Key Takeaways
- • Consultants are using scenario testing to model how AI could fail to meet its growth potential, prompted by an Oct. 22 guidance from The Pensions Regulator warning U.K. pension trustees of hidden AI risks in their portfolios.
- • Meketa Investment Group's September whitepaper "The AI Boom," co-authored by Alison Adams and Frank Benham, outlines four scenarios for how the AI boom could end: a financial break, a rotation, a transition, and a demand disappointment.
- • WTW, Marsh, and Ortec Finance say AI exposure extends well beyond equity concentration, including capex, power and data center demand, semiconductor supply chains, and credit exposure to companies financing AI infrastructure.
- • Marsh's Andrew McDougall argues the drawdown from an AI sell-off could exceed typical historical stress scenarios because of concentration, and PIMCO's Libby Cantrill flags state-level data center restrictions as a possible near-term speed bump.
Summary
Meketa Investment Group's September whitepaper, "The AI Boom," co-authored by managing principal Alison Adams and director of research Frank Benham, examines four ways the AI boom could end: a financial break in which the financing structure fails before the assets it funded generate sufficient returns; a rotation in which capital expenditure is justified in aggregate but value migrates away from established holdings; a transition in which a financing break leads into a longer phase of falling costs and widespread adoption; and a demand disappointment in which usage fails to keep pace with capacity without a credit event. The paper concludes that concentration deserves examination as a risk distinct from any view on whether AI proves transformative. Adams said the firm is developing scenario analysis and historical stress testing around technological revolutions and is tracking exposure through a client-facing AI dashboard.
WTW's Alasdair MacDonald, CIO of its U.K. Investment Advisory business, described AI as a systemic risk affecting all assets. WTW monitors actual index exposure, though exposure shifts over time, including as large IPOs such as SpaceX enter the market, and also tracks thematic indicators such as productivity gains and potential setbacks. He said monitoring requires comparing the actual impact of AI to date with the impact needed to justify current valuations. Ortec Finance's Patrick Tuijp said AI risk extends beyond direct market exposure to second-order exposures, including infrastructure capex, power and data center demand, semiconductor supply chains, credit exposure, and disruption to businesses whose models AI may challenge.
Marsh's Andrew McDougall said AI "cuts across asset classes" and that the firm is prototyping an AI-enabled thematic mapping process to assess public and private assets by revenue sources and theme sensitivity. He warned that given current concentrations, the drawdown from an AI sell-off could be larger than in a typical historical scenario. Marsh's paper "AI: Investing across the value chain" by Max Becker and Dmitri Smolansky argues that AI value is distributed across the ecosystem, though diversification is not a substitute for managing overall exposure. Tuijp said AI is well suited to scenario analysis and that downside cases should be treated as potential broader implications for equities, credit, and the macroeconomy rather than short-term shocks to technology stocks.
PIMCO's head of public policy Libby Cantrill noted in her "Washington Watch" note that more than 40 ballot initiatives restricting data centers have succeeded and that several newly elected Democratic governors campaigned on statewide moratoriums. She said that would likely mean more restrictions and higher costs for hyperscalers, and that a slower buildout could pressure growth, inflation, and rates.
Go deeper than the headline
Research every firm in this story, track the ones that matter and get the daily briefing — free for 7 days.