💰 Money and business · ⏱️ 2 min Latte ●●○○○
What is Circular financing?
📖 The slightly longer version
Circular financing is when a supplier invests money in a customer, and that customer then spends the money buying from the supplier. In AI, chip makers and cloud companies put money into AI labs. The labs then use it to buy chips and computing power from those same companies. Today's example: Ars Technica calls TSMC's $1 billion investment in Anthropic and OpenAI the latest case of this pattern.
Why does it worry people? Sales numbers can look bigger and healthier than the real demand from outside customers. Part of the 'growth' is the supplier's own money coming back. If the customer later struggles, the supplier can lose twice: the investment and the future sales. A bubble is when prices climb far above what things are really worth, and money loops can help hide one.
How the money goes in a circle
- Supplier investsA chip or cloud company puts money into an AI lab.
- Lab buysThe lab spends that money on the supplier's chips or computing power.
- Sales look bigThe supplier reports strong sales, partly from its own money.
- Investors cheerHigher numbers attract more investment, including pension funds.
- Risk buildsIf the lab struggles, the loop breaks and many people lose.
OpenAI's revenue is real and growing fast, but some AI money is going around in circles, and CoreWeave now asks customers to pay up front, a sign of strain. A post on Hacker News argues that retirement savings are fueling an AI bubble. If it pops, normal people's pensions may pay part of the bill, not only the big companies.
⚠️ What to watch out for
Not every supplier investment is a trick: a supplier can have good reasons to back a strong customer, and real demand for AI does exist. The hard part is that nobody outside sees clearly how much revenue is 'outside' money and how much came back around. Whether this is a dangerous bubble or normal growth is disputed, and nobody knows yet.
📰 Where it showed up
This notion came up in A safety alarm rings at Anthropic, days before its IPO. Who is listening?. More lessons: Learn.
🎯 Keep learning
Sources
Written on 2026-11-12. This explanation was accurate on that day. AI changes fast, so check the sources and the date before you rely on it.
Get The AI Daily Grind by email, every morning, in simple words.