BlockBeats News, June 3rd, US stock market KOL Herman Jin (@ShanghaoJin) elaborated in a post yesterday analyzing the four potential risks under the AI investment frenzy: at the micro-company level, the seemingly perfect situation can easily lead to overconfidence, liquidity is abundant but central bank tools are limited, the semiconductor supply chain is out of control, driving up capital expenditure costs, and the slowing revenue growth of models such as Anthropic/OpenAI may shake the Hyperscalers' capital expenditure expectations.
The author pointed out that the demand risk of AI model end (Anthropic/OpenAI) is the most worthy of attention and the most intuitive indicator. Although the current AI bull market appears to have low PE and low foam, in reality, model revenue must maintain high-speed growth in order to rationalize the massive capital expenditure of Hyperscalers (Microsoft, Google, Amazon, Meta, etc.). Once the model revenue growth slows down, "de-intelligence" appears, and the computing power bottleneck leads to token quality/efficiency issues, the market's expectation of the "sustainability of capital expenditure returns" will be shaken.
Once the expectation is shaken, it will directly pierce the "lifeblood" of this bull market, causing the low PE foam to transition from "strong positive feedback" to panic selling. At the same time, "all valuations that have not fallen under the negative news before, supported by confidence, will experience a one-time reset."
