For now, it is a hideous sight. In Cadarache, 60 kilometres north of Marseilles, workers have cleared over 40 hectares of wooded land and moved more than two million cubic metres of soil. However, this scar on the Provençal landscape has been earmarked for greatness. It is where a multinational team of scientists is attempting to build earth's second sun.
As projects go, the International Thermonuclear Experimental Reactor (Iter) could hardly be more ambitious. Its aim is to show that we can control nuclear fusion reactions. This is the same process as generates energy in stars and could, in theory, release up to four million times more power than burning fossil fuels. If Iter works, we'll have solved our energy problems.
But ifs do not come much bigger than that. We do not yet know if it is even possible to build the machine. "Fusion is a big bet - it's not a dead cert," says Steven Cowley, director of the Culham Centre for Fusion Energy, the hub of UK fusion research. The stake for that bet is set at €10bn (£9bn), but that figure is double the original estimate for the project and may rise further; Iter's council was recently presented with just the latest in a series of revised budgets and schedules. Whatever it eventually costs, we will not find out whether the gamble has paid off until 2026, the earliest date for the project's completion.
All this uncertainty and delayed gratification, not helped by the price tag, has generated heat of its own. Iter's critics, who include prominent scientists and Greenpeace International, have argued that the money would be better spent on pressing challenges such as finding ways to increase near-term energy production.
However, the fusion scientists are keen to point out that they are being responsible. It is no use surviving the near term only to find we are faced with a huge energy debt, they argue. World consumption is on course almost to double by 2030. Solar energy and nuclear fission might be more immediately available, but both have their limits. Nuclear fusion's main fuel is derived from seawater, and there are no long-term nuclear waste products. Nothing, they say, would fill the energy gap like this.
Bombard with microwaves
That is what Iter's members - Russia, the EU, Japan, China, South Korea, the US and India - are hoping their 23,000-tonne monster will prove. The jaw-dropping size of Iter is necessary because making commercially viable electricity from fusion depends on economies of scale. Previous successes in smaller reactors have managed to break even, producing as much energy as they consume. But the Cadarache reactor should, according to its designers, give out ten times more power than it takes in.
Operating at 150 million degrees Celsius, ten times hotter than the core of the sun, Iter is certainly going to take in a lot of power. Surprisingly, this kind of temperature is not too hard to achieve. The fuel for Iter is two heavy isotopes of hydrogen called deuterium and tritium. Bombard them with microwaves, magnetic fields and other particles, and they will get hot enough to fuse, releasing energy.
The hard bit comes with the maelstrom created inside the reactor. The high temperature creates a "plasma", a gas of charged particles. Plasma is an engineer's worst nightmare. For a start, it cannot be allowed to touch the reactor's walls; if it does, they will melt, and the whole thing will have to be rebuilt.
The plasma can be held away from the walls using immensely strong magnetic fields, but only - so far - for short periods. This is because the plasma tends to slip around in its magnetic cage, forming areas of high density that can burst through. Even if Iter engineers manage to hold it stable for ten minutes at a time, which is as much as they hope to achieve, the plasma will still shoot out neutrons that can destroy the walls.
This is the frontier where Iter succeeds or fails, Cowley believes. "We're pretty sure we can get out ten times the energy we put in," he says. "But if we have to replace the wall every year, that's going to be a very expensive way to produce electricity."
Once all the engineering problems are overcome, the plant will be able to produce only 500 megawatts of power, equivalent to a single coal-fired power station. Members will then have to build their own fusion reactors using the know-how gained at Iter. Payback will come, so the rationale goes, through these states' privileged position in the trillion-dollar, post-fossil-fuel, global energy market.
It's not an argument that worked for Canada, which pulled out of the fusion dream in 2003. The US also wavered, though it has now committed to paying 9 per cent of the cost. The EU is putting in the largest share, taking responsibility for just under half of the project. Thanks to the strange arithmetic of fusion, however, EU taxpayers may end up paying significantly more than half of the money.
The funding of Iter is a notoriously slippery subject. Roughly 90 per cent of the contributions are due "in kind" - states will contract firms to manufacture equipment for a cost that they do not have to declare to the other states. Even more confusing is that each of Iter's components has been designated as worth a certain number of "Iter accounting units". Members can then choose which components they commission firms in their countries to design and build. This will affect the balance of expenditure; the cost of producing a particular magnet is likely to be far less in China than in Germany, for instance.
Then there is the complexity of the various components. The UK has chosen to build superconducting magnets and the main container vessel for the plasma. These, it turns out, will cost much more to design and build than initial estimates suggested. Cowley maintains this is a good thing: the money will go to UK industries and provide them with engineering challenges that will have their own spin-off benefits, he says.
“We will never really know how much some countries spent," admits Neil Calder, Iter's spokesman. This lack of clarity about the cost may prove to be the project's Achilles heel.
In May, the journal Nature declared it "deeply unfair" to the taxpayers paying for the project and called for "an honest public debate". Science also weighed in, suggesting that fusion's problems could well be intractable. Fusion, said one commentator in the journal, is "the science of wishful thinking".
There is no sign of second thoughts from any of the members, however. According to Sébastien Balibar, a director at France's National Centre for Scientific Research, members stand to gain nothing by halting the project. "Now that Iter has been decided and is under construction, it would be better that it produces useful results," he says.
Michael Brooks is a consultant for New Scientist and the author of "13 Things that Don't Make Sense: the Most Intriguing Scientific Mysteries of Our Time" (Profile Books, £12.99)