Artificial intelligence (AI) concept stocks have been under pressure recently, but hedge fund magnate Thomas Hayes (Thomas Hayes) believes that investors may have overlooked another change occurring in the U.S. stock market: while the S&P 500 index is still dominated by a few large tech companies, a broader range of stocks within the index are brewing a strong rally.
Great Hill Capital Founder and managing member Hayes expects that in the next six months, "the remaining 493 stocks" ( unmagnificent 493 ) among the S&P 500 index constituents, excluding a few large technology companies, may become the main driving force for market gains.
Due to the high concentration of index weights, even if a large number of constituent stocks perform strongly, as long as the stock related to the highest-weighted AI performs poorly, the S&P 500 Index and the NASDAQ Index may not fully reflect the upward momentum within the market.
Hays' judgment is based on the changes in the investment logic of AI. In the past, investors were keen to buy companies that provided chips and computing power for the infrastructure construction of AI. However, with intensified competition and increased supply, the high profit margins of these companies may become unsustainable. In contrast, enterprises that can use AI to improve productivity and expand profit margins may be the more worthwhile investment direction in the next phase.
In an interview on a podcast hosted by David Lin ( David Lin ), Hayes expressed a clear cautious attitude towards infrastructure investment in AI. He believes that a type of "commoditization trading" in AI transactions has come to an end, mainly referring to stocks such as storage chips that previously benefited from the AI infrastructure boom.
Hays explained that the memory chip industry has a clear cyclical nature: when there is a shortage in supply, manufacturers will increase production capacity; however, by the time the additional supply enters the market, demand may have already slowed down, leading to intensified competition and pressure on corporate profit margins. He summarized this process using the logic of commodity markets: "High prices will cure high prices" ( high prices cure high prices ).
He pointed out that the semiconductor and memory chip industries experience similar cycles approximately every four years: a shortage in supply drives up prices and profits, companies then increase production capacity, new competitors enter the market, and by the time the additional supply is released, demand may have already begun to weaken.
In Hays' view, companies that have invested a large amount of capital in technological expansion throughout history may ultimately become losers in this process; those that truly benefit are those that use new technologies to improve productivity and increase profit margins. He believes that the AI industry may also develop along the same path.
Hays also raised doubts about the development costs of cutting-edge AI models. He believes that the cost of developing the most advanced models has become 'out of control', and OpenAI and Anthropic face overly stringent financial requirements in order to generate free cash flows sufficient to support their current valuations.
Rather than bearing the high costs of model development on their own, the fund where Hayes works is using AI internally to improve productivity. At the same time, they have chosen to operate open-weight models, including those from French company AI's Mistral AI, Alibaba's Tongyi Qianwen (Qwen), and Meta's Llama. This reflects his investment philosophy: rather than betting on companies that continuously invest capital in building AI infrastructure, it is better to look for those that can improve operational efficiency by leveraging existing technologies.
In terms of specific investment directions, Hayes is optimistic about low-cost natural gas producers Comstock Resources. The company is located near the Gulf of Mexico export terminals, and there are also some data centers with high power consumption being constructed in eastern Texas. As the demand for electricity from data centers increases, Hayes believes that natural gas producers with cost advantages and favorable geographical locations deserve attention.
Hays also holds a very positive view of Alibaba. He pointed out that Alibaba holds equity in many AI startups in Asia and specifically mentioned its investment in Dark Side of the Moon ( Moonshot AI ). Hays said that Dark Side of the Moon could go public soon, with a valuation of up to $50 billion.
In addition to natural gas and investments related to AI, Hayes is also optimistic about communication tower operator Crown Castle as well as GXO Logistics which is engaged in logistics and warehousing services. The businesses of these companies do not rely solely on selling AI chips or building AI infrastructure; they may instead benefit from economic activities, infrastructure demand, or technological applications.
Hays believes that the market is shifting from the "Big Seven" ( Magnificent Seven ) and other AI core stocks to a broader range of constituents in the S&P 500 index. He observed that defensive sectors such as healthcare, utilities, and consumer goods are beginning to attract funds, and residential builders, which are more sensitive to interest rates, are also showing strength.
This is precisely the reason for what he referred to as a "giant bull market" possibly occurring: even if large tech stocks continue to drag down the index performance, other sectors may still continue to rise beneath the surface of the index.
Hays linked this change to the increased attractiveness of bond investments. He pointed out that the yield on 10-year U.S. Treasury bonds has reached 5.30%, whereas the 5% returns offered by stocks are not guaranteed. For investors, if they are willing to take on the risk of stock price fluctuations but can only obtain returns lower than those of Treasury bonds, some of their funds may shift towards bonds.
In his view, the rotation of funds from large technology stocks to other sectors has already begun, and it is not merely a scenario that may occur in the future. The market value-weighted structure of the S&P 500 index means that a few large constituent stocks have a significant impact on the index's performance; therefore, the overall increase in the index may not reflect the strength of other stocks.
Hays' judgment is not that AI technology is no longer important, but rather that investment opportunities may shift from companies that build technical infrastructure to those that utilize AI to reduce costs and improve efficiency. The capacity expansion cycles in the storage chip and semiconductor industries, the high costs of developing cutting-edge models, as well as the recent attention paid to defensive and interest rate-sensitive sectors, constitute the main basis for his investment approach.
