Microsoft AI CEO Mustafa Suleyman just made the boldest prediction of 2026: every white-collar job that involves "sitting down at a computer" will be fully automated within 18 months.

There's just one problem: the evidence says he's wrong.

In a Fortune interview published Friday, Suleyman told the Financial Times that AI will achieve "human-level performance on most, if not all professional tasks" within 18 months. He specifically named accounting, legal, marketing, and project management as vulnerable. Any task involving computers, he said, will be "fully automated."

But here's what the research actually shows: AI is making workers slower, not faster.

A new study from nonprofit METR found that AI actually made software developers' tasks take 20% longer. Not faster. Longer. The AI tools that were supposed to automate coding instead became productivity drains.

The market seems to agree with the research, not the prophecy. Apollo Global Management's chief economist Torsten Slok notes that while Big Tech profit margins jumped 20% in Q4 2025, the broader Bloomberg 500 Index shows almost no change. "Investors do not believe AI will result in higher earnings outside the tech sector," Slok said.

The legal and accounting sectors that Suleyman targets? Thomson Reuters' 2025 report found lawyers and accountants are indeed experimenting with AI for document review and routine analysis. But results show "marginal productivity improvements" that fall "far short of mass displacement."

The disconnect is staggering. Suleyman plans to build "superintelligence" that can "perform any required job function." Yet 49,135 AI-related job cuts have already happened in 2026 according to Challenger, Gray & Christmas, suggesting companies are eliminating AI roles, not traditional ones.

This echoes the February "SaaSpocalypse" selloff that wiped $150 billion from software stocks after Anthropic and OpenAI launched enterprise AI systems. The promised productivity revolution hit Wall Street like a tsunami. But where are the results?

Suleyman isn't alone in making these bold claims. Anthropic's Dario Amodei warned in May 2025 that AI could wipe out half of entry-level white-collar jobs (though he recently changed his tune citing Jevons paradox). Ford CEO Jim Farley made similar predictions about white-collar job cuts.

Matt Shumer's viral essay in Fortune compared this moment to February 2020 before the pandemic: "This will be more dramatic." Josh Tyrangiel warned in The Atlantic that the U.S. isn't prepared, comparing CEOs' silence to "a shark fin break the water."

But here's the thing: the most powerful AI company's chief executive is predicting total automation while his industry's own research shows AI making workers slower. Either the research is wrong, or the timeline is fantasy.

I've been watching this space long enough to know that AI predictions follow a predictable pattern: breathless announcements, market chaos, then reality. The $150 billion SaaSpocalypse selloff was real. The productivity revolution? Still waiting.

Eighteen months from now, when accountants and lawyers are still using AI as a productivity aid rather than replacement, remember this moment. When Suleyman's timeline passes without the promised automation revolution, ask yourself: why do we keep believing predictions that contradict the evidence?