TPR highlights hidden AI concentration risk in pension fund portfolios

On 22 October 2026, The Pensions Regulator (TPR) published new guidance for trustees on identifying and managing risks in their pension funds. The new risks arise from the increasing concentration of artificial intelligence (AI)-related assets in investment markets. AI-related firms account for a growing share of major equity markets, with many relying on debt financing channels to fund large investments. TPR warns that apparently diversified pension portfolios may contain hidden exposure to common underlying AI-related risks.

TPR expects schemes “to maintain high standards of investment governance when assessing asset allocation”. It asks trustees to consider their portfolio as a whole, rather than reviewing individual funds or asset classes in isolation, and to take proportionate action reflecting their scheme’s circumstances.

Why concentration may be difficult to see

AI-related exposures are spread across a growing range of sectors and asset classes. As a result, exposure is not limited to purely technology-focused investments but can arise in different parts of a portfolio that appear unconnected. TPR's concern is that trustees may underestimate the extent to which apparently diversified portfolios are exposed to common AI-related risks.

If AI adoption disappoints, regulation tightens, or investment in AI infrastructure proves excessive, losses could arise simultaneously across holdings and asset classes that appear unrelated. Investors may therefore find that their portfolios are less diversified than they believed.

TPR also highlights similar concerns identified in the Bank of England's July 2026 Financial Stability Report about increasing connections across the AI supply chain but also the material challenges of understanding the full extent of AI-related exposures. Together, these factors could make investment risks more difficult to assess, particularly during periods of market stress.

TPR notes that concentration risk can arise in both passive and active investment arrangements, meaning that exposure may not always be apparent when portfolios are viewed through individual funds or mandates.

TPR expects trustees to:

• Assess the scheme’s total direct and indirect AI-related exposure across asset classes, including exposure within default arrangements
• Use scenario analysis to test the portfolio under a severe but plausible downturn in AI-related assets
• Consider how a stressed market scenario could affect the scheme's risk profile, including potential losses, liquidity pressures and changing asset correlation
• Review whether diversification would remain effective under stressed market conditions
• Challenge investment advisers and managers on how they identify, measure and manage concentration risk, including whether liquidity or leverage could amplify losses.

Communicating with members

TPR also highlights the potential for members of defined contribution arrangements to make short-term decisions during a material market downturn that may adversely affect retirement outcomes. It says trustees should have a proportionate, clear communication plan that can be implemented quickly and seek advice to ensure their communications are appropriate.

TPR also refers trustees to its earlier guidance, Market oversight: Market volatility and what trustees should do, which sets out its broader expectations during periods of market volatility.