Is AI a bubble? "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: If investment climbs toward 2% of GDP, it could suggest the economy is overweighting AI relative to its productivity returns; equally, if one or more hyperscalers were to cut capex by more than a fifth over the next three to five years, that would mark a sharp turn in sentiment that could trigger a rapid decline among other players. A sustained fall in current enterprise and consumer spending levels would be another warning, especially if foreshadowed by a shrinking Nvidia order backlog. At the same time, the economics need to improve: revenues per dollar of capital should move toward the 0.5–1.0 range. Should the gap fail to close, it would imply that scale is not delivering the expected efficiencies. If valuations start approaching a P/E ratio of 50-60, it would look frothy, since a genuine growth phase should see earnings catching up with prices, not lagging further behind. If internal cash covers less than 25% of capex, data center investment stability comes under pressure. Right now, stability comes from hyperscalers’ strong cash flows. If they stop covering the majority of capex, then more debt and securitization will creep in. Not great given a GPU’s depreciation cycle. If internal funding slips below a quarter of new capex and reliance shifts toward debt and securitization, the sector’s dependence on short-lived GPUs and its exposure to higher interest rates could quickly become destabilizing. My current heuristic is that if two of the five gauges head into red, you’re in bubble territory. Time to sell up, buy the VIX and take some deep breaths. In the year prior to the Panic of 1873, the railroad’s economic strain turned red, accompanied by a decline in funding quality. Anemic revenue growth didn’t help. With the telecoms crash of 2001, revenue growth and funding quality blared red. In the dot-com era, it was industry investment strain and valuations. GenAI isn’t there yet. Racing fast, the engine is whining, but not overheating. How long would it take for two gauges to get into the red? I’ve toyed around with combinations, and most scary scenarios take a couple of years to play out. (And not all scenarios are scary.) That said, so many macro factors, from a recession in the US, to rising inflation, a challenging interest-rate environment, and domestic or international politics, could dampen spirits. While we might not be solidly in bubbleland, it would be hubristic to assume the AI investment cycle is immune to those exuberant dynamics. Onward. For now."