When Jack Dorsey announces the cut of 40% of Block’s workforce (a departure of more than 4,000 employees) on the same day that it presents solid results with a gross profit of $2.87 billion in Q4 2025, we are not dealing with a company in difficulty. We are facing one of the most explicit demonstrations at C-Level scale that the human capital-intensive company model has an expiration date. The stock rose more than 20% in the after-hours, and investors rewarded Dorsey’s stated goal: to reach the $2 million gross profit per employeefour times the pre-pandemic efficiency. But there are two simultaneous causes in this announcement, and confusing them is a strategic error with real consequences for any decision-maker who wants to draw lessons applicable to their organization.
What does this mean? Do we have to run to the lifeboats? We continue with the thesis of the blue pill: we must keep our eyes open so that the wave does not pass over us, but that does not mean that it is a definitive sign of the apocalypse. In episode 2 of the third season of the series Mad Men, the protagonist Don Draper says a phrase that I always keep in mind when news is filled with catastrophic interpretations of changes caused by technology: “Change is neither good nor bad, it simply is.” Let us then silence the trumpets of the apocalypse and analyze what this movement means and what lies behind it.
The ‘botscaling’ thesis
First analyze what happened. The analyst and co-founder of Weblogs Antonio Ortiz precisely anticipates the concept at stake: the philosophy of botscalingthe idea that very small and hyperproductive teams can scale even further than previous generation digital companiesand that the growth will not require staff of hundreds or thousands of employees because the company is thought from scratch with artificial intelligence automating. And he concludes with intellectual honesty that perhaps the skepticism one might have had about those ideas needs to be tempered. It’s a valuable self-correction: the scenario that seemed extreme is playing out right now.
Dorsey put it in precise terms in the call with analysts collected by Bloomberg: «Something happened in December last year where the models just got an order of magnitude more capable.» That is, the same thing that Ortiz maintains: AI models multiplied their capacity and therefore their productivity. It’s not rhetoric. It is the recognition that the profitability threshold of maintaining human capital for certain cognitive functions has been crossed. Block’s own internal tool, called Goosehas been in development for two years, confirming that the operational change is deliberate and structural, not opportunistic.
The pandemic overhire
The temptation to reduce this episode to a simple case of AI adoption immersed in the hype of the agentic tsunami is great, but it is a major diagnostic error. And you have to look at the history: Block tripled its workforce between 2019 and 2022 (from 3,835 to more than 12,500 employees) during the pandemic. The current cut It doesn’t undo even half of that overhiring.. And this is not just a management error: it is the predictable result of a financing model that for years rewarded headcount growth as a proxy for ambition. Dorsey is simultaneously a victim and an actor in that system.
The founder himself acknowledged this in response to the criticism collected by BeInCrypto: «Yes, we over-hired during COVID because I incorrectly built 2 separate company structures (Square & Cash App) rather than 1.» As Ortiz points out, Block has been stagnant in income for a couple of years with irregular profits (explosion in 2024 with 2,897 million dollars, moderation in 2025 with 1,306 million). The combination of structural overhire and AI tool acceleration creates the perfect storm for a radical restructuring otherwise necessary even if AI had not provided the ideal excuse and tool.
‘AI Washing’ and the limits of the efficiency argument
Ethan Mollick, an associate professor at Wharton and one of the most respected academic voices at the intersection of AI and organizational strategy, warns according to SF Standard that effective AI tools are too new and there is little ability to reorganize work around them, making it difficult to imagine sudden 50% efficiency gains company-wide. Mollick’s warning is not an argument against AI adoption; It is a methodological reminder that without solid metrics the efficiency narrative can mask decisions that long-term TCO will end up penalizing.
The phenomenon of AI washing (companies attributing cuts to AI that were already planned or would have happened anyway) is a real risk in this cycle. Sam Altman, CEO of OpenAIhad already warned about this phenomenon. block presents enough evidence for both causes so that no serious analyst can apply Occam’s razor and be left with only one explanation.. And it is worth not forgetting that this cycle is not entirely new: industrial automation, online banking and electronic commerce followed the same sequence. Productivity is first captured by the balance sheet, then transferred by consumers via Price competition, and workers are late for delivery. The current speed is genuinely different. The underlying logic, not so much.
And there is a sad irony in this process: the engineers, product managers and analysts who built Goose and Block’s AI capabilities They are, to a large extent, those who made it possible for the company to operate today with almost half the staff.. Dorsey acknowledges this in his farewell letter: «You built what this company is today.» The phrase is more bitter than it appears.
