On 26 February 2026 Block, the payments firm best known for Square and Cash App, stunned markets and employees by announcing the elimination of almost 40% of its workforce—about 4,000 roles. Jack Dorsey, the company’s chief executive, was explicit: the cuts were not a response to shrinking demand but an intentional pivot to an AI-first operating model in which smaller teams augmented by automation can do the work of much larger staffs.
The move is striking precisely because Block entered the decision from a position of financial strength. The company reported rising gross margins in 2025 and renewed user growth across its consumer and merchant products, undercutting any narrative that the layoffs were a last resort to staunch losses.
What changed was technology. Block has deployed internally developed AI agents—codenamed “Goose”—to write code, triage customer queries and assist decision-making. Management’s calculus, as Dorsey described it, favoured a one‑time, decisive reduction to “reset” the organization rather than a drawn‑out attrition that he argued would sap morale and customer trust.
Investors rewarded that calculus. Block’s stock jumped more than 23% in after‑hours trading, a market endorsement that the short‑term productivity gains from automation are worth the human cost. The signal is loud: even profitable, well-capitalized companies believe they can improve returns by compressing headcount and redeploying capital into AI.
Block’s announcement arrived amid a broader, accelerating pattern of AI‑linked job cuts. In 2025 firms such as Pinterest, Dow, Chegg and others cited AI or automation as drivers of layoffs, and Challenger, Gray & Christmas recorded roughly 55,000 U.S. job losses attributed to AI that year. The debate that followed coalesced around a simple question—are companies genuinely replacing labour with mature, production‑ready AI, or are executives using “AI” as a more palatable euphemism for cost cutting?
There is evidence on both sides. Some AI deployments demonstrably reduce the need for repetitive tasks: code generation tools and automated customer triage can shrink hiring requirements for entry‑level programmers and support staff. Yet other cases show limits. Salesforce’s aggressive replacements prompted erosion in long‑term customer relationships, and Klarna, having cut staff in the name of automation, later rehired service workers when AI proved inadequate for complex problems.
The controversy was fed by a provocative external study, the “2028 Global Intelligence Crisis,” which modelled a cascade of economic effects from widescale AI adoption: rising profits and “ghost GDP” disconnected from household incomes, the collapse of intermediary services, falling white‑collar employment and strains on credit systems. The report—presented by its author as a thought experiment—briefly rattled markets, helping trigger sharp declines in the shares of incumbents such as IBM and American Express even as hardware suppliers to AI continued to prosper.
The immediate macroeconomic worry is intuitive. If productivity gains accrue to capital while wages stagnate or fall, consumer spending could weaken, unemployment might push more people into gig work and downward pressure on wages could erode the very demand that underpins firms’ revenues. Winners will be companies that control algorithmic platforms and compute infrastructure, while middlemen and margin‑dependent service providers risk secular profit erosion.
History cautions that technology rarely produces a neat, predictable outcome: ATMs did not eliminate bank branches and the internet created industries no one had forecast. Still, Block’s decision marks a turning point in the narrative. Automation of high‑value white‑collar work forces policymakers, corporate leaders and workers to confront not whether jobs will change, but how the gains from change will be distributed and governed.
That choice — how to bind technological progress to broad‑based economic security — will determine whether AI becomes a tool of shared prosperity or a vector of structural dislocation. For now, Block’s 4,000 layoffs are less a conclusion than a signal that the debate about automation, markets and social policy has moved from hypothetical to operational.