Interconnected risks are the focus of a new report from Swiss Re Institute and the London School of Economics (LSE), titled “The age of interconnected risks.” In the report, Swiss Re Institute and LSE state that The average number of risk interconnections identified by companies has risen by 24%, with AI emerging as a more strongly connected node. For insurers and reinsurers, the report also sets out what such links mean for concentration of exposure.
What the Company Risk Disclosures Reveal About Linked Risks and AI
The finding rests on a defined sample. Describing its method, Swiss Re Institute and LSE write: “we analyse risk disclosures in the 10-K filings of 91 Fortune-100 companies with multinational operations.” The report treats those filings as the record of what companies themselves identify as connected risks.
The headline figure and the AI finding come from the same passage of the report. Swiss Re Institute and LSE say the average number of risk interconnections identified by companies has risen by 24%, and they describe AI as a more strongly connected node within that network of links.
The report also points to the number of companies flagging individual themes. According to Swiss Re Institute and LSE, The number of corporates highlighting AI, climate and social instability risks has doubled or even tripled since 2019. The sentence groups AI with climate and social instability, and it is attributed in the report to corporate disclosures rather than to the authors’ own scenario work.
The authors also look beyond single companies to the systems around them. They write that The scope for more costly systemic crises has likely risen as economies have become more dependent across multiple domains beyond the financial system, such as the real economy, critical infrastructure and technology. The language is hedged in the report itself, which says the scope has “likely” risen.
Supply Chains and Digital Concentration Raise Accumulation Questions for Re/insurers
Supply chains get specific attention in the document. Swiss Re Institute and LSE call them another critical nexus and say that (geo)political, tariff, climate, pandemic and cyber risks increasingly interact and reinforce one another, creating multiple pathways through which supply chains can be disrupted.
The report then moves to concentration in digital and financial infrastructure. It states that In 2024, three firms controlled 70% of the world’s cloud infrastructure. On payments, it says that With three companies processing 97% of global credit card transactions, a payment system outage would likely disrupt the daily flow of financial exchanges.
The insurance reading of those concentrations is spelled out directly. Swiss Re Institute and LSE write: “For re/insurers, such concentration creates accumulation risk, as a single outage or cyber event can trigger simultaneous claims across otherwise unrelated insureds and multiple lines of business.”
Readers following the same theme elsewhere may want to look at InsuraBeat’s coverage of the IAIS mid-year global monitoring exercise on systemic risk and of Munich Re’s plan to acquire a cyber insurtech.
The report also touches on financial-sector structure. It says that Non-bank financial intermediaries (NBFIs) account for more than 50% of global financial assets. On AI financing, Swiss Re Institute and LSE note that AI-related deals accounted for more than one third of private credit (PC) deals, according to the report.
Exposed Locations for US Data Centres and Taiwan’s Chip Plants
The report ties AI to physical assets and to the hazards around them. Swiss Re Institute and LSE state that More than a quarter of US data centres are in areas with at least three large-hail days a year, and more than 40% are in zones of significant to very high tornado risk.
On the value at stake, the report says that Individual AI data centre campuses can exceed USD 10bn in replacement value and reach as much as USD 50bn in exceptional cases. For semiconductors, it adds that Replacement values for a large semiconductor plant can reach USD 20‒30bn.
Location is the other half of the semiconductor point. According to Swiss Re Institute and LSE, In Taiwan, about 88% of semiconductor plants are in areas of extreme seismic risk. Taken together, the passages describe high replacement values sitting in hazard-prone areas, which is the combination the report links to accumulation.
Natural catastrophe losses supply the wider backdrop. Swiss Re Institute estimates global insured natural catastrophe losses are rising by 5‒7% annually, driven by exposure growth, urbanisation, and intensifying climate hazards, the report states. It adds that Exposure growth alone accounts for more than 80% of weather-related insured loss growth globally.
Further InsuraBeat reporting on climate and capital includes EIOPA’s description of heatwaves as a climate hazard for Europe’s insurers and the Geneva Association’s modelling of fragmentation and reinsurance capital.
Report Sets Out What Insurance Can and Cannot Absorb
On resilience, Swiss Re Institute and LSE are direct about the role of cover. They write that greater insurance penetration provides an important source of systemic resilience. Elsewhere, the report cautions on capacity: rising frequency of secondary perils and greater loss accumulation risks may make capital more selective in the places where exposure is growing fastest.
The report draws a line around what the private market can carry. In its words, Systemic risks – such as pandemics, extreme cyber accumulation and wider critical infrastructure failures – by definition exceed the private-sector risk-bearing capacity. It adds that Preserving the capacity to transfer risk will remain as important as reducing risk itself.
The role it assigns to carriers is specific. Swiss Re Institute and LSE say that Re/insurance can help identify and price critical concentrations, incentivise investment in resilience and diversify residual risks, while also signalling where systemic exposures may require public-private solutions.
The report also calls on supervisors. It says that supervisory stress testing frameworks should incorporate cross-domain scenarios, involving AI together with market synchronisation, critical infrastructure failures and cyber incidents. On the policy backdrop, it adds that Limited fiscal space and (geo)political fragmentation can weaken international crisis responses.
For regulators’ views on technology dependence and AI, InsuraBeat has also covered the FSB’s publication of responses on sound AI adoption. The full text of the Swiss Re Institute and LSE report is available from Swiss Re Institute.
