/
Published on 12.03.2026
TLDR: Block slashed its workforce from 10,205 to under 6,000 in February 2026, framing the layoffs as AI-driven efficiency rather than cost-cutting. Wall Street rewarded the move with a 24% stock jump, but the narrative deserves far more scrutiny than it is getting.
Jack Dorsey's fintech company Block did something in February 2026 that should make every technology leader pause and think very carefully. They eliminated roughly 4,000 positions, nearly 40% of their entire workforce, and did not apologize for it. Instead, they posted it as a strategic victory: AI made those jobs unnecessary. The stock price jumped 24% on the day of the announcement. That reaction alone tells you everything about where the incentive structures currently sit in public markets. Investors are not rewarding companies for building sustainable workforces. They are rewarding companies for demonstrating they can do the same work with fewer humans.
The productivity claims are specific enough to be worth examining. Block says engineers using their internal AI tooling now ship 40% more production code than they did six months prior. Non-technical staff reportedly build their own workflow applications in hours, bypassing IT departments entirely. The company committed 130 million dollars to AI infrastructure for 2026, funded largely by the salary savings from the people they let go. That is a fascinating feedback loop: fire people, take their salaries, invest in the technology that replaces them, then report higher per-capita productivity. The math works on a spreadsheet. Whether it works in practice over twelve to eighteen months is an entirely different question.
What is missing from this narrative is any serious discussion of institutional knowledge loss. When you remove 4,000 people from a 10,000-person organization, you are not just removing headcount. You are removing the humans who understand why certain architectural decisions were made, who know which edge cases the documentation never captured, who carry the context that makes complex systems maintainable. AI tools can generate code faster, but they cannot tell you why the payment reconciliation service has that one weird timeout configured at 47 seconds instead of 30. Someone knew. That someone might have been in the 4,000. Block's CTO apparently wrote an internal AI manifesto two years before the cuts, which suggests this was planned and methodical. But planning a reduction and successfully executing one without degrading system reliability are two very different skills.
The broader pattern here is worth watching closely for anyone in a technical leadership role. Block killed its entire Web3 division to fund AI instead, which is at least intellectually honest. Most companies try to do everything at once. Dorsey made a clear bet: agents over tokens, AI over crypto. Whether that bet pays off depends entirely on whether their internal AI tooling is genuinely as capable as claimed, or whether the productivity numbers are being measured during the honeymoon period before accumulated technical debt and knowledge gaps start creating real operational problems.
For architects and engineering leaders, the lesson is not that you should slash your teams in half. The lesson is that companies are now publicly positioning AI-driven workforce reduction as a competitive advantage, and your board has noticed. If you do not have a credible answer for how your team is leveraging AI tooling to increase output, someone else will provide that answer for you, and their version will probably involve fewer of your people. The smarter move is to get ahead of it: instrument your own productivity metrics, demonstrate AI-augmented gains transparently, and make the case that your team with AI is more valuable than a smaller team with AI. Because the Block playbook is now in the water, and every CFO in the industry is reading it.