On the basis of these gauges, genAI remains in a demand-led, capital-intensive boom rather than a bubble. But booms can sour quickly, and there are several pressure points worth watching
Subscribe for new posts
Blog posts only. No commonplace entries. Never sold or shared.
The truth is that bubbles are only recognizable in retrospect; what may appear to be a “boom” has to “burst” in order to be a bubble. However, identifying signs that might indicate that a boom is likely to be a bursting bubble is extremely valuable.
Acknowledging this, the authors cover a lot of ground and come up with five factors to watch: economic strain (capex / GDP %), industry strain (investment / revenue), revenue growth (revenue doubling time in years), valuation heat (price / earnings ratio), and funding quality (a composite index).
Their conclusion — AI is still firmly in “boomb” territory, rather than looking like a bubble. Helpfully, they also provide a few warning signs to watch for: if investment in AI approaches 2% of GDP, sustained drops in enterprise and consumer spending, P/E ratios approaching 50-60, and if internal cash covers less than 25% of capex. Two of these would push the authors into the danger zone.
Is this a perfect study? Probably not. Is it possible they’ve misread the data? Certainly. But what is certainly true is they’ve taken the question seriously, looked for unbiased indicators, tested their validity qualitatively and quantitatively, and even shared how they might be wrong.
We need a lot more of this and a lot less of the smug, overconfident, tweet-length assertions that grab most of the headlines today.