Fusion energy wasn't new back in 1972, when physicist Robert Goldston first arrived at the Princeton Plasma Physics Laboratory as a graduate student. His colleagues were ecstatic over generating one tenth of a watt of electricity for one hundredth of a second--enough to make a light bulb flicker. "We thought that was so amazing," he says, "we held a big party."

More than three decades later, the celebrations are still going on, but no bulbs are lit. Last week a research group called ITER, which includes Japan, Europe and the United States, announced that it would build the world's first large-scale fusion reactor in Cadarache, France. If all goes well, the plant will make 500 megawatts for 400 seconds--enough to run air conditioners in a small city for a few minutes. The trouble is, scientists will have to pump even more energy than that just to get the plant warmed up. Net output: negative.

Commercial fusion plants might already have been pumping electricity onto the grid if funding levels hadn't been cut in the 1980s, when oil was cheap and it wasn't a high priority, says Goldston, now director of the Princeton lab (an ITER participant). It would have come just in time for soaring demand for energy. Unlike conventional nuclear plants, which split atoms, fusion plants combine (or fuse) hydrogen atoms to form helium--at no risk of blowing up and without any nasty materials that could be used in a bomb. And they emit no carbon. Now it may take another 30 years to bring them online. Time enough for plenty of parties, by candlelight.