The AI trade could rapidly unravel – and one hedge fund is preparing for the fallout

Selwood Asset Management CIO Karim Moussalem warns that the AI boom constitutes a speculative bubble with significant risks, citing rising energy prices impacting AI scalability, overly optimistic depreciation of AI infrastructure overstating profits, and potential regulatory scrutiny. He is positioning his fund by going long energy-intensive commodities like uranium and copper, viewing them as undervalued direct beneficiaries of AI's power demands. This bearish outlook is reinforced by other prominent investors, including Leon Cooperman and Georges Debbas, who also express concerns over "ridiculously high" AI valuations and the sector's long-term sustainability.

Analysis

A contrarian view on the AI-driven market rally is emerging, articulated by Selwood Asset Management's CIO, Karim Moussalem, who characterizes the current environment as a "speculative mania" and a "total retail driven frenzy." The core of this bearish thesis rests on two primary vulnerabilities: overstated profitability and energy constraints. Moussalem argues that AI hyperscalers are using optimistic depreciation schedules of six to eight years for capital expenditures, while rapid innovation from firms like Nvidia (NVDA) suggests a one to two-year cycle would be more appropriate, implying current profits are "vastly overstated." Concurrently, the scalability of AI is threatened by rising energy prices and grid capacity limitations, with Moussalem noting that energy costs are going "parabolic." This has led his fund to pivot towards commodities, establishing long positions in uranium and copper, which are viewed as direct, undervalued beneficiaries of AI's immense power demand. He specifically highlights a "huge deficit looming" in the uranium market. This cautious stance is corroborated by other market figures; Leon Cooperman notes AI valuations are "ridiculously high," and BNP Paribas' Georges Debbas suggests owning portfolio protection makes fundamental sense due to structural risks heading into 2026.

Key Decisions for Investors

• Investors should critically re-evaluate exposure to high-valuation AI and technology stocks, particularly unprofitable firms which are up 120% since April lows, given the risk of a speculative bubble and potentially overstated earnings.
• Consider initiating or increasing long positions in commodities essential for AI's energy infrastructure, such as uranium (e.g., URA, URNM) and copper (e.g., CPER, COPX), which offer a direct play on the theme at more reasonable valuations and benefit from a strong supply-demand outlook.
• Closely monitor US energy price trends, hyperscalers' reported capital expenditure and depreciation policies, and any increase in regulatory scrutiny of large tech companies, as these are identified as key potential catalysts that could trigger a correction in the AI trade.