AI summary and analysis
Marks responded directly to the AI bubble issue. He did not simply give a binary answer, but split it into four levels: corporate investment behavior, investor psychology, AI infrastructure and market pricing.
Key points
- He believes that bubbles are first of all a psychological state, not just valuation numbers; when investors ignore prices, risks and uncertainties because of fear of missing out, risks begin to accumulate.
- AI may belong to a technology bubble that can promote social progress, but this does not mean that all participants and high-priced assets can make money for investors.
- Marks distinguishes between good and bad bubbles: some bubbles accelerate infrastructure construction, but capital losses may still be borne by investors chasing prices higher.
- He is particularly concerned about circular transactions, large capital commitments and "lottery-style" expected return calculations, as these may weaken real commercial return constraints.
- This memo is very suitable to be placed next to the AI theme on the stock page to remind investors that the long-term correctness of the industry does not mean that the current price is safe.