The trillion dollar AI bubble
Big tech is pouring hundreds of billions into AI but the money coming back is tiny, and a crash like the dotcom bust could hit everyone, not just tech investors.
- AI companies would need to make $600 billion a year to justify current spending, but most are burning cash — OpenAI, Anthropic, and xAI are all losing billions while raising more.
- Unlike normal software, every AI query costs real money, so more users makes things more expensive, not cheaper.
- 90% of CEOs say AI will change their business, but only 1 in 4 can explain how it actually makes money — it looks like corporate peer pressure, not strategy.
- Data centers are eating power and water at huge scale, with some single facilities using as much water as a town of 30,000 to 50,000 people.
- A third of what workers paste into AI tools is now sensitive company data, and most companies have no controls to stop it — researchers call it the largest uncontrolled corporate data leak ever.
- Chipmakers like NVIDIA, AMD, and TSMC are the real winners, and the shortage they fuel is now pushing up prices on laptops, phones, and gaming PCs.
Outlook: If AI revenue does not catch up to spending soon, a dotcom-style correction could wipe out huge value across retirement accounts and index funds, not just tech stocks.