Humans for machines? Inside AI-driven layoffs

"For which of you, intending to build a tower, does not first sit down and count the cost, whether he has enough to finish it?" - Luke 14:28 

When you look towards 2040, what do you see? In an unprecedented move on March 31, 2026, at 6am, Oracle laid off between 2,000 and 30,000 employees globally.

There was no meeting. Just an automated email from Oracle that said, in the cleanest corporate language money can buy, today is your last working day.

By the time those employees finished reading, the system access was allegedly revoked immediately, and the company had already announced billions more in Artificial Intelligence (AI) infrastructure investments.

Same company, same quarter and same signature. They fired the humans to pay for the machines. History will decide whether that was a vision or a high-stakes bet. 

The Oracle layoff was roughly 18 per cent of its entire workforce to fund AI and cloud infrastructure expansion.

Amazon, Microsoft, and Salesforce followed the same playbook within the same 60-day window. In 2025 alone, AI was cited as a factor in nearly 55,000 layoffs in the United States, concentrated almost entirely in the tech sector. 

Now, here is the number that should make every CEO, board member, and policymaker stop: McKinsey calculates that the global compute value chain will need to invest $5.2 trillion (Sh671.16 trillion) into AI data centres by 2030.

Gartner puts worldwide AI spending at nearly $1.5 trillion (Sh193.61 trillion) in 2025, rising above $2 trillion (Sh258 trillion) in 2026. 

That is not an investment. That is a bet placed on a technology that has not yet shown a full return, and the maintenance bill arrives every year. Before the machines turn profitable, before the return justifies the risk and before anyone knows, with certainty, whether the bet was right. 

What if the math does not close? What if humanity spends the equivalent of several national GDPs, eliminates a generation of skilled workers, and discovers somewhere in the 2030s that it would have been cheaper, more reliable, and far more adaptable to keep training and paying humans?

History has a quiet way of laughing at civilisations that moved too fast and counted too slow. 

What It Means for Business

The direction is not wrong; technology costs always fall, and labour costs always rise. The internet once cost a fortune and lived only in universities. Today, it runs through the pockets of a boda boda rider and a mama mboga (greengrocer). That trajectory will repeat with AI. 

What is happening is not a typical recession layoff cycle. It is a fundamental rewiring of how companies allocate human capital in an AI-first world. The executives who understand this are not celebrating. They are counting. 

The real business question is not whether to adopt AI. It is whether your adoption curve outpaces your liability curve.

Companies that dismantle human capacity before AI capacity is genuinely ready will face an execution gap. That gap appears in customer failures, ethical lapses, and decisions that no model was trained to make, with the judgement that the moment demands. 

The leaders who win this decade are not the ones moving fastest. They are the ones building AI and human capability in parallel, knowing the handover must be earned, not declared. That is foresight leadership. 

What It Means for Policy

Kenya and the broader African policy class must watch this with open eyes. If automation displaces too many people too quickly, the economic fallout could exceed the gains. Human workers are also customers, and they need income to participate in the economy. 

This is not a technology debate. It is a social contract debate. Governments that allow corporations to export labour decisions without a workforce transition framework will inherit the consequences alone.

Education, reskilling, and labour law reform are not peripheral concerns. They are the policy infrastructure of the next decade. 

Kenya's rural electrification story is the right analogy. Connecting the country village by village was slow, expensive, and unglamorous. But it built an economy. AI transition policy needs the same stubborn patience. 

What It Means for People

The people who navigate this well will not necessarily be the most technical. They will be the most adaptive. The ones who read rooms, carry judgement, hold institutional memory, and make the call no model can make alone. That remains, for now, irreplaceably human. 

Afterthought

Oracle's 6am email should be a wakeup call. Not into panic, but into the discipline of counting the cost before the tower is half-built and the money is gone. The question Luke asked 2,000 years ago is the most strategic in any boardroom today. The answers are coming. Decisions are made on the radar screen, but the future is yours

- The writer is a human-centred strategist and leadership columnist