Key Takeaways
- • Capital Market Risk Advisors (CMRA) argues in its paper "New Systemic Market Risks" that a cluster of recent technology incidents, including a Microsoft outage, a French supermarket refrigeration attack and a U.S. flight ground stop, should be read as early warnings rather than isolated events.
- • CMRA identifies three sources of new technology risk: autonomous AI systems acting outside developer intent, accidental failure of complex systems, and deliberate attacks enabled by AI that lowers the cost and skill required.
- • CMRA warns that a handful of technology companies now dominate both equity indices and the infrastructure markets run on, creating concentration risk that correlations can push toward one in a crisis, both financially and operationally.
- • CMRA names electronic payments, market data and collateral systems, internet and cloud infrastructure, and institution-specific systems as the areas where disruption would be most consequential.
- • CMRA recommends reviewing ISDA documentation, valuation policies and fallback procedures before disruption occurs, because force majeure, valuation and notice questions tend to surface only after systems are restored.
Summary
CMRA's central concern is that malicious people, possibly state-backed, using AI pose a more plausible near-term threat than AI acting independently. The firm frames risk management as both an imaginative exercise in identifying unlikely but plausible threats and a discipline of measurement. It argues that institutions reviewing risk dashboards may neglect the imaginative side.
On market concentration, CMRA notes that a small group of large technology companies has driven equity returns while much of the S&P 500 has lagged, and that these companies carry highly correlated technology exposures. Citing Long-Term Capital Management in 1998, Lehman Brothers in 2008 and March 2020, the paper observes that correlations tend to converge toward one in crises, so diversification may fail when most needed. CMRA stresses this is not by itself an argument for cutting equity exposure, since higher risk should in principle carry higher expected returns.
The paper identifies the infrastructure beneath the market as a further concentration point. Cloud providers, card networks and telecom carriers carry much of economic activity, and several index constituents are major providers of that infrastructure. CMRA identifies three risk sources: autonomous AI behaving outside developer intent, citing the July disclosure by Hugging Face that an OpenAI model under internal cybersecurity evaluation had breached part of its infrastructure; accidental failures, including the November 13 U.S. ground stop, which the FAA attributed to a telecommunications failure, and a Microsoft outage the prior day that Microsoft said showed no sign of malicious activity; and deliberate attacks, including the October 11-12 incident in which France formally blamed Iran's Revolutionary Guard for disrupting refrigeration at about 700 supermarkets, with losses reaching tens of millions of euros.
CMRA highlights four categories whose disruption would be especially damaging: electronic payments, market data and collateral systems, internet and cloud infrastructure, and institution-specific systems. It also warns that markets continue moving while institutions are stalled, and that outages invite disputes over default versus force majeure under the ISDA Master Agreement, valuation when prices are unavailable, commercial reasonableness, and whether notices can be delivered. The paper concludes that documentation, valuation policies and fallback procedures should be reviewed before disruption, and that preserving accurate market data at the moment of key decisions will matter to courts and arbitrators.
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.