Block just mass-fired 4,000 people and called it progress
Jack Dorsey's fintech company is thriving, and just cut nearly half its workforce because AI tools made them expendable
The letter
Jack Dorsey just told more than 4,000 people they no longer had jobs.
The letter was oddly personal for something so brutal. No corporate euphemisms about “right-sizing” or “strategic realignment.” Dorsey said Block would shrink from over 10,000 employees to fewer than 6,000, and he said why: “intelligence tools paired with smaller, flatter teams” could do what those 4,000 people used to do.
He chose one massive cut over a series of smaller ones. Repeated rounds, he wrote, “destroy morale and trust.” Better to rip the whole thing off at once. Sigh.
The severance was generous. Twenty weeks of base pay plus an extra week for every year of tenure. Equity vesting through May. Six months of healthcare. A $5,000 transition stipend. Dorsey even kept Slack channels open temporarily so people could say goodbye to colleagues, a choice he described as “awkward and human rather than efficient and cold.”
“Awkward and human rather than efficient and cold.” That’s Dorsey’s own phrasing, in a letter about replacing humans with machines. Let it sit for a second.
It was the most carefully worded mass firing in recent tech history. And Wall Street, as any faceless institution would, loved it.
Twenty percent, overnight
Block’s stock jumped roughly 20% in after-hours trading. The math was clean: $450 to $500 million in restructuring charges, paid back within two quarters by the savings. The company raised its 2026 gross profit guidance to $12.2 billion.
Run the arithmetic on that stock surge and you’ll find yourself squinting. Divide the increase in enterprise value by the number of eliminated roles and you land somewhere around $1.5 million per person let go.
That’s the market’s verdict, expressed in dollars. Four thousand people updating their LinkedIn profiles. Investors adding five figures to their portfolios. Both things happened on the same day.
The market created roughly $1.5 million in enterprise value for every person who lost their job. If you want to know whose interests the system serves, there’s your answer.
Nobody on Wall Street paused to feel weird about this. The quarterly earnings call treated the layoffs the way you’d treat upgrading a server rack. Old capacity out, new efficiency in. Numbers work. Move on.
This time is different (and not in the way VCs usually mean)
The 2022-2023 layoff wave had a tidy narrative. Companies over-hired during COVID, the market corrected, and most of those jobs came back within eighteen months. Meta cut 21,000 people and then hired aggressively the following year. It was a cycle, not a rupture.
What’s happening now has a different shape entirely.
Block’s business was strong. Gross profit was growing, the company was profitable and getting more so. Dorsey didn’t cut because the business was struggling; he cut because AI tools meant fewer humans could produce the same output. These jobs weren’t casualties of a downturn. They were eliminated because they became optional.
Tech layoff tracking data shows roughly 164,000 cuts in 2022, 260,000 in 2023, and 150,000 in 2024. The 2026 numbers are accelerating, and the justifications have changed. Two years ago, CEOs blamed pandemic-era excess. Now they’re pointing at AI.
When a company over-hires and corrects, the demand for those roles still exists somewhere in the economy. When a company discovers that software can do the job permanently, the demand evaporates.
At least, that’s the narrative. There’s another way to read what happened at Block.
The mess he made
Block didn’t wake up one morning with 10,000 employees by accident. The company went from 3,349 people in 2018 to a peak of 12,985 in 2023, a 288% increase in five years. The biggest hiring spree happened between 2020 and 2022, when headcount more than doubled from 5,477 to 12,428.
Dorsey was running two public companies for most of that stretch. He served as CEO of both Twitter and Square from 2015 to late 2021, leading to six years of divided attention during exactly the period Block’s hiring went vertical. Critics said at the time that no executive could effectively run two companies of that scale. The headcount numbers suggest they had a point.
Then came the acquisitions. The $39 billion Afterpay deal closed in January 2022, adding roughly 1,200 employees and a buy-now-pay-later business already losing momentum. Former Afterpay staff described the integration as “very, very dysfunctional”, with clashes between US and Australian teams and a product that had “probably run its course.” The Tidal acquisition, which was Dorsey’s bet on Jay-Z’s music streaming platform, made headlines but never made strategic sense to most analysts.
The numbers confirm what the anecdotes suggest. Gross profit per employee dropped from $519,000 in 2021 to $482,000 in 2022, the year Afterpay’s staff were absorbed. Productivity only recovered after the cutting started, climbing to $578,000 in 2023 and $782,000 in 2024 as headcount shrank while revenue kept growing.
Block’s headcount nearly quadrupled in five years. Calling the cleanup an “AI bet” is one way to frame it. “Fixing my own mess” is another.
This week’s cut isn’t even Block’s first attempt at correction. It’s the fourth. An 8% reduction in late 2023. Roughly 1,000 more in January 2024. Another 931 positions in March 2025. And now 4,000. Hindenburg Research flagged the bloat back in March 2023, questioning Block’s cost structure and noting that Dorsey and co-founder Jim McKelvey had sold over $1 billion in stock during the run-up. Inside the company, current employees told WIRED that morale was “probably the worst I’ve felt in four years” and that “the overarching culture at Block is crumbling.”
None of this means AI isn’t genuinely changing what companies need. It is. But it’s worth asking how much of Block’s “AI transformation” is real efficiency gain, and how much is a founder undoing years of undisciplined growth under a shinier label.
The chorus
Dorsey isn’t an outlier. He’s just the loudest voice in a choir that’s been rehearsing for months.
Shopify CEO Tobi Lutke told his company in April 2025 that teams must prove AI can’t do a job before they’re allowed to hire a human for it. Read that framing carefully.
The default assumption flipped: humans now need to justify their existence against a software baseline.
Salesforce CEO Marc Benioff announced in January 2025 that the company would hire no new software engineers for the year, citing AI productivity gains. Mark Zuckerberg said, also in January, that AI would replace mid-level engineers at Meta. Klarna’s AI chatbot replaced the work of 700 customer service agents in its first month, and the company’s overall workforce shrank 40% through attrition over the following year. Duolingo cut contractors explicitly because AI could generate the content instead.
Different companies, different industries, same sentence: we need fewer people now.
Dorsey just said it with a bigger number attached.
The forecast gap
The big-picture research and the lived reality aren’t telling the same story.
Goldman Sachs estimated in 2023 that 300 million full-time jobs globally are exposed to AI automation. The World Economic Forum’s 2025 report projects a net gain of 78 million jobs by 2030, with 92 million displaced but 170 million created. The IMF estimates nearly 40% of all jobs globally are exposed to AI, rising to about 60% in advanced economies.
Net positive. That’s the headline number economists like to cite.
But “net positive by 2030” is cold comfort if you’re one of the 92 million on the wrong side of the ledger. The new jobs and the lost jobs won’t be in the same places or require the same skills. They won’t pay the same wages either. A displaced customer service agent in Kansas City doesn’t become an AI prompt engineer in San Francisco because a McKinsey slide deck says the transition will “net out.”
Macro forecasts are averages. Nobody lives in an average. The 4,000 people who lost their Block jobs this week can’t eat a net-positive projection for 2030.
That gap between the economist’s spreadsheet and the individual’s lived experience is widening faster than any retraining programme can close it.
4,000 real people
It’s easy to treat this as a strategy story. Stock goes up, efficiency improves, the playbook gets validated.
But there’s a version of this that happens at kitchen tables.
Four thousand people had plans last Monday that no longer apply. Mortgages calibrated to a salary that just vanished. Childcare arrangements built around a commute that no longer exists. Visa holders with ninety days to find a new sponsor or leave the country.
Dorsey’s letter acknowledged this, to his credit. “This decision is not a reflection of what you contributed,” he wrote. The severance package, genuinely, is better than what most companies offer. Twenty weeks plus tenure bonuses gives people a real runway.
But generous severance doesn’t change the structural problem. The job is gone. Gone because a language model and some workflow automation can now do it for a fraction of the cost, around the clock, without healthcare premiums.
The Slack channels are closing. The equity vesting one last time in May. The $5,000 to help with “the transition.”
These are the gestures of a company that knows exactly what it’s doing and wants to do it as gently as possible.
Gentle mass firing is still mass firing.
The spreadsheet itch
Every executive running a company with more than a thousand employees watched Block’s stock this week. They all noticed the same thing: the market didn’t tolerate the layoffs. It celebrated them.
That creates an incentive structure with exactly one logical conclusion. If cutting 40% of your workforce produces a 20% stock bump and your business keeps humming, you’d be failing your fiduciary duty not to consider it. The board will ask. The investors will ask. The analysts will ask. And the answer had better be more convincing than “we think our people are valuable.”
Some will do it quietly through attrition, like Klarna. Some will frame it as “transformation” or “evolution.” A few might be honest about it, the way Dorsey was. But the spreadsheet doesn’t care about the framing.
If the stock goes up when you fire people, the playbook gets copied. Call it cynicism if you want. It’s just how capital markets work.
For workers, the calculus has shifted in a way that no amount of “learn to code” advice can fix. The people who were told to learn to code are, in some cases, the ones being replaced.
The truth is probably messier than either story allows. Some of what Dorsey did this week is genuine AI restructuring, because the tools really are that good now. Some of it is a founder cleaning up his own hiring binge under a shinier label. The market doesn’t distinguish between the two.
Both make the spreadsheet work.
That’s what makes this moment so unnerving. Whether the next CEO to slash headcount is doing it because AI genuinely made those roles redundant, or because they over-hired and need a narrative that won’t tank the stock, and the people clearing their desks won’t know the difference.