The $3.5 Trillion Question: Is AI a 1999 Bubble — or Is It Different?

Robert ShillerA bubble is a self-reinforcing feedback loop where rising prices generate the very story that justifies them.

• Shiller's warning is psychological, not technical: prices rise → a narrative forms to explain them → the narrative pulls in new buyers → prices rise again. AI's narrative ("the most important technology since electricity") is powerful precisely because it may be true.
• His CAPE work showed that the market's mood, not just its earnings, sets prices — and mood is mean-reverting. The more a valuation depends on the story staying loud, the more fragile it is to the story merely getting quieter.
• The test Shiller would apply: strip the narrative and ask what the cash flows alone justify. If the gap between "story price" and "cash-flow price" is enormous, you are not buying a business — you are buying a mood.

Philip FisherThe wise investor pays a fair price for an outstanding company rather than a bargain price for a mediocre one.

• Fisher is the counterweight to the bubble chorus. He built his record buying superb growth businesses at prices that looked expensive — and being right because the growth was durable, not fleeting. A high multiple is not, by itself, a verdict.
• His discipline was qualitative and brutal: does the company have a long runway, real management depth, and an advantage competitors can't easily copy? Applied to AI, the question becomes which players have a durable edge versus which are renting a temporary lead with cheap capital.
• The Fisher trap to avoid: using "great companies deserve premiums" as a blank cheque. A premium is justified by durable growth you can articulate — not by a stock that has already gone up.

Warren BuffettRisk comes from not knowing what you're doing. Be fearful when others are greedy.

• Buffett's first filter is not valuation — it is competence. If you cannot confidently model a company's economics a decade out, the correct position size may be zero, regardless of how exciting the story is. Most AI valuations fail this test honestly applied.
• His second filter is price. "Be fearful when others are greedy" is not a market-timing call; it is a reminder that the crowd's enthusiasm is the moment to demand a bigger margin of safety, not a smaller one.
• And his quiet point for index investors: you may already own the AI trade without choosing it. Knowing your real exposure — through the concentration in your funds — is itself an act of discipline.