Amazon Hits $3 Trillion While Firing 30,000 People

Amazon just became a $3 trillion company. That took slightly more than two years after hitting $2 trillion in June 2024, putting it in the rarefied company of Apple, Microsoft, Alphabet, and NVIDIA. The stock jumped 4.7% on the news. Investors were thrilled. Investors are always thrilled when numbers go up.

Here's the thing about the numbers going up: Amazon has laid off roughly 30,000 people since late 2025. About 16,000 of those were corporate positions, cut in January alone. Since 2022, the company has eliminated 57,000 roles—roughly 16% of its corporate workforce. That's not restructuring. That's sustained, methodical contraction dressed up as optimization.

Amazon Web Services grew 37% last quarter, the strongest in more than four years. So the company is spending aggressively—$220 billion in capital expenditure planned for 2026, with an additional $20 billion deployed beyond the original plan. All of that money is flowing into cloud infrastructure, AI chips, and data centers. The investment thesis is clear: build the machines, not the people.

Wall Street loves this story. Amazon's consensus rating is "Strong Buy," with a mean price target suggesting nearly 30% upside. Analysts frame the layoffs as "strategic repositioning," which is the professional term for "doing more with fewer people while keeping the same bonuses." The market has decided efficiency is a virtue when it means higher margins and lower payroll.

This isn't unique to Amazon. Meta and Visa are cutting staff too. The S&P 500 hit new records. The Dow crossed 54,000. Tech companies are simultaneously making record capital investments in AI infrastructure while reducing headcount to improve efficiency. They're buying computers at scale while selling human beings at a discount.

There's a genuine paradox here, but not the kind Wall Street wants to discuss. The paradox isn't that a $3 trillion company can't afford to keep people employed. It's that we've decided efficiency means cutting the thing that makes profit possible—human judgment, experience, institutional knowledge—and replacing it with a bet on what machines will do next. We've optimized for one thing and called it optimization for everything.