An international plan to build a nuclear fusion reactor is being threatened by rising costs, delays and technical challenges.
Emails leaked to the BBC indicate that construction costs for the experimental fusion project called Iter have more than doubled.
Some scientists also believe that the technical hurdles to fusion have become more difficult to overcome and that the development of fusion as a commercial power source is still at least 100 years away.
At a meeting in Japan on Wednesday, members of the governing Iter council reviewed the plans and may agree to scale back the project.
'Size of a battleship'
On a windy construction site in the south of France, the lofty scientific goal of developing nuclear fusion as a power source is starting to take on a more substantial form.
Covering an area of more than 400,000 square metres, workers have built a one-kilometre-long earthen platform on which the experimental reactor will sit.
"This is going to be the world's biggest science experiment," says Neil Calder, Iter's head of communications.
"This is a vast global project to show the scientific feasibility of fusion as a limitless source of energy.
"On top of this platform we are going to build 130 buildings. The main building will contain the Iter machine itself.
"It will be huge - the size of the Arc de Triomphe in Paris - and it'll weigh about the same as a battleship - 36,000 tonnes of metal and instrumentation."
Iter was formally launched in 2006 as collaboration between the European Union, the United States, Russia, Japan, China, India and South Korea. The plan was to build the world's most advanced fusion experiment within 10 years for a budget of $6bn (£3.6bn).
But the grand scheme has been dogged by soaring costs caused by more expensive raw materials and increases in staff numbers. Emails seen by the BBC indicate that the total price of constructing the experiment is now expected to be in excess of $16bn (£10bn).
Professor Sebastien Balibar is research director for the French national research laboratory in Paris. He says that if the rising price of Iter is met by cutting back other research programmes that would be a disaster for science.
"If Iter is built on money having to do with energy or oil, that is perfectly good, I hope it works and in one hundred years I hope we know how to control a fusion reaction. But if it is taken from the public support of research in physics or biology then I would be very upset," says Professor Balibar.
Costs are not the only problem; Iter is also beset by huge technical challenges.
Fusion takes place when a superheated gas called a plasma reaches a stage called ignition, where hydrogen atoms start to fuse with each other and release large amounts of energy. Iter aims to achieve this but only for a few minutes at a time.
MIT professor Bruno Coppi has been working on fusion research in Italy and the United States for many decades. He believes that Iter is the wrong experiment; it is too costly, will take too long and may not deliver fusion. He says we should be looking at other options.
"We are pressed for time, the climate situation is worse. I think we should go with a faster line of experiments. Iter should admit its limitations and it will give a limited contribution to fusion, but to get to ignition you need to follow a different road," he says.
Another huge hurdle is how to contain gases that are 10 times hotter than the Sun. The materials required simply haven't been invented yet.
Professor Balibar explained: "The most difficult problem is the problem of materials. Some time ago I declared that fusion is like trying to put the Sun in a box - but we don't know how to make the box.
"The walls of the box, which need to be leak tight, are bombarded by these neutrons which can make stainless steel boil. Some people say it is just a question of inventing a stainless steel which is porous to let these particles through; personally I would have started by inventing this material."
Failure a possibility
In Provence, the scientists working on Iter say they have faith that the project will deliver the most effective path to fusion.
Dr Norbert Holtkamp is the man tasked with building the machine.
"Iter is a step that will demonstrate whether fusion is viable. But whether it is easy then depends on the cost of energy at that time on the cost of oil, but certainly Iter has the potential.
Dr Holtkamp recognises that Iter is a scientific experiment - and as such it has the possibility of failure.
"Any project can fail, especially if it's one of a kind or the first of its kind. It would be irresponsible for any scientist or project manager to say that in a science project it cannot fail."
The rising costs of construction and technical challenges are to be reviewed at a meeting of the Iter council in Japan on Wednesday and Thursday. It is possible that by the end of this year, a new scaled-down version of Iter will be agreed.
Dr Holtkamp says the view that the project is to be scaled down is wrong.
"Fusion is not going to be the alternative in the next 20, 30 or 40 years, that is correct. But there needs to a long term plan; 40 years is little more than a generation. We need to think about the next generation and the many after that."
Professor Balibar says that the end result of the ballooning costs and increasing technical challenges will be a further slowing of the path to fusion.
"The consequence of all these difficulties is that it's not going to be tomorrow that one succeeds with fusion. But the energy problem and the climate problem are urgent," he says.
"The global warming is now - one needs to find a solution immediately, one cannot wait 100 years. The solution to the climate and energy problem is not Iter, (it) is not fusion."
While fusion offers a long-term hope of securing energy supplies, the changing climate and the pressing need for greener energy may ensure that renewables get greater political support in the short to medium term.
Ultimately fusion may be a technological dream that is just too hard to turn into reality. And Iter, in a beautiful setting in the south of France, may become the graveyard of a good but impossible idea.