While the long-term prospects for AI remain bright, increasing signs in the short term suggest that spending in the AI sector may slow. While this could help alleviate inflation, it could also negatively impact capital expenditures, a driver of economic growth. For investors, this means caution is advised regarding overvalued stocks:
1. Infrastructure and regulatory bottlenecks.
Many announced AI projects face practical difficulties due to power and water shortages, and regulatory pressure is increasing. For example, Texas and New York are discussing or implementing restrictions on data center construction, reducing the feasibility of infrastructure plans.
2. Fund flows in the semiconductor sector also suggest potential risks. Semiconductor ETFs (such as SOXX) have experienced parabolic expansion, and the daily rebalancing mechanism of highly leveraged ETFs significantly amplifies volatility during market downturns.
3. Market narrative shifts from euphoria to skepticism.
Recent market reactions to positive news (such as Intel's earnings report or Oracle's massive deal) have been lukewarm, indicating over-leveraging and significant valuation skepticism. The narratives of media and businesses are shifting from "fear of missing out" to a focus on actual returns;
4. China's low model training costs, policy support, and economies of scale have resulted in cheap computing power,
threatening the profits of related companies. The decline in the token price index reflects potential user frustration after limited experimental results, which may lead companies to slow down related spending;
5. Given the risk of a slowdown in infrastructure construction, investors are advised to prioritize completed or near-completed projects, or take advantage of wider credit spreads by purchasing bonds in this sector. (Source: Academy Securities)
