Current tech valuations look far less stretched than at the dot-com peak: the tech sector trades at a 22 forward P/E versus 55 in 2000, while the S&P 500 is at 20.4 versus 25 then.
Nvidia — often cited as the AI bubble symbol — now trades at roughly its 2019 P/E after about $1 trillion of market-cap losses this year, suggesting earnings have caught up with price.
Market concentration also looks less alarming in context: the top five stocks are about 27% of the S&P 500, near late-1960s levels, and U.S. top-10 concentration is still lower than in several major overseas markets.
U.S. private AI investment is about 20 times China’s, which the analyst frames as overinvestment rather than proof of a bubble, arguing such capital misallocation has often built the infrastructure for later winners.
The conclusion is conditional, not absolute: the analyst says today does not resemble a 1999-2000 bubble, though capital can still be misallocated and the bull market will eventually end.