When Jack Dorsey announced last week that Block — the parent company of Square, Cash App and Afterpay — would do just that Reducing its workforce by 40 percentAfter cutting more than 4,000 jobs and cutting its headcount to fewer than 6,000, the capital markets’ reaction was immediate and brutal in its clarity: Block’s stock rose more than 22 percent in after-hours trading. Analysts described it as a kind of decisive moment. Dorsey described this as inevitable, and warned that the majority of companies You will follow suit During the year.
It’s not just a restructuring story. This is one stress test of the defining question of the AI era: Has AI gone from being a productivity booster to a structural headcount reducer?
What a rewarding market it actually is
The stock rally delivers an important and, frankly, somewhat uncomfortable message. Investors aren’t rewarding Block because AI has conclusively proven it can run fintech at half the human cost. They are equivalent to the margin thesis. Block has guided its adjusted operating profit margin to hit 26 percent in 2026, up from 17 percent in 2025. That’s a very interesting number, and capital markets are pricing in wishful thinking.
However, the reality is that much of what is celebrated is expected competence rather than mere demonstration efficiency. A Harvard Business Review study, published in January, found that companies often lay off workers It is largely based on the expected capabilities of artificial intelligence– Not its proven performance. After years of massive investment in infrastructure, markets are under pressure to find returns. As a result, they reward the signal of ambition as much as the reality of practical implementation. This distinction is huge.
Is the collar shrinkage starting?
Yes and no. The honest answer is that this is real in the structural sense, and it is accelerating in the strategic sense.
AI already automates some categories of administrative work: writing code, compliance documentation, synthesizing data, and routing customer inquiries. These are real efficiencies. Ban reportedly automated large portions of its software engineering processes before making these cuts. This is not fiction.
But a 40% headcount decline at this point in AI adoption is ahead of the curve to say the least. Human judgment, situational reasoning, institutional knowledge, and the kind of adaptive problem solving that fintech requires at scale have not yet been reliably replaced. Operating payments infrastructure across multiple geographies, and managing regulatory relationships and trust at the consumer level carries layers of human responsibility and accountability that current AI systems do not fully capture.
Dorsey may be right that others will follow. But the fastest followers may not be the ones who fail the most or the most resilient. Companies that seize this moment as a competitive signal to reduce their headcount without true operational readiness for AI are taking real risks, not only to their business execution, but also to their institutional knowledge base.
What Dorsey got right all along
Despite justified skepticism about the timing, it would be intellectually disingenuous not to acknowledge how this matter was handled. The severance package, which is said to include at least 20 weeks of base salary with additional compensation based on length of service, is among the most generous in recent tech history. Dorsey’s internal memo was direct and transparent. He didn’t just hide behind the typical “restructuring” language. He openly cited AI, took ownership of the decisions and treated his employees with economic dignity.
In an environment full of subtle, ambiguous, right-sized messages and ambiguous termination terms, one should appreciate this clarity. Leaders who are honest about transformational disruptions—even uncomfortable ones—gain a different kind of trust from the marketplace, from employees, and from the public. This is a leadership attitude that more executives should learn from.
What this means for the AI ecosystem
For me, this moment represents an architectural reorganization. This is a fundamental change in where value is captured across the AI stack. With each change, it is helpful to look below the model layer to fully understand it.
Models, computation, and APIs have become centralized and commoditized. Most importantly, we are seeing the emergence of the Temporary Agent Operating System (TAOS), an execution infrastructure layer that enables AI agents to orchestrate workflows, maintain state, and execute consistently across complex operational environments. This is not a feature of the models. This intelligent operating layer is what will see AI evolve from a mere productivity tool to a true hybrid workforce enabler – and it’s growing fast.
Block’s move is a bold bet. If AI tools succeed, Dorsey will seem like a visionary. But when operational gaps arise – in governance, in product excellence, in customer trust or all of the above – accountability will be just as transparent.
The lesson for this industry is simple. Don’t just copy the percentage. Understand what a temporal proxy operating system can do for you sustainably – and design your human organization around what it can’t do.
Youssef Khalili is the Global Head of Transformation and CEO of the Middle East and Africa region at the company Quantitywhich works to develop cutting-edge digital employee technology.
