Fusion power was produced in these [previous] experiments - the record now is about 10 MW - but only for a second or two.

During these experiments the whole experimental plant was consuming more than 100 MW, so no prizes for these experiments as a power source! The problem is that these experiments are too small to generate sufficient power to make a power source, whereas ITER would be in the size range where a reactor could operate. Based on extrapolation from existing experiments, ITER should generate 500 MW of fusion power for the range between several hundred seconds up to continuous operation. This power shows up as heat which is cooled away. In principle, you could connect up a steam turbine and generate electricity with this heat (ITER will not do this for efficiency, cost and reliability reasons), which might give about 150 MW of electrical power. The power taken to run ITER is about 110 MW electrical, so ITER could sell about 40 MW electrical back to the grid. Of course a commercial scale reactor would need to be just a bit bigger (or work a bit better - it is not yet clear which - hence the need to construct ITER), so that a fusion reactor would have typically 4000 MW of fusion power (1300 MW electrical power) and hence sell ~1000 MW to the grid.

The problem with hot fusion is rather getting it to work at all, since it requires special conditions which are easy to go outside, causing the reaction to die out. Then of course although the fuels are cheap, the capital investment up front is rather high, requiring a lot of confidence from the investors that it is going to work (hence the worldwide experimental development programs which spreads the risk of failure so everyone's cost is minimised, and so that when we eventually get to building power reactors we can be confident they will work). Thirdly for fusion to not create radioactive waste that hangs around for a long time (>100 years) existing structural steels and other materials will have to be developed a bit to rid them of problem isotopes (but here already ITER is pretty good with most of its materials available for recycling after 100 years storage) and to test them so the nuclear regulators will sanction their use.

Actually where fusion goes and whether it gets developed at all depends on the taxpayer and his elected representatives. They have taken the initiative to put together programmes of fusion research around the world and most of these programmes lead on to our machine, ITER, as their next logical step. ITER is however just an experiment so they have taken this decision because presumably they foresee an energy shortage coming up in the medium term and want to collaborate on developing one way to overcome it. How well fusion stands up against the alternatives is something we in the business cannot decide. We can only advise.

So the simple answer is that we are almost ready to go ahead and build ITER. It just needs the commitment of the governments and therefore the taxpayers concerned. We can only advise them that we think we are ready, can do it and that it is essential to do it for longer term energy (and hence economic) stability. It's then up to them. The building of ITER would take 10 years, and 20 years are foreseen for operation, so we could probably start designing DEMO by 2020, and if needed operating it by 2035. This puts the first commercial fusion power station (the step after DEMO) on line at the earliest in about 2050.

ITER is not an electric power producing reactor. In some respects, like plasma size, ITER is like a prototype power reactor, but in others it is far away. That's why the first commercial implementation of fusion has to remain in the middle of the next century. There is still lots more engineering to do to make the device reliable and economically competetive.

ITER will therefore not produce any electrical power, except in a token experiment using a test blanket module to show it is possible. To use conventional steels in our experimental plant (necessary to get nuclear licensing) we have to operate at relatively low temperatures. These are not the best for steam raising so we have saved ourselves the cost of a turbogenerator, thermal store, etc. ITER will also be a pretty unreliable energy source at the beginning, so this also makes sense from that viewpoint. The heat will be rejected to the environment, probably via cooling towers. In the follow up device - DEMO - conventional steam raising and power generation will be used.

Up to now fusion experiments have addressed only the scientific feasibility of fusion, with little attention to how it can be turned into a practical electrical power source. ITER is the first really "nuclear" fusion experiment, which from the outset has to be designed to satisfy anything mother nature and the safety regulators can dream up, as well as to be repairable when it fails. The only convincing experiment is one conducted in the real environment, and ITER will be the first time prototypical fusion power reactor technologies will be put together with plasma physics to try to make a viable power source. No doubt we will get some of the technology wrong, and make some items at unnecesary cost, but we will only find that out by trying this step. Further speculation from today's knowledge base will tell us nothing.

ITER will really be the first fusion device that produces more energy than it consumes, but for a limited time. Experiments so far (like JET, TFTR or JT-60) have attempted to reach "breakeven" - a point where the fusion power produced by reactions in the plasma is equal to the power you are putting into the plasma from outside to keep it hot. However although going close, they have not achieved this. ITER will produce many times more power from fusion than it consumes as power input to the plasma, just because of the scaling to larger size. Bigger plasmas are easier to keep hot than small ones. However, be careful in your understanding, I am talking about thermal power (i.e energy transfer rate - J/s - at a given time) and not electrical energy (i.e J). ITER will not be used to generate electrical energy (i.e electricity), except on a token small scale to indicate that we can do it. That is because we would have to design a higher temperature blanket around the plasma to do this efficiently, which it is not worth designing (we will test our ideas on ITER but in special test cells which also include prototypical tritium breeding) and installing turbogenerators (which would cost more money). We would also have to operate with short times between pulses or in steady state, and we don't know how to do that yet and want to keep costs to a minimum for this experiment. The thermal energy we create when the plasma is operating will be dissipated into the environment via cooling towers. However we will need electrical energy to drive our systems and heat the plasma using microwaves and neutral particle beams. So in this sense only the machine afterwards (DEMO) will be a net producer of electricity (if that is what you mean by energy). However if you add up all the electrical inputs to the plant in ITER and convert them to thermal energy, and compare that with the thermal energy we can produce operating our pulsed device as often as we can, even ITER will produce more thermal energy than it consumes.