Eduard van Gelderen
Just recently, I attended an interesting conference organized by With Intelligence, an investment data platform. One of the panel discussions was on private markets, and allocators and CIOs on the panel expressed their interest in allocating to private assets. The rationale behind their comments was not new: Private markets provide interesting returns and offer investment opportunities you will not find in public markets. Placing these comments in a Total Portfolio Approach perspective, private assets seem to be the logical choice. But are they?
Last January, Scientific Infra & Private Assets, an EDHEC venture (recently acquired by the PEI Group), released its capital market assumptions and compared those with other CMAs of large asset managers. SIPA’s estimate of expected returns for infrastructure assets was more than 11%, which compares to a range of 7% to 11% for its peers. For private equity, the numbers are more than 12% versus a range of 10% to 13%. Based on these numbers, the CIOs are right: With the yield on 10-year U.S. Treasurys hovering just above 4%, the returns on private assets seem to be rewarding. Moreover, real assets make up a large percentage of modern economies and play an important role in the build-up of our digital infrastructure, energy solutions, smart cities and other parts of a sustainable society. The CIOs seem to be right on this point, too.
It almost feels similar to a Goldilocks environment: What can go wrong? Obviously, a war in the Middle East, which hopefully will end soon. The essence of this column remains the same.
Short Term vs. Long Term
Lately, many corrections in market prices and valuations are due to one theme: artificial intelligence (apart from the recent war in the Middle East). In the public markets, we’ve seen the dominance of the Magnificent 7 large-cap technology stocks, driven by the demand for AI. At times, the question is: Are we too optimistic about this?
The New York Times, in August 2025, published an article stating that the billions of investments in AI had not paid off at all. An MIT study claimed that 95% of AI projects fail to deliver measurable profit impact within six months. Jessica Pollock, a researcher at, FCLTGlobal, responded to these news items by claiming that it does not make much sense to focus on such a short investment horizon. She is absolutely right: There might be some quick wins, but the real benefits are not the obvious productivity gains, but the wholesale redesign of business processes. This takes longer than six months. But harm from the focus on short-term results was already done, as stock prices tumbled, with only two of the companies—Alphabet and Nvidia—outperforming the S&P 500 Index in 2025.
More recently, we see articles popping up emphasizing the disruptive nature of AI for certain industries. This time, the emphasis is on private equity portfolios and the exposure to software companies. Since the late 1990s, the introduction of software-as-a-service companies offered a business model with recurring—and growing—cash flows, making software companies interesting investments for private equity. The story is that new entrants to the industry, using AI, are providing attractive alternatives to traditional SaaS options. This is normal competition, but for sure, several software companies did not see this coming. The problem for the investor is that the tradability of these software companies, owned by private equity firms, is limited, which will lead to potentially significant write-offs.
Can We Predict the Disruption?
It’s not just the software companies that face competition from AI: Eventually every industry will be impacted. How is AI going to change, or even disrupt, the business models of the portfolio companies to which investors have exposure? And why is it so difficult to see the disruption coming?
In large part, this is because the AI labs from only a handful of companies are laying out the AI foundation, and developments move very quickly. Most people simply do not know what the research agenda of these AI labs look like. Available AI tools are built on top of this foundation and only get smarter crowding out human judgment. In this dynamic environment, it is relatively easy to understand what productivity gains are possible with AI, but significantly more complex to form an opinion on how AI will impact decisionmaking processes when the models do not provide sufficient explanation. Hence, the loud call for explainable AI models. In this context, Matt Shumer, general partner in Shumer Capital and the CEO and co-founder of OthersideAI, stated on social media: “[T]he gap between public perception and current reality is now enormous, and that gap is dangerous … because it’s preventing people from preparing.”
Several large asset owners, such as the California State Teachers Retirement System, publicly expressed their interest in AI and started to train their analysts and portfolio managers to think about the disruptive effect of AI on their portfolio companies. But this effort is easier said than done when developments move so fast. They are confronted with an enormous challenge, not only in terms of demand and growth—which, for example, are driving the profitability of data centers—but also in terms of competitiveness of business models.
Worth the Risk?
In line with Shumer’s statement, ignoring this challenge is very bad advice, and CalSTRS’ CIO, Scott Chan, explained the fund’s reasoning.
“While each market disruption is different, history tells us this is when diversification and dynamic allocation pay off—in the public and private markets,” Chan says. “We are in the early stages of the AI transformation, and there will be future winners and losers that are impossible to predict today, so a diversified portfolio that can adapt dynamically to opportunities and risks as the transformation evolves is the best way for us to manage these risks.”
It is clear that Chan falls back on a traditional investment belief: diversification. But this is not the same as moving to a passive approach. A passive investor will just buy the market, which implies that it will have both winners and losers in the portfolio—an unsatisfying thought. An active manager needs to think carefully about diversification in an AI world. Interestingly enough, AI models can help to group exposures with similar AI exposures and sensitivities!
But there is a ‘hidden’ message in Chan’s statement, too, when he says he wants to adapt dynamically to opportunities and risks; the illiquidity in private markets is not helping. Other investors than CalSTRS are starting to express that they, too, see a trade-off between the illiquidity premium in private markets and an agility option in the public markets. After all, it is easier to adjust a public market portfolio in case AI scenarios turn out differently than expected. Being stuck in an investment that is harmed by the AI developments worries investors and, as such, the private equity story does not stand on its own.
I would argue that for many investors, the value of the agility option currently exceeds the value of the illiquidity premium. This will be the case as long as investors do not fully understand the disruptive nature of AI. Think about this trade-off in the context of Total Portfolio Approach!
Eduard van Gelderen served as the head of research at FCLTGlobal in 2025 after spending more than six years as the CIO of PSP Investments in Montreal. Prior to his role at PSP, he worked for the investment office of the University of California and was CEO of APG Asset Management in the Netherlands. He recently launched Brave Foresight, an investment management consultancy company focusing on innovation and artificial intelligence.
This feature is to provide general information only, does not constitute legal or tax advice, and cannot be used or substituted for legal or tax advice. Any opinions of the author do not necessarily reflect the stance of CIO, ISS Stoxx or its affiliates.
Tags: Artificial Intelligence, Asset Allocation, Eduard van Gelderen, Private Markets, Scott Chan, Total Portfolio Approach