Is it time to liquidate the position of "NVIDIA"?

"Perspective" is a newly launched lightweight data graphic column by 36Kr - using data to analyze trends and pictures to present key points.

"Talk is Cheap. Show me the data."

Author: Fan Liang

Recently, there have been continuous negative news in the AI sector. Michael Burry, the prototype of the "Big Short," heavily shorted AI giants Palantir and Nvidia in the third quarter. Meanwhile, SoftBank also announced that it had cleared its position in Nvidia and turned to investing in OpenAI.

Chart by 36Kr

These actions have intensified the market's concerns about the AI bubble. Data shows that Nvidia's market value has expanded rapidly, and its proportion in the total market value of US stocks has reached a historical high. The ratio of its market value to the US nominal GDP has significantly exceeded that of Cisco at the peak of the Internet bubble.

Chart by 36Kr

However, the market has not yet reached a consensus on the "bubble theory." A strong counter - evidence is that as of October 15, the proportion of short positions in Nvidia has been continuously declining this year, indicating that the short - selling power has decreased instead of increasing. In addition, historically, Burry's predictions often only come true in the short term, and his long - term success rate is not high.

Chart by 36Kr

Chart by 36Kr

Turning our attention to the Asian market, the situation is different. In the A - share AI sector, there has been capital rotation and valuation correction. Overall, it is far from reaching the "bubble" level. Its total market value only accounts for about 3.3% of the A - shares, having a limited impact on the broader market. In the South Korean market, the rising stock prices of semiconductor giants such as Samsung and SK Hynix are mainly driven by the recovery of memory chip prices, following more of the industry cycle logic rather than a simple reflection of the AI bubble.

Chart by 36Kr

In summary, the current discussion about the AI bubble is essentially a game between historically high valuations and explosive high growth. The valuation levels of US AI leaders do need to be vigilant, but their solid profitability behind is a strong support. The key to the future market lies in whether the commercialization path of AI applications can be smoothly verified. Before that, high market volatility may become the norm.