The smartest VCs stopped funding AI startups. They are buying old, boring companies instead.

For two years, everyone chased the next AI startup. The real money quietly went the other way.

General Catalyst put $1.5 billion into what they call AI roll-ups. The play is simple. Buy a profitable services company that already has customers. Automate its work with AI. Use the savings to buy the next one.

Thrive raised another billion for the same move, and OpenAI took a stake in it. This is not a side bet. It is where serious capital is going.

One of these companies, Long Lake, reached $100 million in EBITDA in under two years. Then it agreed to buy a travel company for billions. It did not invent anything. It bought something that already worked and made it run leaner.

The lesson is not about AI. It is about where AI creates value. Not in building another model from zero. In taking a business that already makes money and making it cost less to run.

I build companies for a living. For the first time, buying one looks smarter than building one.

If AI can cut a company's costs in half, is it cheaper to build a new company or to buy an old one?