A boom-bust signal preceded the dot-com crash and the 2008 housing crisis. It may now be worth watching for in the artificial intelligence (AI) trade too.
The signal is known as a “funding gap,” and the Wall Street Journal traces it through two prior busts. Columnist Jonathan Weil argues the AI trade depends on that same flow of capital.
The Boom-Bust Signal in 2000 and 2007
Weil lists several other risks facing the AI trade, from data center power shortages to disruption by smaller rivals. He argues that a closing capital window remains the most immediate threat to watch.
In 2000, “burn rate” and “funding gap” entered common use as the dot-com bubble deflated. Pets.com kept raising cash despite selling products at a steep loss, until the money stopped and it folded months later.
A similar pattern emerged in November 2007. Freddie Mac slashed its dividend and sold $6 billion in preferred stock. The move aimed to shore up capital after mounting mortgage losses.
Freddie Mac still could not raise the common equity investors wanted. That gap helped set up the crisis that ended in government conservatorship the following year.
Is the AI Trade Showing the Same Cracks?
OpenAI and Anthropic are burning cash with no public profits, according to the Journal analysis. OpenAI’s ability to keep raising money leans partly on continued backing from Nvidia (NVDA) and SoftBank.
Anthropic’s IPO timeline has already shifted from October to November. Holtec Nuclear, a supplier to AI data centers, paused its listing this month too.
Still, the Journal is careful to note these delays look like early caution, not proof the funding window has closed. Oracle’s (ORCL) debt-fueled data center spending depends on OpenAI continuing to pay for compute. That ties the wider AI trade to a handful of privately held labs.
The Journal argues booms can end suddenly or unwind slowly. Both prior busts began the moment funding gap re-entered the conversation. Whether that phrase returns for AI, it adds, is something the market will likely only recognize in hindsight.
