Nuclear fusion and nuclear fission share a name and a fundamental scientific category. Beyond that, according to those now building insurance frameworks around them, the similarities largely end.
Soheil Karimian, portfolio manager at Tokio Marine GX (TMGX) in London, has spent the past nine months developing NC Fusion, an insurance facility for nuclear fusion technology launched in October 2025 in partnership with Northcourt, a specialist managing general agent (MGA) within the Optio Group.
His starting point is that the insurance industry's instinct to reach for existing nuclear frameworks when assessing fusion risk is precisely the wrong approach.
The stakes are becoming harder to ignore. In March 2026, the UK government committed more than £2.5 billion to fusion over five years and called on insurers to provide proportionate cover outside traditional nuclear pools and exclusion clauses. In the US, regulators are developing a dedicated framework for fusion machines separate from nuclear reactors. Both moves signal that fusion is moving closer to commercial reality and insurers are expected to keep pace.
The distinction matters more than many in the insurance market appreciate. Both technologies are nuclear in the scientific sense, but their risk profiles diverge sharply. Fusion involves the handling of tritium, a radioactive isotope, but its waste has a maximum longevity of around 12 years, fundamentally different from the legacy contamination produced by fission, which extends well beyond a human lifetime. Fusion also avoids the meltdown risk associated with fission and does not produce long-lived radioactive waste.
Regulation has caught up with that reality. Both the UK and the US now classify fusion under health and safety legislation rather than nuclear acts, meaning the industry is managed as non-nuclear for regulatory purposes. For the insurance market, that classification is the critical starting point. "How that industry needs to be managed and how that risk is managed is non-nuclear," Karimian said. "That's the line that really matters."
The more immediate challenge for the insurance market is the absence of any historical loss data. Most lines of business, even relatively new ones, exist alongside comparable risks that provide at least some basis for modelling. Fusion does not.
"You're talking about taking something from lab into commercial scale and not really knowing what losses you're potentially going to be looking at," Karimian said. "There's no getting away from the fact that it requires the industry to model exposures in a different way."
In practice, this means underwriters must get close to clients at a very early stage, before commercial deployment and before meaningful loss data exists. Karimian described a process that begins with sitting down with clients' engineers to understand where risks are identified and how they are being mitigated, and then working outward from there to map the supply chain.
"The supply chain is really significant," he said. "Where is it? Are there any bottlenecks? Who are they working with? From an operational and business interruption perspective, that is critical. From a delayed startup for construction, that is critical."
The approach draws on risk engineering principles from high-hazard industries, downstream oil and gas in particular, where underwriters are accustomed to the nuances of technically complex risks and supply chain dependencies.
Karimian is candid about the fact that the coverage map for fusion is still being drawn. NC Fusion is designed to scale as the technology matures, spanning pre-commissioning, construction, and operational phases, but the specific products that fusion companies will need are still being determined through active dialogue with the industry.
"We are at such an early stage," he said. "We are talking to them about what coverages they actually need. Specific nuclear wordings are not appropriate, given the different regulatory classification. But the answers are going to be different depending on which company you are talking to and where they are in their development."
That variability reflects where the market is. A company still in the research phase has different requirements from one approaching construction or commissioning. The financial products that make sense at each stage, from performance bonds and surety to credit and liability cover, differ accordingly. TMGX's goal over the next 12 months is to develop products that fusion companies actually need and that the broader market will support, working openly with peers rather than in isolation.
For Karimian, fusion is a test case for how insurers approach emerging technologies more broadly. As new industries scale rapidly, he argues, carriers will increasingly need to build products before claims histories exist.
"Looking at something like fusion and being prepared to engage really early, it has already given us a blueprint for what other industries we can do that with," he said.
Karimian points to the rapid development of artificial intelligence as a reminder that technological change rarely follows predicted timelines. Fusion may still be years away, but he argues there is no reason insurers should wait to engage. "If you told me 12 months ago the level at which AI was going to be operating, I think we would have all just laughed," he said.
"If you wait until they have a solution and they say, by the way, we need insurance, you're just a supplier at that point," Karimian said. "If you want to be part of the solution, now is the time."
