Key Takeaways
- • Thomas Hayes, founder and managing partner of Great Hill Capital, predicts a "monstrous bull market" will brew beneath the S&P 500 over the next six months, led by the "unmagnificent 493" rather than the Magnificent Seven.
- • Hayes argues the first phase of the AI boom rewarded infrastructure suppliers of chips, memory, cloud computing and data center equipment, while the second phase will reward traditional companies that use AI to cut costs and expand margins.
- • Named picks include natural gas producer Comstock Resources (CRK), telecom tower operator Crown Castle (CCI), logistics provider GXO Logistics (GXO), and Alibaba (BABA).
- • Hayes cites the U.S. 10-Year Treasury yield at roughly 5.3% versus an expected stock return of roughly 5%, which he says makes bonds more attractive and pressures richly valued tech stocks.
- • Hayes acknowledges risks, noting AI infrastructure revenue is still growing and a strong earnings beat by semiconductor or data center companies could reverse the rotation quickly.
Summary
Thomas Hayes, founder and managing partner of Great Hill Capital, argues that investors focused on the recent selloff in AI-themed stocks may be looking in the wrong place. He predicts that over the next six months a "monstrous bull market" will form beneath the surface of the S&P 500 Index, with leadership coming from the other 493 companies in the index rather than the Magnificent Seven that dominated the past two years, which he calls the "unmagnificent 493."
Hayes's core thesis is that the first phase of the AI boom rewarded infrastructure suppliers selling chips, memory, cloud computing capacity and data center equipment. He believes that trade has become crowded, and that intensifying competition and heavy capital expenditure could compress margins for those players. He cites the cyclical pattern in memory and semiconductors, where high prices prompt capacity expansion that later coincides with cooling demand, and describes this script as recurring roughly every four years. The second phase, in his view, will favor traditional companies that apply existing AI models to customer service, warehouse routing, administrative work and inventory management, capturing cost savings directly. He is skeptical of frontier model developers such as OpenAI and Anthropic, arguing their development costs are very high and the free cash flow needed to justify valuations is extremely demanding. He notes Great Hill Capital uses self-deployable open-weight models, including Mistral AI, Alibaba's Qwen and Meta's Llama, for internal productivity.
Hayes says he is gradually exiting parts of the semiconductor industry and rotating into natural gas, logistics, telecom infrastructure and defense. His named picks are Comstock Resources, which he views as a low-cost natural gas producer near Gulf Coast export terminals and East Texas data centers under construction, and Crown Castle, GXO Logistics and Alibaba, which he sees as a lower-cost route to Asian AI exposure. He also mentions Alibaba's investment in Moonshot AI and suggests a possible listing at a $50 billion valuation, presented only as a potential scenario rather than a confirmed IPO plan.
Hayes observes that rotation is already underway. In the October 16 U.S. session, technology and communication services led declines while utilities and energy held up, and the Dow Jones Industrial Average posted a weekly gain while the S&P 500 and Nasdaq closed lower. He names healthcare, utilities, consumer staples and rate-sensitive homebuilders as sectors drawing renewed capital attention. He also points to the U.S. 10-Year Treasury yield of about 5.3%, which compares with an unguaranteed expected stock return of roughly 5%, pressuring high-priced tech stocks to prove that AI spending will convert into cash flow.
The article notes risks to the view, including continued revenue growth at AI infrastructure players and the possibility that earnings beats from semiconductor and data center companies could reverse rotation quickly. Investors will watch earnings from Meta, Tesla and major chipmakers. Hayes's thesis would be supported by broader market breadth, with healthcare, industrials, consumer staples, utilities, financials and small- and mid-caps strengthening together.
Hayes's core thesis is that the first phase of the AI boom rewarded infrastructure suppliers selling chips, memory, cloud computing capacity and data center equipment. He believes that trade has become crowded, and that intensifying competition and heavy capital expenditure could compress margins for those players. He cites the cyclical pattern in memory and semiconductors, where high prices prompt capacity expansion that later coincides with cooling demand, and describes this script as recurring roughly every four years. The second phase, in his view, will favor traditional companies that apply existing AI models to customer service, warehouse routing, administrative work and inventory management, capturing cost savings directly. He is skeptical of frontier model developers such as OpenAI and Anthropic, arguing their development costs are very high and the free cash flow needed to justify valuations is extremely demanding. He notes Great Hill Capital uses self-deployable open-weight models, including Mistral AI, Alibaba's Qwen and Meta's Llama, for internal productivity.
Hayes says he is gradually exiting parts of the semiconductor industry and rotating into natural gas, logistics, telecom infrastructure and defense. His named picks are Comstock Resources, which he views as a low-cost natural gas producer near Gulf Coast export terminals and East Texas data centers under construction, and Crown Castle, GXO Logistics and Alibaba, which he sees as a lower-cost route to Asian AI exposure. He also mentions Alibaba's investment in Moonshot AI and suggests a possible listing at a $50 billion valuation, presented only as a potential scenario rather than a confirmed IPO plan.
Hayes observes that rotation is already underway. In the October 16 U.S. session, technology and communication services led declines while utilities and energy held up, and the Dow Jones Industrial Average posted a weekly gain while the S&P 500 and Nasdaq closed lower. He names healthcare, utilities, consumer staples and rate-sensitive homebuilders as sectors drawing renewed capital attention. He also points to the U.S. 10-Year Treasury yield of about 5.3%, which compares with an unguaranteed expected stock return of roughly 5%, pressuring high-priced tech stocks to prove that AI spending will convert into cash flow.
The article notes risks to the view, including continued revenue growth at AI infrastructure players and the possibility that earnings beats from semiconductor and data center companies could reverse rotation quickly. Investors will watch earnings from Meta, Tesla and major chipmakers. Hayes's thesis would be supported by broader market breadth, with healthcare, industrials, consumer staples, utilities, financials and small- and mid-caps strengthening together.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.