The AI gold rush has a body count.

Oracle is laying off 30,000 employees - roughly 20% of its workforce - to redirect funds toward AI infrastructure and data centers. It's one of the largest tech layoffs of 2026, and it crystallizes the brutal tradeoff companies are making: fire the people who built your business, bet everything on the next hype cycle.

Larry Ellison is betting billions that AI infrastructure will print money. Maybe he's right. Oracle has always been good at reading where enterprise spending is headed, and every Fortune 500 company is scrambling to build AI capabilities. But 30,000 people just lost their jobs so Oracle can build data centers to rent GPU clusters.

This is the same playbook we saw during the cloud transition: cut costs in legacy businesses, reinvest in the new thing, hope you time it right. It worked for Microsoft with Azure. It worked for Amazon with AWS. But it didn't work for IBM, which spent the 2010s restructuring and layoffs and never caught up.

The difference is that cloud infrastructure had obvious, proven use cases. Companies needed to host applications somewhere, and cloud was cheaper and more flexible than running your own data centers. AI infrastructure is different - we're still figuring out what the killer apps are. Enterprise spending on AI has been massive, but actual ROI remains unclear.

Here's what bothers me: Oracle employs some of the best database engineers in the world. These aren't interchangeable code monkeys - they're people with deep expertise in distributed systems, transaction processing, and reliability engineering. The skills that made Oracle dominant in the database market.

And now they're being fired to fund... what, exactly? GPU clusters that Nvidia and hyperscalers already provide? AI workloads that might not materialize? This feels less like strategic transformation and more like Ellison chasing the latest hype with employee livelihoods.

The technology is real. Data centers need to be built. But I can't shake the feeling that in five years, we'll look back at 2026 as the year companies fired their institutional knowledge to chase an AI bubble.