The Pulse
UK Pensions Regulator Flags Hidden AI Exposure in Portfolios
The UK Pensions Regulator has issued guidance asking pension trustees to assess how AI-related investments may concentrate risk across apparently diversified portfolios. It recommends whole-portfolio exposure checks, stress tests and scruti
AI.info Team · Written with AI · fact-checked ·
The UK Pensions Regulator is asking trustees to look beyond AI’s promise of productivity gains and examine how the technology could tie seemingly unrelated investments to the same market risk. In guidance published on 22 October 2026, the regulator warned that pension portfolios may be more exposed to an AI downturn than their mix of funds and asset classes suggests. It urged trustees to measure that exposure and test how their schemes would fare under stress—not to make a blanket retreat from AI-linked assets.
AI’s growth has become a portfolio question
The Pensions Regulator, or TPR, points to the scale of investment behind AI and the spread of AI-related companies across financial markets. Stanford’s 2026 AI Index says global corporate AI investment more than doubled in 2025, with private investment rising 127.5%. TPR says companies linked to AI also account for a large share of major stock-market indices and raise money through varied equity and debt channels across jurisdictions.
That breadth matters because AI-related exposure does not stop at technology firms. TPR’s trustee guidance names infrastructure, energy, real estate and financial assets as sectors that can become exposed to the same AI-driven expansion. If adoption disappoints, rules tighten or investment in AI infrastructure proves excessive, the regulator says losses could hit holdings across those sectors at the same time.
Market-cap funds can conceal overlapping bets
TPR highlights two ways a scheme can accumulate exposure without making an obvious, direct bet on AI. In passive market-capitalisation-weighted funds, the weight of AI-related companies can rise as their market values increase, particularly in portfolios with large growth-asset allocations. A scheme’s exposure can therefore grow even if its holdings appear spread across a broad index.
Active managers pose a different measurement challenge: several funds may hold the same companies or respond to the same market forces, while each manager assesses concentration within an individual fund. TPR warns that this can obscure repeated holdings and correlations across the scheme as a whole. The regulator’s concern is not that every AI-linked holding will fall together, but that conventional labels may give trustees too little visibility into shared risks.
TPR wants stress tests across the whole scheme
Ben Gunnee, TPR’s Executive Director of Market Oversight, said:
“That is why trustees should assess their AI-related exposure, test portfolio resilience and challenge advisers now to understand how market developments could affect their investments under different market conditions.”
Ben Gunnee, Executive Director of Market Oversight, The Pensions Regulator
The guidance asks trustees to identify direct and indirect AI-related exposure across all asset classes, then test a severe but plausible downturn in AI-related assets. Those tests should examine possible losses, liquidity demands and whether correlations change enough to push the scheme beyond its risk appetite. Trustees should also ask advisers how they measure concentration, what assumptions they use in stress tests and whether liquidity needs or leverage could amplify losses.
Defined-contribution schemes face a member-response risk
TPR’s advice extends beyond investment managers’ models. A sharp downturn could prompt members in defined-contribution schemes to make short-term decisions that crystallise losses or leave them out of a later recovery, the guidance says. Trustees should have a clear, proportionate communications plan ready to use if markets fall.
The regulator calls for action proportionate to each scheme’s circumstances; it does not prescribe a uniform allocation or a blanket sale of AI-related investments. Its immediate test for trustees is whether they can see the scheme’s combined exposure—and whether their diversification would still work if market relationships changed under pressure.