"Huh - 60-year-old technology," the giant, silver-haired septuagenarian car guy growls as a huge twin-bladed Chinook helicopter screams overhead. Bob has Googled himself a spotters' list of the latest Marine firepower and armed himself with a telescope as big as his forearm. The Hollywood Marines drive past in three roaring LAV-25s that drown out all thought and conversation; Bob's Aviators sweep the beach.

A slighter and more retiring man sitting beside Bob leans almost excessively forward. This is Dr Larry Burns, GM vice president of research and development and strategic planning. This should have been Larry's show, but Larry is profoundly hard of hearing and, in this racket, he cannot hear a thing. He doesn't need to, however, as Bob is answering everything today. Bob has taken over and one wonders who chose this venue under clattering skies to launch the Sequel.

After all, it was Larry's department that created the hydrogen fuel-cell, drive-by-wire sports utility vehicle we are here to drive, as well as its 2002 predecessors, the Hy-Wire and the fantastically innovative Autonomy 'skateboard'.

These were, and to some extent still are, the absolute acme of hydrogen fuel-cell-powered transport. As Christopher Borroni-Bird, Sequel's British programme director, explained back in 2002, "Electric power means a new sort of car, built in modern factories, supplied by high-tech, low-cost suppliers. GM is creating a new world order of personal transport."

The idea behind Autonomy was that all the major functions in a vehicle would be controlled by electrical wire, including brakes, steering and, of course, the drive system, delivered via tiny motors in the wheels. There would be no direct connection between brake pedal and brakes, steering wheel and wheels or accelerator and engine. Driver inputs would be interpreted and controlled by computer, and functions such as four-wheel steering and braking, as well as anti-lock, anti-skid, traction control and emergency brake assist, would be simply a line of code in the car's electronic brain. The skateboard-like chassis contained all the main vehicle functions and could be fitted with a range of demountable bodies, allowing you to have an SUV on holiday, a sports car at weekends and an MPV during the week.

Far-fetched? Futuristic? You bet. This was thinking of the highest order, perpetrated by Larry Burns's brilliant 500-strong team of engineers, scientists and thinkers. Even in the heady and profitable days of 2002, it offered a way out of the mature, smoke-stack industry problems of the motor-makers. Recently, these reached a nadir for Ford, with the news that the House of Henry could lose as much as $9bn (7.1bn) in 2006, while GM is also in severe financial trouble. Negotiations on a merger between the two collapsed earlier this month.

The Autonomy project offered a chance for the car industry - and especially GM - to reinvent itself as a modern, low-polluting, lights-off factory operation. But that was then and this is now. As Lutz explains, Sequel is now just a means to an end, and that end is not fuel-cell cars but "an all-electric architecture where all forms of engines as well as fuel cells can be used".

He explains, "The thinking is that the hydrogen infrastructure might not arrive, but we have an architecture that we could use for all engines. We are fixing [parts supplier] Delphi and saving costs of $2bn a year. We are reducing our healthcare and pension expenditure and our workforce.

"Our ongoing fixed costs will be lower by $9bn a year and that gives a lot of daylight. Some of that will go into increasing profits, but we are more than aware that our 20-year decline is partly a result of not allocating money to the business."

Lutz doesn't rule out hybrids, but says GM is more than impressed with the performance of lithium-ion batteries, which offer fast, high-power, 'memory-free' recharging. "The real issue is petroleum," he says, "and the real objective is electric drive, whether it's powered with a fuel cell or a lithium-ion battery. Hell, we just want to get out from under the oil companies."

Lutz blames the American government for GM's disenchantment with its world-beating fuel cell. "The US government is dragging its feet over the hydrogen infrastructure," he says, adding that GM remains committed to producing one million fuel-cell cars profitably - but that might be in China, for Chinese mmarkets. "China is building loads of nuclear power stations," he says, "and we know that nuclear can produce almost fossil-free hydrogen, and the Chinese government is really keen to get involved."

Apart from this obvious cooling on the primacy of fuel-cell research, you have to boggle at the thought of GM returning to all-battery technology after the debacle of the EV-1, a lead-acid battery-powered car produced between 1996 and 1999. More than 1,110 of these sleek, 80mph coupes were built and 800 were leased out to customers, but if you took away the subsidies from the US government, each car would have cost GM just under $1 million.

Eventually, the company recalled and crushed most of them. The EV-1 was the subject of this year's nonsensical conspiracy movie, Who Killed The Electric Car? But listening to Lutz, it is hard not to disagree with Wally E Ripple, a research engineer interviewed in the film, who suggested that because there is a trillion dollars worth of oil still left in the ground, representing over 100 trillion dollars' worth of business for car-makers and oil companies, there is little incentive to develop a viable electric (or fuel-cell) car.

Perhaps he's right. For all the current fixation with Peak Oil and America's 'addiction to oil', no one has done a trustworthy, well-by-well audit of what oil is left under the ground, so no one really knows. Last week, oil prices started to fall to about $60 a barrel from their July high of $78, and the heads of Exxon and Saudi Aramco, the oil company with the world's largest output, made calming noises about the state of current stocks.

They claimed, respectively, that the end of oil was nowhere in sight and that at current production rates there was a century's worth of crude left. Well, they would say that, wouldn't they, but as long as we have no way of verifying what stocks remain, their opinions are at least as valid as any other.

Some Western analysts might argue that Bob Lutz is right, and that however much oil is left we should eke out supplies by using a mix of hybrid, electric and fuel-cell power. But there are other issues. Reversing climate change will require a reduction in the burning of fossil fuel and/or the sequestration of carbon dioxide. Energy security has become more of an issue since Russia shut off its natural gas supplies to Ukraine in a price row earlier this year, and since Osama bin Laden vowed to attack Western oil installations.

And the rules of the game are changing. Since April 2003, the California Air Resources Board (Carb) has given car-makers the option of meeting part of their mandatory zero-emissions targets by producing fuel-cell vehicles. Each company has to produce a sales-weighted number of fuel-cell vehicles to contribute to a total of 250 in 2008, rising to 2,500 between 2009 and 2011, 25,000 between 2012 and 2014, and 50,000 between 2015 and 2017. Battery vehicles can be used to replace up to 50 per cent of that requirement.

It's a tall order, and to comply GM is having to send 62 of the 100 fuel-cell Equinox SUV test vehicles, but the American giant is late to the party. Rivals Toyota and Honda already have large fleets of fuel-cell vehicles on test and Takeo Fukui, Honda's chief executive, is committed to putting a fuel-cell car on sale within five years. DaimlerChrysler has a test fleet of more than 100 fuel-cell vehicles and there's even a DaimlerChrysler fuel-cell bus in London, running on route RV1 between Covent Garden and London Bridge.

GM is falling behind, and when you are behind you can no longer dictate the terms of the debate. All the good things that GM was hoping might fall into its lap as a result of its fuel-cell research could be up for grabs once again.