Non-peak perils cross US$100bn - and the reinsurance maths no longer works

Munich Re's Monte Carlo data shows secondary perils hitting a scale that high attachment points cannot fully contain

Non-peak perils cross US$100bn - and the reinsurance maths no longer works

Reinsurance News

By Mark Rosanes

In 2025, for the first time, insured losses from non-peak perils crossed US$100 billion in a single year. Munich Re put the figure at US$104 billion at the Monte Carlo Rendez-Vous. That is not just a record. It signals a permanent shift in the scale of frequency-driven losses.

Non-peak perils, including hailstorms, floods, severe convective storms, and wildfires, were once treated as attritional risks whose individual events were manageable. The category crossed a threshold that previously required a major hurricane or earthquake to reach. Global insured losses also passed the US$100 billion mark for the sixth consecutive year, even without a US hurricane landfall in 2025.

The California wildfires in January 2025 were the single largest event. Economic losses reached US$54 billion, the greatest ever recorded for that category. Europe also saw wildfire activity near major cities including Bordeaux, Marseille, Madrid, and areas of Sicily, though urban centres were largely spared.

Attachment points under pressure

The reinsurance response to the secondary peril surge of 2022 and 2023 was to raise attachment points sharply, keeping most non-peak losses with primary insurers. Even so, secondary perils accounted for 92% of global insured catastrophe losses in 2025, according to Swiss Re Institute, with total insured losses across the market still reaching US$107 billion.

The challenge runs deeper in the current market. Property catastrophe rates fell around 16% globally following mid-year 2026 renewals, the steepest annual decline since the late 1990s. Business above attachment points is being priced lower. That transfers secondary peril frequency risk to primary insurers without addressing the aggregate exposure that accumulates across a full season.

The secondary peril label itself is under scrutiny in treaty and modelling frameworks calibrated to an older loss environment. Thomas Blunck, a member of Munich Re's board of management, pointed to the sector's broadening role. "The value of reinsurance has never been more evident than it is today. A resilient reinsurance sector is capable of absorbing increasingly complex and globally interconnected risks," he said.

Cyber: a widening coverage gap

The protection gap in cyber runs differently from natural catastrophe risk, but the underlying problem is the same: exposure is growing faster than coverage. Munich Re's own research found that 89% of companies feel inadequately protected against cyber attacks, a figure drawn from its own survey data.

Swiss Re has framed a sharper problem: whether cover already in place is enough for the losses businesses need it to absorb.

AI is accelerating the pressure on both sides. One in four cyber breaches were AI-enabled in the first half of 2026, according to Swiss Re data. The situation drove up both claims frequency and severity. Contract standards for AI-related risk lag even further behind cyber, where modelling frameworks and coverage definitions have at least begun to form.

Heat adds a further dimension, Munich Re reported. Heatwaves are causing growing losses to agriculture, infrastructure, and supply chains. Attribution remains harder than for storms or floods, but the economic damage is accumulating across multiple lines.

Capital is not the binding constraint

Munich Re's Stefan Golling identified the real challenge. "Volatility is not a temporary phenomenon. Our mission is to pool our expertise, capacity and innovative strength to help our clients remain resilient, adapt successfully to change and navigate the new risk landscape with confidence," he said.

Reinsurance capital has grown at around 5.8% per annum over the past eight years, according to Munich Re, and the sector entered 2026 with ample capacity. The binding constraint is not how much capital exists. The question is whether the modelling and attachment structures are calibrated to a world where US$100 billion from non-peak perils is the baseline.

Related Stories

Keep up with the latest news and events

Join our mailing list, it’s free!