|
Matthias Knab, Opalesque for New Managers: Ask more than 500 asset owners what worries them most about artificial intelligence and they will tell you, in order: that valuations and capital expenditure are compounding on each other, that the market has become dangerously concentrated, and that everyone now depends on a handful of mega technology providers.
Then ask them what they did with their money last year. They bought more of it.
Morningstar published the fifth edition of its Asset Owner Perspectives Survey on 16 September, covering more than 500 institutions across 11 countries in North America, Europe and Asia-Pacific, with over $20 trillion between them. The respondents are pension funds, insurance general accounts, outsourced CIOs and family offices. What the survey documents, more clearly than its headline suggests, is a group of investors who have identified their single largest portfolio risk with some precision and increased their exposure to it anyway.
The American reversal
The swing in US allocations is the cleanest number in the study. Last year, 23% of asset owners increased their US exposure while 32% cut it. This year those figures have essentially inverted: 31% increased, 24% decreased.
This happened in spite of the stated concerns rather than in the absence of them. Fifty-one percent cited concerns about the Trump administration, 34% regulatory uncertainty and 30% global conflict. They bought anyway, and Morningstar attributes the decision to market resilience, AI-driven growth and benchmark weightings.
That third item deserves more attention than it usually receives. Benchmark weightings are not a view. They are the absence of one. An institution that increases US exposure because the index told it to has not made a judgment about American assets, it has declined to make a judgment about the index. Given that the same institutions rank market concentration as their second-biggest AI concern, the mechanism by which they became more concentrated is worth naming out loud.
Three worries that are really one worry
Asked to rank the most material global investment issues, respondents put rising inflation first at 76%, the evolving generative AI landscape second at 71%, and energy and supply chain crises third at 66%.
These are not three separate problems, and the survey shows that asset owners know it. Nearly six in ten said AI-driven energy demand could push up energy costs and inflation. Read together, the top three concerns describe a single feedback loop: AI buildout drives energy demand, energy demand drives costs, costs drive inflation. AI-related environmental concerns more than doubled year on year, from 12% to 25%, which is the fastest-moving number in the entire study.
So the position is not that asset owners are unaware. They have mapped the chain of causation. They have simply concluded, or been structurally obliged to conclude, that being underweight is the greater risk.
The diversifier that may not diversify
Private markets are the other major finding. Asset owners expect their allocation to rise from 19% of AUM today to 23% within five years. The most ambitious are in Europe, targeting 25%, against 23% in Asia-Pacific and 21% in North America. That European lead is itself notable, given how often European institutions are characterised as the cautious end of the market.
The reasons given repay a second look. The leading motivation, at 56%, is diversification versus public markets. Higher expected returns follow at 42%. Third, at 29%, is access to specific market themes, and the themes Morningstar names are AI electrification and data transmission.
Which is to say that a meaningful slice of the private markets allocation is not a diversifier at all. It is the same AI trade the respondents have just flagged as their concentration risk, bought through a less liquid wrapper. Whether that is diversification or duplication depends entirely on what the public book already owns, and it is not a question the survey asks.
The obstacles, meanwhile, have not moved. Liquidity risk is cited by 63%, transparency issues by 43% and limited data availability by 28%. These are the same three problems the industry has been discussing for a decade. Allocations are rising regardless.
The number that fell off a cliff
The most striking single data point in the survey is buried in the sustainability section, and it is a governance number.
The share of asset owners regarding business ethics as financially material fell 24 percentage points in one year, from 68% to 44%. Morningstar records this as the largest year-over-year drop in the study. Governance materiality overall fell from 54% to 45%, and social from 49% to 39%. Environmental factors held broadly steady, with 56% saying they have become more material.
Set that against the same survey's finding that governance concerns among leading AI companies are part of what asset owners are weighing. Governance at AI firms is a worry. Business ethics as a materiality factor has collapsed. Both statements come from the same respondents in the same questionnaire.
The generous reading is that the drop reflects headline and policy risk rather than conviction, which is Morningstar's own suggestion, and that institutions have simply stopped saying out loud what they continue to price. The less generous reading is that a category which was never firmly anchored in financial analysis has proved easy to put down when the politics turned. Either way, a 24-point move in twelve months tells you the number was soft to begin with.
The regional picture complicates the usual narrative. ESG integration above 75% of AUM sits at 18% globally, down two points, but still well above the 12% recorded in 2024. Underneath that, North America rose from 37% to 43% while Europe fell from 48% to 43%. The two regions have converged precisely, moving in opposite directions to get there.
Nobody is waiting for the regulator
Confidence in regulation has been in freefall for four years. Net positive views fell from 55% to 39% this year. Just 4% of asset owners now regard regulation as a major help, against 28% in 2022.
Data providers were the only group to gain ground, with 34% now looking to them to improve ESG data, up from 29%. They remain behind regulators at 38%, but the gap has narrowed by 16 points in a year. Readers should note that the survey was conducted by Morningstar Indexes and Morningstar Sustainalytics, which is to say that a data provider has found growing demand for data providers. The finding is plausible and the direction of travel is consistent with everything else in the study, but it is not a disinterested observation.
On stewardship, direct engagement remains the primary method of active ownership, but the share treating it as a top priority dropped from 39% to 30%. Collective engagement rose from 16% to 20% and public policy engagement from 16% to 19%. Institutions appear to be concluding that the problems they care about are systemic, and that meeting companies one at a time is not the efficient route to addressing them.
What it adds up to
Lindsey Stewart, Morningstar's director of institutional insights, describes asset owners as "navigating a market increasingly shaped by artificial intelligence", weighing innovation against valuation, concentration and governance.
Navigating is the polite word. What the survey actually shows is a large group of sophisticated institutions that can describe the risk in detail, can trace its transmission into inflation and energy costs, can name the handful of companies on which the whole structure rests, and are increasing their exposure to all of it, partly because it is working and partly because the benchmark leaves them little choice.
Another curious finding: Asset owners are nearly twice as likely to apply AI bottom-up, at 38%, as top-down, at 20%, and they use it mostly for internal workflow and processing rather than to generate investment insight. The institutions with the largest exposure to the AI trade are, for the most part, not yet using AI to analyse it.
|
RSS
