A much-delayed nuclear fusion project involving more than 30 countries is ready to assemble the world’s most powerful magnet - a key part of efforts to generate clean energy by smashing atoms together at super-high temperatures.

The International Thermonuclear Experimental Reactor (ITER) project, based in southern France and backed by the United States, China, Japan, Russia and the European Union, needs the magnetic system to create an “invisible cage” to confine super-hot plasma particles that combine and fuse to release energy.

ITER said late on Wednesday that the final component of the system - the central solenoid - had been completed and tested by the United States, and assembly was now underway.

“It is like the bottle in a bottle of wine: of course the wine is maybe more important than the bottle, but you need the bottle in order to put the wine inside,” said Pietro Barabaschi, ITER’s director general.

The magnet was originally scheduled for completion in 2021, but has been beset by delays.

“To be behind schedule by four years after 10 years of effort shows just how troubled this project is,” said Charles Seife, a professor at New York University who writes about nuclear fusion.

Barabaschi said the “crisis” was now over and construction was proceeding at the fastest pace in ITER’s history. The start-up phase of the project will begin in 2033, when it is scheduled to start generating plasma.

He said ITER proved that countries could still cooperate despite geopolitical tensions.

“They have a very, very strong cohesion of objectives and for the time being I see no sign of a withdrawal from anyone.”

Fusion investment has been growing, with dozens of initiatives currently underway. Several private start-ups have said they can build commercial fusion reactors within a decade.

Barabaschi said he was sceptical but supportive of the dozens of ventures in development across the world.

“We already know that we can get fusion,” he said. “The question is, are we going to get fusion in such a way that it would be cost-effective?

“I am quite sceptical that we will be able to achieve this within, say, one or even two decades. Frankly speaking, it will take more time.”