Imagine a nuclear industry that can power America for decades using its own radioactive garbage, burning up the parts of today's reactor wastes that are the hardest to dispose of. Add technology that takes nuclear chaff, uranium that was mined and processed but was mostly unusable, and converts it to still more fuel. Then add a global business model that makes it much less likely that reactor by-products such as plutonium will find their way into nuclear weapons in countries like Iran, even as economical nuclear-power technology becomes available to the whole world.

That is the alluring triple play the Bush administration hopes to turn with the Global Nuclear Energy Partnership (GNEP) it unveiled earlier this year, a proposed long-term research and development program almost as audacious as the Manhattan Project. The basic fuel-reprocessing concepts at its heart have been kicking around for the better part of a half-century. Now they are being touted anew as a way to provide plentiful carbon-free fuel for an energy-hungry world threatened by human-induced climate change.

Under the plan, for which the administration has requested $250 million for the fiscal year beginning October 1, the United States and certain partner countries would process spent nuclear fuel using new techniques that would turn some of it into more fuel and minimize the amount requiring disposal. The United States and its partners would also lease reactor fuel to other countries, which would then return their spent fuel to be reprocessed.

The technology could exploit uranium far more efficiently: Phillip J. Finck, associate director at Argonne National Laboratory near Chicago, says it could extract up to 100 times as much energy from uranium as is now possible. With the waste now piled up at reactors around the United States, the theory goes, GNEP could produce all the electricity the country will need for decades, maybe even centuries -- assuming enough of the necessary new reactors could be built. That would eliminate about a third of all U.S. carbon dioxide emissions (roughly the portion that today comes from fossil-fuel power plants). All this while reducing waste and thwarting the diversion of fuel to nuclear weapons.

In practice, though, in the best scenario GNEP would take decades to develop, and in the worst it might produce nothing; it could turn out to be a nonstarter on technical grounds, or the technology could be economically uncompetitive with other carbon-free sources of electricity. And the program could undermine a more modest and achievable goal: resuscitating a nuclear industry that hasn't launched a successful reactor project since 1974.

Today, a public once wary of nuclear energy has opened up to it as a possible answer to global warming. New reactor designs similar to those used in today's commercial fleet -- but said to be safer and more efficient -- are already approved or under review by the U.S. Nuclear Regulatory Commission. Utilities are in various stages of planning at least 16 such reactors (see "Stirrings of Renewal" chart) and may file applications with the NRC as early as the end of next year.