In "Bubbles and Crashes", Goldfarb and Kirsch offer hypotheses for what conditions must obtain (and not) for there to be a "bubble" (think "Tulips"; late-90s Internet, etc.) They then test their hypothesis by seeing whether a bubble actually arose following the introduction of a variety of technologies, from radio to transistors, nylon hose to rubber, automobiles to transistors, and many more. Their factors seem robust, and were able to predict both bubbles that did arise (eg internal combustion engine, radio, Internet commerce,) and those that didn't (eg diesel engine, tv, mainframe commerce.)
Their theory involves the interplay of 4 factors:
1. High uncertainty
2. Compelling narrative
3. Novice investors
4. "Pure play" market offerings (at least some.)
AI fits all of these perfectly.
Thus... AI is currently a bubble.
Don't scream and snarl like I'm swearing or uttering blasphemies. A Bubble is a Bubble. It's why there continues to be a frantic race, and almost uncontrolled spending. Because even though there is a powerful narrative, and belief exists that a concrete value proposition will arise; it hasn't yet. So investors, including novice investors, are buying shares of companies like Google and Meta, and frankly novice institutional investors are pouring billions into private funding rounds. (G&H "novice" investors can include otherwise experienced investors with zero experience in the new space.)
This is a bubble beyond anything remotely seen in history; very likely multiples of ALL combined bubbles prior. Before it ends, many investors are likely to lose their shirts; if AI provision is largely a utility, its profits can't rise much above a fraction of costs, and capital entry costs. But all the majors (OpenAI, Anthropic, Google, Meta, OpenAI) are locked in fierce feature and efficiency/scale battles; none are likely to emerge decided winners (eg... from which electric wholesaler providing you with power do you find your electricity most juicy and satisfying???).
Until a relatively stable market emerges, particularly with stable costs and revenue, we will remain in this crazed state of meteoric valuations. BTW... who were the most boring and uninteresting firms in the 90s? Some boring book retailer right? And a bit better search engine? Who'd buy stocks in that, right??? 😼
Act carefully. Invest carefully. And above all, don't pretend you know things you don't...