Two reports from S&P Global Market Intelligence point to evolving trends in private equity investment in artificial intelligence, and the associated risks that the general partners may not have fully taken on board.
According to one report says private equity investment into industrials, driven largely by enthusiasm for AI-related propositions, is on track to reach a five-year high, with some US$82.06 billion already invested worldwide by the end of May. If the momentum is maintained, the year-end total could exceed the $160.47 billion invested in industrials in 2022.
Admittedly, some of this momentum is driven by investment in defence and aerospace sectors. But most is from AI. The report quotes Marc Rowan, chief executive officer of Apollo Global Management Inc., saying: “This is infrastructure. This is energy. This is energy transmission. This is an energy transition. This is advanced manufacturing. This is defence, and it’s AI and data.”
The second report warns of limited partners’ concentration risk in private equity overcommitment to large AI funding rounds. LPs may be overexposed to AI assets as their GPs execute parallel investments regardless of what that does to their investors’ overall portfolio exposure. Proportionately small LP commitments to each fund may aggregate into uncomfortably large sector bets across several funds.
Huge commitments to single bets have a way of unwinding in time. Consider the recent reset in market value for leading technology companies. The so-called Magnificent Seven tech giants — Alphabet, Amazon, Apple, Meta, Microsoft, NVIDIA and Tesla — shed around $2.3 trillion in market value in June as markets soured on their AI-related capital expenditure plans. One wonders how soon another tweak in the AI proposition may bring similar value destruction to other assets exposed to the technology.
Furthermore, many of the assets now being invested by private equity in its dive into AI infrastructure are just that — infrastructure. This has historically been considered as a separate asset class precisely because its operational demands and time horizons are very different from regular private equity investment in corporates.
Are all the eager private equity investors into AI-related industrials and infrastructure really going to be able to manage value-accretive value creation? And what about the risk of another reset in valuations once the AI model shifts once again? All those bumper investments in industrials in 2026 could be put at risk.
Private equity is absolutely supposed to be a long-term long-dated asset class, investing across the short-term shifts in the markets. This also ought to mean some robust indifference to short-term investment fads. With AI, one really has to wonder.
