On July 22, 2026, engineers at the ITER construction site in Cadarache, southern France, closed the final joint on the ninth and last sector of the ITER tokamak’s vacuum vessel — the steel chamber that will contain the fusion plasma once the reactor is operational. The completion was marked at a “KOREA-ITER Fusion Day” ceremony attended by roughly 150 people, and it closes out a manufacturing effort that ran across four of ITER’s seven member parties.
What the vacuum vessel is, and why finishing it matters
The vacuum vessel is the double-walled, D-shaped steel chamber at the core of the ITER tokamak. It holds the vacuum and the deuterium-tritium plasma the reactor is designed to heat past 150 million degrees Celsius, and it doubles as a primary confinement barrier around the fusion reaction. ITER’s vessel is built from nine wedge-shaped sectors, welded together into a single donut-shaped (toroidal) structure roughly 5,000 tonnes in total. Each finished sector — combining the raw steel forging with integrated thermal shields and superconducting magnet components — stands about 11.3 metres tall, 6.6 metres wide, and weighs approximately 400 tonnes on its own.
Completing sector 9 of 9 does not mean ITER is ready to operate — assembly of magnets, cryostat, and other systems around the vessel continues — but it removes one of the largest and longest-lead manufacturing dependencies from the critical path. The vacuum vessel has effectively been the pacing item for tokamak assembly for years, so its completion is a genuine, checkable milestone rather than a promotional one.
HD Hyundai’s role: one contractor, four sectors, two different national allocations
South Korea’s HD Hyundai Heavy Industries manufactured four of the nine vacuum vessel sectors, under two separate contracts. The company won the original contract for two sectors allocated to South Korea’s domestic ITER procurement share in 2010, then, in 2016, secured an additional contract to build two more sectors that had originally been allocated to the European Union’s procurement share. HD Hyundai delivered all four sectors by 2024, and the last of them was integrated into the completed vessel structure ahead of the July 2026 ceremony.
That reallocation detail is worth pausing on for a research-administration audience: ITER does not simply commission a single prime contractor to build the machine. Each of the seven members funds and manufactures an agreed share of components domestically, “in-kind,” through its own national domestic agency, and components move between countries and are welded together on-site into a single machine. When one member’s industry cannot meet a manufacturing schedule, ITER’s governance structure allows a procurement package to be reassigned to a different member’s domestic agency and industry — here, from the EU’s package to Korea’s, executed by the same contractor already qualified on the design. That is a live example of how a multinational, in-kind procurement model actually functions in practice, not just in treaty text.
The governance and export-control angle
ITER is not a company or a single-country national laboratory. It is a treaty-based intergovernmental organization with seven members: the European Union (via Euratom), the United States, Japan, China, Russia, India, and South Korea — a membership list that includes NATO allies, the EU, and both China and Russia within a single formal collaboration. Each member contributes the large majority of its share “in-kind” rather than in cash: hardware manufactured domestically to ITER’s technical specifications, then shipped to Cadarache and integrated by the ITER Organization.
For CASRAI’s audience — research administrators, research security officers, and international-collaboration managers — that structure raises the exact questions that sit at the center of research security and export-control practice:
- Technology transfer across export-control regimes. Fusion hardware, superconducting magnet technology, and precision manufacturing techniques can intersect with dual-use and nuclear-adjacent export-control frameworks (in the US, for example, the Export Administration Regulations and International Traffic in Arms Regulations touch technologies of this kind). A treaty organization whose members are simultaneously allies and, in other bilateral relationships, subject to restrictive export controls toward each other has to resolve technology-sharing and IP terms at the organizational level, through the ITER Agreement itself, rather than leaving it to be negotiated bilaterally between individual member states component by component.
- IP and design ownership in a shared-procurement model. When a component originally allocated to one member (the EU) is reassigned to and manufactured by another member’s industry (South Korea), the underlying design data, specifications, and resulting intellectual property have to be governed by pre-agreed organizational rules, not renegotiated per shipment. This is the kind of consortium-agreement and IP-allocation question that shows up, at smaller scale, in any multinational research grant or collaboration agreement.
- Domestic agencies as the compliance interface. Each ITER member operates its own “domestic agency” (Fusion for Energy for the EU, US ITER for the United States, and equivalents for the other five members) that sits between the member government and the ITER Organization. For research-security and compliance staff, that domestic-agency layer is the practical model worth studying: it is how a seven-party treaty organization with genuinely mixed export-control relationships among its members manages technology flow without requiring every bilateral pair of members to negotiate controls directly.
None of this is unique to fusion research. It is a large-scale, unusually visible instance of a governance problem every institution running an international research collaboration has to solve at some scale: how do you move technology, data, and components across a group of partners who are not all bound by the same export-control relationships to one another, without breaking the underlying regulatory obligations any single member owes its own government.
What comes next
With the vacuum vessel sectors complete and integrated, ITER’s remaining assembly work centers on the surrounding magnet and cryostat systems and the broader machine-assembly sequence before the project’s planned move toward first plasma operations. HD Hyundai and its domestic-agency counterparts described the milestone as a demonstration of the manufacturing quality control needed to build fusion-scale components to ITER’s tolerances — a capability with applications well beyond this one project, in future fusion-energy manufacturing supply chains.
