Super Bowl LX just happened. Nearly a quarter of the ads were for AI companies. OpenAI, Anthropic, Google, Amazon, Meta—all spending $8-10 million per 30-second spot.
Companies that burn billions in losses pay millions to tell you their products are revolutionary.
I've seen this before. So have you, if you were coding in 2000.
The Dot-Com Parallel Nobody's Talking About
January 2000: A leaked essay called "bubble.com" warned that internet companies were catastrophically overvalued. Nobody listened. Two months later, the market peaked and then lost 75% of its value over 18 months.
Super Bowl XXXIV (January 30, 2000) was packed with dot-com ads. Pets.com with the sock puppet. E*Trade. Seventeen different internet startups are burning venture capital on advertising.
Pets.com had its IPO eleven days after the Super Bowl. The company shut down nine months later.
The Numbers Don't Make Sense
OpenAI is reportedly discussing a funding round that would value it at $830 billion—more than Coca-Cola, Chevron, or Costco.
The company projects:
- $14 billion in losses for 2026
- $115 billion in cumulative losses through 2029
- $20 billion in revenue against $1.4 trillion in compute commitments
That's 1.4% of commitments covered by revenue.
HSBC analysis: Even if OpenAI hits $200 billion in revenue by 2030, they'll still need an additional $207 billion in funding to stay in business.
Microsoft lost $440 billion in market capitalization in a single day after investors questioned its AI spending plans.
The Debt Bubble Under the AI Bubble
Companies supplying infrastructure to OpenAI have taken on $96 billion in debt to fund operations. The big five tech companies (Amazon, Google, Meta, Microsoft, Oracle) added $121 billion in new debt in 2025 alone—four times their historical average.
Oracle is being sued by bondholders for concealing the amount of debt they'd need for an AI buildout. Their credit rating is approaching junk status.
This is exactly how bubbles work: companies with unproven business models borrow unsustainable amounts, then spend desperately on advertising to create the appearance of legitimacy.
The Tools Aren't Going Away
I use Claude Code daily. AI coding assistants are genuinely useful when used correctly. But they're tools that require human judgment, not replacements for it.
The valuable tools will survive the crash. The rest won't. Companies that burn billions while losing money won't make it. The ones with sustainable unit economics might.
What This Means for Developers
Short term: Keep using AI tools while they're heavily subsidized. Learn what works and what doesn't. Build things you couldn't build before because the tedium was too high.
Long term: The opportunities after the crash are probably more interesting than what we're working on now. The dot-com crash gave us Google, Amazon, and modern web development. The AI crash will leave us with whatever actually works.
The code still needs to be written. The bugs still need to be found. The architecture still needs human judgment.
That's not changing, regardless of how many AI companies burn through their funding. That said, I am writing this article in a blog pair-programmed with Claude. Grammarly checked it and offered suggestions. AI is everywhere, and it can't be stopped, no matter the economy.
And yeah, the image is also AI 😄
Originally published at intelligenttools.co
Top comments (0)