Mapping the Fusion Supply Chain

A few weeks ago the Fusion Industry Association put out its 2026 supply chain report, launched at the first-ever fusion supply chain trade show in New Mexico. The headline numbers: fusion companies spent $538 million with suppliers in 2025, up 24%, and project $681 million this year. Three quarters of surveyed suppliers invested in expanding fusion capacity in 2025, anywhere from $30k tooling upgrades to $65 million facilities.

That’s real money moving through a supply chain most people don’t know exists. So I tried to map it.

Click around. Each box is a link in the chain, and the ones with meters show how many fusion companies told the FIA that segment is a constraint.

What stands out

The bottlenecks aren’t where I expected. The exotic stuff (superconducting tape, tungsten) is scaling. What’s binding right now is almost boring: power electronics (48% of companies constrained), heat management (44%), vacuum vessels and pumps (32% each). Fusion is competing for switchgear and converters against grid upgrades, data centers, and EVs, and it’s the smallest customer at the table.

But the number that actually matters is the forward-looking one: 48% of fusion companies named fuel cycle systems as their biggest future concern. Which brings me to a conversation I keep thinking about.

Fuel is a business model

When I spoke with David Bryon at First Light Fusion, the thing that stuck with me was how far ahead they are on breeding, and what that’s worth. In February, First Light validated a tritium breeding ratio of 1.8 for their FLARE concept, the highest reported to date, using a liquid bath of plain natural lithium. At the design point, one plant would generate a 25 kg annual tritium surplus.

Here’s why that’s wild. The world’s civilian tritium inventory is about 20 kg, essentially all of it a byproduct of CANDU fission reactors (mostly in Canada), and it sells for $30,000 to $40,000 a gram. Run the naive math on a 25 kg surplus at today’s prices and one plant’s fuel byproduct is worth $750 million to $1 billion a year. Every other fusion company that reaches commercial scale needs startup tritium from somewhere.

Obviously the naive math is naive: if breeding works at scale, the price collapses. That’s the point. But in the window where most of the industry needs tritium and almost nobody has surplus, making fuel isn’t a subsystem. It’s its own business, the way enriched uranium became its own industry decades before anyone called it a supply chain.

The FIA report’s framing is that the “chicken and egg” gridlock between fusion companies and suppliers is finally easing. My takeaway is slightly different: the chain is filling in from both ends, with materials suppliers pivoting in from other industries and EPC firms circling from nuclear. The fuel cycle is the gap in the middle where a company could still build a moat.

Sources: FIA 2026 Fusion Industry Supply Chain Report · World Nuclear News coverage · First Light Fusion FLARE announcement · Science on the tritium supply problem