slip
Dorsey's most consequential statement was not about Block. It was about everyone else.
"I think most companies are late. Within the next year, I believe the majority of companies will reach the same conclusion."
Read that sentence carefully. It is not predictive. It is prescriptive. It is framed as an inevitability, which makes it function as an instruction. When the CEO of a $50 billion company tells the market that every other company should be doing what he just did, he is not making an observation. He is lowering the reputational cost of the action for every CEO who follows.
This is how norms shift. Not through regulation or gradual cultural evolution, but through a prominent actor publicly declaring that a previously risky decision is now safe. Every CEO who was considering AI-framed layoffs but hesitating over the PR risk just got permission from a peer. The herd has a leader now.
The problem is that the technology does not support the narrative at the scale being implied. A Harvard Business Review analysis published in late 2025 found that most companies laying off workers in the name of AI are doing so based on the technology's potential, not its demonstrated performance. Forrester's 2025 workforce survey found that nine out of ten companies do not have mature enough AI capabilities to meaningfully fill the roles they are cutting. The gap between "AI can theoretically do this" and "we have deployed AI that reliably does this" is wide — and the layoffs are happening on the theoretical side of that gap.
The junior developer hiring collapse was partly driven by the same logic: companies cutting roles that AI might replace, before AI had actually replaced them. The permission slip accelerates this pattern. It tells every CFO in every boardroom that the market will reward the cut now, regardless of whether the AI works later.
The Steelman — When the Cuts Make Sense
The strongest version of the counterargument deserves a fair hearing.
First: Block genuinely overhired. Dorsey admitted it. Cutting redundant organizational structures is operational discipline, not cruelty. A company that maintains two parallel engineering orgs for products that share infrastructure is wasting shareholder capital and employee potential. If you have 10,000 people doing 6,000 people's work, the honest thing is to say so and restructure.
Second: AI tools are changing what is possible with smaller teams. A well-equipped senior engineer with Copilot, Cursor, and Claude can produce output that previously required a team of three or four. The productivity gain is real — not speculative, not theoretical, but visible in shipping velocity and code output at companies that have invested in the tooling. The market rewarding efficiency is not a perversion. It is the market working.
Third: Dorsey's directness is preferable to the alternative. Most CEOs wrap layoffs in euphemism — "right-sizing," "organizational optimization," "strategic realignment." Dorsey said he overhired, said AI changed the equation, and said other companies should do the same. There is something to respect in a CEO who states his reasoning plainly rather than hiding behind consultants and press releases.
These points are real. And they are insufficient.
The efficiency argument works for one company in isolation. It breaks when every company executes the same playbook simultaneously. If Block cuts 4,000, and eBay cuts 800, and the next twenty companies follow Dorsey's advice — the aggregate effect is not "efficiency." It is a coordinated labor market shock that no individual company's AI deployment can absorb.
Forrester's 2025 data suggests this is already happening. Fifty-five percent of employers who made AI-attributed layoffs report regretting the decision. Forrester's workforce analysts predict that roughly half of all AI-attributed job cuts will be quietly reversed — workers rehired, often offshore, frequently at lower wages, sometimes as contractors rather than employees. The "AI efficiency" story, at aggregate scale, may turn out to be a wage arbitrage story wearing a technology costume.
The individual optimization is rational. The collective outcome is a market that rewards firing people first and asks whether the AI works later.
What the Playbook Actually Says
Step back from Block specifically and look at the pattern Dorsey just demonstrated. It is a three-step sequence, and it is replicable.
Step one: Frame the cuts as AI-driven. This triggers the Goldman Sachs premium instead of the default penalty. The market reads "AI restructuring" as permanent cost reduction, not cyclical distress. Same layoffs, different label, opposite stock reaction.
Step two: Pair with higher forward guidance. Block's announcement came alongside strong Q4 numbers and a margin expansion narrative. The cuts are not positioned as a response to weakness. They are positioned as a bet on strength — the company will do more with less, and the savings drop straight to the bottom line.
Step three: Publicly encourage peers to follow. This is the move that transforms an individual corporate decision into an industry norm. When Dorsey tells every CEO to do what he did, he normalizes the strategy, reduces the reputational risk for followers, and creates a self-reinforcing cycle: each company that follows validates the decision for the next.
Dorsey executed all three in a single announcement. The market response confirmed that the playbook works. As I discussed in The New Gilded Age, the structural incentive in the current economy consistently rewards capital concentration over labor investment. The AI layoff playbook is the latest and most efficient expression of that incentive: fire people, call it technology, raise guidance, and watch the stock price prove you right.
The optimal CEO behavior is now legible. Cut headcount. Attribute it to AI. Issue higher guidance. Tell your peers to follow. Stock goes up. Repeat.
The Question the Market Cannot Price
Return to the $68 million September event. Block flew thousands of employees to a company celebration. They met their colleagues. They built relationships across teams. They absorbed the institutional knowledge that only transfers in hallways and after-hours conversations — the understanding of why a system was built a particular way, which edge cases matter, what the last outage taught the team.
Five months later, those people were gone.
The market priced the efficiency gain with precision: 24 percent, in a single day. What the market cannot price is what left with those 4,000 people. Institutional knowledge does not appear on a balance sheet. Team cohesion does not show up in a quarterly earnings call. The cultural signal — that the company will celebrate you in September and cut you in February — does not have a line item.
Forrester's prediction that half of AI-attributed layoffs will be quietly reversed suggests the market already suspects the efficiency is overstated. But the reversals will happen quietly, over years, in offshore offices and contract arrangements that do not trigger the same headlines. The 24 percent stock premium is immediate and visible. The rehiring is gradual and invisible. The incentive structure is clear about which one matters more.
When the market rewards you for firing people, and the most prominent CEO in the room tells every other CEO to follow — the question is not whether the AI works. The question is what the market is optimizing for.
Dorsey got a standing ovation from Wall Street. The 4,000 who left got a severance package and a LinkedIn notification.