Fusion startups are partnering with defense companies, too

Tomorrow’s utilities are not the only potential customers for fusion companies. RTX has invested in laser developer Xcimer, while General Atomics has contracted Tokamak Energy to develop superconducting magnets. Before selling electricity, these companies are taking their technology into another market. How much can the resulting capital and manufacturing experience contribute to building a power plant?

What Xcimer and RTX announced in September 2026 was an investment and an agreement to explore technical cooperation. RTX Ventures, the group’s venture capital arm, took a minority stake in Xcimer. The companies also signed an agreement to explore cooperation in the defense sector. The individual investment amount was not disclosed. The companies’ announcement

Tokamak Energy’s relationship with General Atomics, announced in October 2025, is different: a contract to develop high-temperature superconducting magnets. The UK company is participating in a defense research program as a magnet technology supplier. Its customer is General Atomics; Tokamak Energy itself does not have a direct contract with the US government on this program. Tokamak Energy’s announcement

Grouping both deals under “defense money flowing into fusion” obscures that distinction. An equity investment buys a stake in a company. A development contract pays for work a customer wants done. The former is not sales revenue, and announcing the latter does not establish that a finished product has been delivered.

Fusion development builds more than the equipment needed to produce a reaction. It also develops the ability to manufacture components, assemble them and test them repeatedly. Lasers and magnets are enabling technologies that may allow companies to offer those capabilities to other customers.

What these companies aim to supply is not a completed fusion power plant, but laser and magnet technology developed along the way. Signing a contract and completing and delivering a product are separate stages. These partnership announcements do not demonstrate that fusion electricity generation has been achieved, either.

In a September open letter, Xcimer’s founders described a strategy of developing energy and defense businesses together: building manufacturing capacity, supply chains and talent while generating revenue before the first power plant starts operating. They argue that the engineering work in the two fields has substantial overlap. The founders’ letter

The interesting part is the attempt to secure work for customers without waiting for electricity sales. Alongside raising the next round of development capital, the company seeks to gain manufacturing and delivery experience by fulfilling orders. This remains a stated strategy, however. It is not evidence that the defense business already generates sufficient profits to fund a power plant.

Separately from RTX’s investment, Xcimer also receives support through the US Department of Energy. In October 2026, it announced an agreement for up to $30.5 million in additional funding under DOE’s Milestone-Based Fusion Development Program.

Payments depend on completing pre-agreed technical and commercialization milestones and having them independently verified. The announced ceiling should not be treated as cash received in full on the announcement date. Xcimer’s DOE funding announcement

Equity investment, customer-funded development and milestone-based public payments can all support a company. But each counterparty expects something different, and each source has different conditions for payment. Adding the headline amounts together does not reveal how much money a company can spend freely or which commitments it must prioritize.

Following investment announcements alone says little about what customers actually want to buy. Development contracts make it clearer which technologies a fusion company is trying to turn into a business.

Does winning customers in another market necessarily accelerate fusion development?

If manufacturing facilities and testing experience can be shared, work for other customers may help build capabilities that would otherwise have to be established solely for a power plant. This is the kind of synergy Xcimer’s founders describe.

Customers also have their own deadlines and requirements. If fusion development and another market’s product development demand different work, companies must balance the use of people and equipment. That is a question for evaluating diversification, not a claim that either company is experiencing internal problems. Partnership announcements alone do not disclose staffing allocations, profit margins or how much money flows back into fusion research.

A prominent partner’s name is therefore no guarantee of being closer to success. What matters next is how the contracted work progresses, and how its manufacturing experience and revenue connect to fusion development.

The first thing a fusion company sells may not be electricity. It may sell components and development work while building the organization that can eventually deliver a power plant. To judge whether that route truly advances fusion, we need to follow delivery records alongside progress on the power plant itself.