Global insurtech funding reached US$2.44 billion in the second quarter of 2026, the highest quarterly total since Q2 2022. The result, reported by Gallagher Re, extends a two-quarter rebound from the sector's 2021 peak of US$15.8 billion.
AI-focused companies attracted 99.1% of all insurtech funding during Q2, raising US$2.42 billion across 95 deals. Every funding round above US$5 million in the quarter went to an AI-focused company.
Mega-round activity drove the recovery. Deals of US$100 million or more accounted for US$1.67 billion, or 68.4% of the total, the strongest large-scale fundraising quarter since Q4 2021. Earth imagery company ICEYE alone raised US$520 million in a Series F, taking up roughly a fifth of the quarter's total funding.
The concentration of capital is becoming more pronounced. Deal count reached 107 for the quarter, the highest since Q1 2024, yet early-stage funding fell 51.8% quarter-on-quarter, from US$548.5 million to US$264.2 million.
The Q2 AI funding share of 99.1% has risen from 95.2% in Q1 2026, indicating the concentration is accelerating. If early-stage funding continues to thin while mega-rounds dominate, the pipeline of new insurtech options available to insurers and reinsurers is narrowing.
Andrew Johnston, global head of insurtech at Gallagher Re, said the pattern reveals a paradox at the heart of the current funding cycle. "Capital availability is clearly not a problem. And yet we are seeing something of a paradox: at a time when AI is supposed to be making things cheaper, individual insurtechs seem to be raising, then burning through, more cash than ever," Johnston said.
Insurance and reinsurance companies participated in 27 technology investments during the quarter, down from 32 in Q1 2026. The overall funding recovery was driven by venture capital and private equity rather than insurer-backed capital.
Alongside its funding analysis, the Q2 report examines the insurance implications of the data centre buildout underpinning the AI economy. For the insurance and reinsurance markets, the scale of this infrastructure creates both an opportunity and a structural challenge.
Swiss Re Institute projects global data centre insurance premiums to rise from US$10.6 billion currently to US$24.2 billion by 2030. Individual AI-optimised facilities can reach US$20 billion or more in construction cost alone. Financing institutions are pressing for coverage at full replacement value, yet the insurance and reinsurance markets can support only a fraction of that limit at competitive pricing.
The risk profile of these assets stretches across more than 20 lines of business and across every stage of the project lifecycle. Coverage obligations begin before construction, with environmental liability and title insurance. They extend through construction all-risks, marine cargo for high-value chips and servers, and commissioning exposures.
In live operations, the dominant risks shift to business interruption, service level agreement breaches, cyber and power supply failures. Standard property policies typically require physical damage to trigger a claim. Power failure, cooling failure, and software misconfiguration are the most common causes of outages, yet none produce physical loss under conventional policy language.
Non-damage business interruption extensions address this gap but are not yet standard market offerings. Gallagher Re describes them as one of the most commercially material product development opportunities in data centre insurance.
Swiss Re found that more than a quarter of US data centre capacity sits in areas with elevated hail exposure. Over 40% falls within zones with material tornado risk. A single hyperscale campus can concentrate tens of billions of dollars of insurable value in one location.
Johnston said the scale of the opportunity is unlike anything the industry has produced in recent years. "The AI boom is creating one of the largest new pools of insurable assets the industry has seen in decades. Every major data centre being built today will require insurance during construction and throughout a multi-decade operational life," he said.
