Key Takeaways
- • Hedge funds maintain significantly higher overall exposure to the artificial intelligence trade compared to mutual funds, which remain underweight the sector despite increased infrastructure-related allocations in 2026.
- • Sharp divergence exists on mega-cap technology stocks: hedge funds increased holdings in Microsoft and Amazon during Q2 2026, while mutual funds reduced exposure to both companies.
- • Hedge funds reduced positions across semiconductor and AI-adjacent names including Nvidia, Broadcom, Lam Research, Marvell Technology, Applied Materials, and Advanced Micro Devices, while mutual funds took opposing views on AMD, Micron Technology, and similar semiconductor plays.
- • Both investor groups found consensus on 12 infrastructure-focused companies spanning power generation (American Electric Power, NiSource, Xcel Energy), data centers (CoreWeave, Flex, Sanmina), and related technology infrastructure providers (SiTime, Seagate Technology, Talen Energy, AXT, Bloom Energy, Legence).
- • According to Goldman Sachs' Q2 institutional portfolio analysis, mutual funds remain below benchmark allocations in AI-related infrastructure spending despite year-to-date increases in this category.
Summary
Goldman Sachs' analysis of institutional portfolios during the second quarter of 2026 reveals a widening strategic divergence between hedge funds and mutual funds on artificial intelligence positioning. Hedge funds have maintained their core commitment to the AI trade with continued strong exposure, while mutual funds lag the broader theme despite recent efforts to increase infrastructure-related holdings.
The most pronounced split emerged in mega-cap technology equities. Hedge funds accumulated positions in Microsoft and Amazon during Q2, viewing these companies as beneficiaries of sustained AI capital deployment. Mutual funds, by contrast, reduced their exposure to both stocks, signaling a more cautious stance on these higher-valuation names. This divergence reflects fundamentally different risk appetites and portfolio construction philosophies between the two investor classes.
Semiconductor and memory companies presented an equally stark picture of misalignment. Hedge funds broadly retreated from positions in prominent AI-adjacent chipmakers—Nvidia, Broadcom, Lam Research, Marvell Technology, Applied Materials, and others—during the period. Mutual funds took the opposite tack, with a majority of managers increasing holdings in Advanced Micro Devices, Micron Technology, and Sandisk, even as hedge funds reduced exposure to these same names. This positioning suggests mutual funds may be seeking value in cyclical semiconductor exposure as hedge funds rotated capital elsewhere.
The investor groups found meaningful common ground around infrastructure enablement. Goldman identified 12 companies attracting simultaneous buying from both hedge funds and mutual funds in Q2. This consensus list spans multiple infrastructure categories: power utilities American Electric Power, NiSource, and Xcel Energy; data-center operators and infrastructure providers including CoreWeave, Flex, and Sanmina; semiconductor equipment and materials suppliers SiTime and Seagate Technology; energy infrastructure player Talen Energy; and smaller technology infrastructure names AXT, Bloom Energy, and Legence. This alignment reflects mutual recognition that physical and electrical infrastructure represents a durable, less-contested investment thesis within the broader AI trade.
Mutual funds have increased allocations to AI infrastructure companies during 2026, yet according to Goldman's assessment, these weightings remain below benchmark levels, suggesting further room for expansion if conviction deepens. The divergence between hedged and traditional institutional allocators underscores the ongoing bifurcation of the AI investment narrative—from consensus infrastructure plays to highly contested positioning in semiconductor and software equities.
The most pronounced split emerged in mega-cap technology equities. Hedge funds accumulated positions in Microsoft and Amazon during Q2, viewing these companies as beneficiaries of sustained AI capital deployment. Mutual funds, by contrast, reduced their exposure to both stocks, signaling a more cautious stance on these higher-valuation names. This divergence reflects fundamentally different risk appetites and portfolio construction philosophies between the two investor classes.
Semiconductor and memory companies presented an equally stark picture of misalignment. Hedge funds broadly retreated from positions in prominent AI-adjacent chipmakers—Nvidia, Broadcom, Lam Research, Marvell Technology, Applied Materials, and others—during the period. Mutual funds took the opposite tack, with a majority of managers increasing holdings in Advanced Micro Devices, Micron Technology, and Sandisk, even as hedge funds reduced exposure to these same names. This positioning suggests mutual funds may be seeking value in cyclical semiconductor exposure as hedge funds rotated capital elsewhere.
The investor groups found meaningful common ground around infrastructure enablement. Goldman identified 12 companies attracting simultaneous buying from both hedge funds and mutual funds in Q2. This consensus list spans multiple infrastructure categories: power utilities American Electric Power, NiSource, and Xcel Energy; data-center operators and infrastructure providers including CoreWeave, Flex, and Sanmina; semiconductor equipment and materials suppliers SiTime and Seagate Technology; energy infrastructure player Talen Energy; and smaller technology infrastructure names AXT, Bloom Energy, and Legence. This alignment reflects mutual recognition that physical and electrical infrastructure represents a durable, less-contested investment thesis within the broader AI trade.
Mutual funds have increased allocations to AI infrastructure companies during 2026, yet according to Goldman's assessment, these weightings remain below benchmark levels, suggesting further room for expansion if conviction deepens. The divergence between hedged and traditional institutional allocators underscores the ongoing bifurcation of the AI investment narrative—from consensus infrastructure plays to highly contested positioning in semiconductor and software equities.
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.