Panel gaps are one of those details you cannot unsee. Once somebody points out that the seam between a hood and a fender runs wider on one side than the other, you will be checking it on every car you walk past for a week.<\/p>\n
The tolerance people argue about there is two or three millimeters, on a panel you can lift with one hand.<\/p>\n
Now run the same argument on a wedge of stainless steel weighing about 1,100 metric tons. There are nine of them, and they have to close into a circle. That is the job at a site in Provence, and on Monday, July 27, the sixth of those nine was scheduled to start its trip into the pit.<\/p>\n
The machine is ITER, the fusion reactor that seven members are building near Cadarache in southern France. Its plasma chamber, the vacuum vessel, is not cast in one piece. It arrives as nine wedges that get stacked into a hole in the ground and welded shut later.<\/p>\n
Five are already down there. The one moving Monday is the last sector Korea built, and ITER marked the occasion<\/a> with a ceremony at the Assembly Hall doors on July 22.<\/p>\n The thing being lifted is not one component. ITER calls it a sector module, and it covers a 40-degree slice of the finished chamber.<\/p>\n At its core sits the vacuum vessel sector itself, a double-walled stainless steel shell weighing 440 tons. Wrapped around that is a thin thermal shield. Bolted to either side are two D-shaped superconducting toroidal field coils at roughly 360 tons apiece.<\/p>\n Add the rigging and the load hanging off the cranes has run between 1,300 and 1,400 tons on recent lifts. ITER’s assembly hall has two 750-ton overhead cranes that work as a pair, rated together at 1,500 tons, so the margin is thinner than you would like for something you cannot drop.<\/p>\n The way the weight gets shared is worth a look. Four crane hooks carry the module, each one reading around 350 tons on the operator’s screen, tied into a single lift point by a yellow steel beam called the Dual Crane Heavy Lifting Beam. That beam is what lets a load the size of a small apartment block rotate in mid-air.<\/p>\n Anyone who followed Britain sliding a 500-ton reactor vessel into Hinkley Point C with 40 millimeters of room<\/a> already knows the genre. This is the same sport with more pieces and no second attempt.<\/p>\n Here is the part that actually improved, and it improved because people got better at it rather than because somebody bought a new tool.<\/p>\n In June 2025, when the second module was set down beside the first, the crews left an offset of 100 millimeters between them. That gap got closed afterward, but it was not where they wanted to be.<\/p>\n For the third module, in November 2025, the target was 50 millimeters. Trouble with the metrology systems then handed the team an opening to push further, and ITER says they landed it with roughly 10 millimeters of offset<\/a> by leaning heavily on human observation rather than instruments.<\/p>\n Ten millimeters. On an object that weighs as much as a small warship, guided in the last stretch by people watching.<\/p>\n The January 2026 lift went tighter still, in a different direction. As that module rose, the clearance between the top of the load and the support beam above it shrank to 0.4 millimeters. Mathieu Demeyere, the ITER construction manager who has supervised these operations, said afterward the crews now know they “can adjust to extremely narrow margins.”<\/a><\/p>\n They are going to need it. The pit gets more crowded with every module, and later this year one of them has to be lowered between<\/em> two sectors already in place for the first time.<\/p>\n ITER’s reputation is delay. That reputation was earned, and nothing here erases it.<\/p>\n But the assembly line itself has been speeding up. The consortium doing the work took 7.4 months to turn its first sector into a finished module and get it in the ground. That figure is now down to 5.5 months, mostly by running tasks in parallel instead of one after another.<\/p>\n The knock-on effect is the part worth noticing. ITER has moved the expected date for the final sector module going into the pit from December 2027 to June 2027<\/a>. Jens Reich, the organization’s machine assembly program manager, described the gain as six months that “create flexibility for the overall baseline,” a cushion for whatever goes wrong next.<\/p>\n Reich had already put a marker down back in November. If the pace held, he said, “we will complete sector installation on schedule in 2027.”<\/p>\n The headcount tells the same story from another angle. The industrial consortium running assembly has gone from 90 people on site to 220, and picked up responsibility for module sub-assembly, in-pit assembly and pre-weld positioning along the way.<\/p>\n None of that changes the machine’s calendar. Under the baseline ITER adopted in 2024<\/a>, research operations start in 2034, full magnetic energy comes in 2036, and the reactor does not run on deuterium-tritium fuel until 2039. That last one is four years later than the previous plan, and it came with about 5 billion euros in extra cost.<\/p>\n The sector modules sit on what project managers call the critical path, which is a polite way of saying everything else is waiting on them.<\/p>\n Once all nine are in the pit, Westinghouse takes over. The company signed a $180 million contract with ITER in June 2025<\/a> to weld the nine sectors into one continuous torus, all of them simultaneously. Westinghouse calls it the most intensive stage of the assembly, and it cannot begin a day early.<\/p>\n There is also a 1,000-ton magnet sitting in the same building with nothing to do. The central solenoid, built in California and topped out in June when the last 110-ton coil went onto the stack<\/a>, is the piece that fires the plasma current. It cannot be moved into the middle of the reactor until the nine sectors are standing in a ring around it.<\/p>\n So the wedge going in this week is not a photo opportunity with a big crane. It is the item unlocking the next two jobs on the list.<\/p>\n Korea manufactured four of the nine sectors, and HD Hyundai Heavy Industries built them. Europe supplied the other five. Once Monday’s module is down, every Korean sector is in the ground and the three still to come are European.<\/p>\n Hyuk-Woon Kwon, South Korea’s ambassador to France, kept his summary at the ceremony narrow, which is more than fusion press events usually manage. The road is long, he said, but “steady progress is being made.”<\/p>\n That is roughly the honest read. ITER is still the slow, expensive, institutional answer to fusion, and private outfits are aiming at the same physics on much shorter timelines. Commonwealth Fusion Systems has its compact SPARC tokamak about 75% built outside Boston<\/a> and is chasing net energy gain by 2027, twelve years ahead of ITER’s fuel date.<\/p>\n What ITER has that the renderings do not is 440 tons of double-walled steel in a hole in Provence, five more like it alongside, and a schedule that moved the right direction for once. Whether the machine ever earns the money spent on it is a 2039 question. Whether the crews can stack nine wedges to within 10 millimeters is a 2027 one, and that one is going fine.<\/p>\n","protected":false},"excerpt":{"rendered":" Panel gaps are one of those details you cannot unsee. Once somebody points out that the seam between a hood … <\/p>\nWhat actually goes down the hole<\/h2>\n
The gap between pieces went from 100 millimeters to about 10<\/h2>\n
Assembly is running early, which is not the ITER anyone remembers<\/h2>\n
Nothing else in the machine moves until all nine are down<\/h2>\n
Korea is finished. Europe has three sectors left.<\/h2>\n