hile carmakers agonize over the variety of options for investments in fuel economy, the long-term hope is clearly hydrogen. Industry consensus puts the hydrogen horizon at a minimum of 15 years, with other guesses at two to four decades. This timeline matters. One scenario conceived by NRDC compared aggressive fuel economy (40 miles per gallon by 2012 and 55 miles per gallon by 2020) to early hydrogen adoption (100,000 fuel-cell vehicles per year by 2010 and 2.5 million per year in 2020). This is a very favorable development pace for hydrogen, yet implementing aggressive fuel economy would still save almost 25 times as much oil by 2020.

Larry Burns, the vice president of advanced product development for G.M., sees first light for hydrogen in 2010: "I'm talking about fuel cells that are the same cost and durability as gasoline propulsion systems," he says confidently. G.M. believes that whoever cracks the fuel cell puzzle first will assume a commanding height in the global economy. Burns sees huge opportunity in the fact that 88 percent of the world's people do not own a car or truck. He looks at booming markets in India and China and sees dollar signs. He understands that the way cars are currently made, and the rate at which they consume fuel, cannot support the kind of exponential growth he anticipates. And he insists fuel cell cars will perform so much better than current vehicles that they will drive significant market change.

Burns is not alone in seeing this potential, but a recent report by the American Physical Society identifies more than a few critical technology gaps: A panel of physicists concluded that hydrogen production needs a fourfold to tenfold improvement in cost and efficiency, while fuel cells themselves must improve 10 to 100 times in cost, size, durability, and ease of manufacture.

Widespread acceptance of hydrogen cars will require widespread availability of hydrogen. In the United States, 175,000 filling stations sell petroleum; 25 sell hydrogen. Cost estimates for this endeavor spiral into the hundreds of billions of dollars. The potential of hydrogen to be a climate-change savior also assumes environmentally sustainable production, but in the short run much of this energy source will likely come from fossil fuels, releasing still more greenhouse gases.

These question marks, and the wide consensus that Burns is a decade off the mark, haunt an otherwise compelling dream. G.M.'s pursuit of fuel cells is, in effect, a high-stakes poker game with a 10- figure ante. "I think it's a mistake for G.M. to plunk down a billion-dollar bet on hydrogen and fuel cells in the next 10 years," says NRDC's Hwang, who sees much better business prospects in more conventional technology. "It's easy to get seduced by the sexy technology, the water vapor coming out of the tailpipe." And he worries that G.M. may be trying to dodge federal action on fuel economy in the present with its seductive vision of the future.

"Maybe it's reasonable that people are skeptical, given some of the history of the auto industry," says Burns, in passing reference to its unseemly resistance to such public benefits as seat belts, air bags, and emission-control devices. But he protests that the fuel cell critics are not in his labs. "We have not seen one thing yet that suggests the roadmap we have [for affordability] can't be realized by 2010," he says. His real concern is the competition: Toyota, Honda, Nissan, Ford, Ballard, VW, United Technologies, Hitachi. All of these companies are developing fuel cells, and Toyota is rumored to have outspent even G.M. "I don't intend to blink," Burns says. "We're not doing this because of a CAFE head fake. We're doing it because of the competitive threat of somebody else figuring it out."

But Hwang at the NRDC doesn't buy it. "The engineers may be well intentioned, but G.M.'s lobbyists, and its PR machine, are as Machiavellian as they come," he says.