The Bubble You Are Inside Does Not Feel Like One

Someone who was in the room the last time watched a stock collapse down to nothing wrote a piece this week comparing the dot-com moment to where we are now with AI. The argument is careful, and the conclusion is unsettling: the signals that precede a burst are present, and this time the fallout could be worse.

I have read a lot of these comparisons. Most of them feel academic. This one did not, because it forced me to ask a question I have been quietly sitting with for months: are the companies our teams are working inside building on the real thing, or on the narrative?

Here is what I know from being close to the work. The AI investment case is not uniform. Microsoft, Alphabet, Amazon; these are profitable businesses generating real cash, embedding AI into products people actually pay for. That is not dot-com math. But the infrastructure layer, the chip plays, the AI-adjacent platforms, a lot of them are trading at valuations that assume a straight line from "we are building AI stuff" to "we are making money from AI stuff." That straight line has never existed in tech. Not once. And I'm not sure that it does now.

What I watch for in client conversations is simpler than stock multiples. I watch for the projects that exist because someone had to say yes to something AI-related, and the projects that exist because someone solved a real problem and AI happened to be the right tool. The ratio is still uncomfortably low, but not non-existent. The urgency to deploy is outpacing the discipline to deploy well, and that's where the risk shows up for me. When the cycle turns (it always turns, it's just a matter of when), the companies that built on urgency will have the hardest time explaining what they got for it. It's the thing that may actually fall down - that's where I'm seeing the bubble sit right now.