When discussing AI investment, Tom Lee pointed out a harsh but crucial fact: in the next decade, no more than 10% of AI companies may truly become good investments.
This is remarkably similar to the dot-com bubble of 2000. Countless internet companies disappeared then, but the mere 2% of survivors (Amazon, Google) were enough to significantly outperform the S&P 500 in terms of investment returns, even though the other 98% failed.

AI will also exhibit the same structural characteristics:

• 90%+ of businesses will eventually prove that their business model is unsustainable;

• However, the explosive returns of the top winners will cover all losses and still bring significant excess returns.

• Therefore, instead of betting on a single company, it is better to build an "exponential portfolio" that covers the entire AI ecosystem.