Eric Takamura signs off the computer and picks up his coat and keys. No sense putting this off. He steps across the hall to the conference room doorway and tells Joel Jermakian he's heading downtown. The meeting is going to be at some restaurant -- "Chef Goff's," he says.

"Chef Geoff's," Jermakian corrects him.

Whatever. Takamura shrugs his trench coat higher on his shoulders and walks out into the late afternoon, knowing the traffic's going to be awful on the Dulles Toll Road into Washington. But he dreads the meeting more than the drive. It's not so much that his future is on the line again; after nine years of helping to run a small company, he's used to the feeling of looming disaster.

What really worries Takamura about this meeting is that it's supposed to be an informal "chat session," a chance to mingle with a group of bigwigs from investment firms. If they like him, maybe they'll invite him to a more formal session, where he would get two minutes to make his pitch and persuade them to invest in his company. He needs the money. If the company can just keep it together for one more year, the technology he and a few colleagues have poured their lives into might finally be on track for something big.

And so tonight, to clear the first hurdle, he has to chat.

Takamura can rebuild a set of brakes or calculate the vibration of a moving vehicle. He can change a diaper, fix drywall and hang out all night drinking and dancing with friends. But Takamura doesn't chat.

So he smokes to relax as he drives his 1995 Toyota Corolla, wearing his good suit and a fresh haircut. He picks up one colleague and meets another at the downtown restaurant. They seem to be early, though really they're just unfashionably on time. As the room fills up, as people with name tags pinned to their suits greet each other like old friends, Takamura and his two colleagues stand in a knot, clutching drinks and glancing over their shoulders at the milling crowd.

Most of these other people are here to promote their companies, too, but they're a different breed. They peddle software or Internet services. This is their element: networking, handing out business cards and spouting punchy descriptions of the amazing, innovative, can't-miss ideas that will soon generate tons of cash for savvy investors. They'll shake hands with anybody, but the real objective is to find a guest with a little green tab at the bottom of his or her name tag. Green equals cash: The tab-wearers are from venture capital groups looking to put money into hot properties. It's like the late 1990s all over again.

This is a scene Takamura and his colleagues have never experienced. Their company has a clunky name, New Generation Motors, and it makes a mundane-sounding product: electric motors. But there is nothing mundane about what they believe their product can do. This motor is so efficient, they believe, it could help revolutionize the auto industry at a time when gas prices are zooming to record highs. It could make electric cars practical, or help put gas-electric hybrid cars in every driveway, or power buses that don't pollute.

If this were a gathering of college engineering students instead of Internet executives, the guys from New Generation would probably be the center of attention. In the insular, ultracompetitive world of solar car racing, they're stars. Top students from around the globe build and race solar-powered cars, and virtually all the best U.S. college teams use electric motors from New Generation.

Somehow, the New Generation guys have to make these people -- drinking free booze, nibbling on chicken wings and comparing Internet domain names -- understand what they have to offer. For that, Takamura is counting on Anubhav Sethi, a recent George Washington University grad who majored in business and hired on at New Generation a year ago.

Sethi knows the language of business. Unlike Takamura, he seems comfortable in his suit, and his wide, slightly pudgy face radiates sincerity and competence. Clutching a zip-up leather organizer, Sethi agrees to start working the crowd as Takamura and his colleague, GW professor Nabih Bedewi, scurry back to the bar.

The room, lit by candles that give the wood paneling a fleshy glow, is now full, and a few lazy ceiling fans aren't quite enough to keep the air from growing heavy with body heat and aspiration. Sethi dabs a handkerchief at his forehead. He hasn't actually done this networking thing before with such a big group.

He turns into the path of another guy working the room. There's an awkward moment as they stoop to check name tags, and Mark Lewyn introduces himself as chairman of a company called Paxfire Inc. What, he wants to know, is New Generation Motors?

"We make electric motors," Sethi says.

"Sounds like a good thing," Lewyn says. He waits, and it dawns on Sethi to ask what Paxfire is, and Lewyn replies: "I would describe us as an Internet recycler." He lets the line sink in. "We take mistypes, capture that and turn it into money." If users make typos entering an Internet address or topic, Paxfire guesses what they were looking for and routes them to a search engine, saving time and effort. His is a snappy synopsis, honed for cocktail patter. After a moment, he moves on.

Sethi mops his forehead again.

Another guy shows up almost immediately, a balding, sandy-haired man wearing a dark blazer and gray slacks as naturally as some people wear track suits. He's Tim Meyers, and he has a green tab on his name tag. He's on the group's selection committee, which decides which companies get invited back for a formal shot at investment gold.

"Tell me, what is the main use of your motor?" Meyers says. "I'm a software guy, so when I saw that I'm like, 'I need help!'"

Sethi begins explaining the company's technology, including the percentage of energy the motor captures and the unusual method used to control output. Originally from India, Sethi speaks beautiful English, but he speaks softly and with a slight accent. In the roar of the room, he is almost impossible to hear. Sweat is now pouring down his forehead.

"Okay," Meyers interrupts, "so give me a use."

Sethi mentions a deal the company has to build three-wheeled "autorickshaws" in India, and Meyers has seen these vehicles on trips abroad, but then Sethi is off again into an explanation of how the motor works.

"Give me the efficiency level," Meyers interrupts again. "'Cuz I'm a software guy. I'm not an engineering guy."

Sethi wipes his forehead. He explains how much more efficient their motor is than others. Meyers nods. "Okay, so it's a significant improvement," he says. He chats a moment longer, mentions that he's traveling soon to India, then shakes hands and begins to move away. At the last moment, Sethi realizes he's about to commit a giant faux pas. He saves himself by tapping Meyers on the arm and offering a business card, which causes Meyers to produce one in return.

When Takamura and Bedewi materialize a moment later, Sethi has a sheen of sweat like he's been doing jumping jacks. Bedewi says someone has told him they should talk to a guy named Tim Meyers. "I just talked to him," Sethi says.

"He's the chairman of the committee," Bedewi says, "so we need to gang up on him."

They go stand near the group where Meyers is now holding court. Takamura, hovering behind Sethi and Bedewi, nurses a bottle of Heineken with a napkin wrapped around it. They don't chat. They wait.

They're used to biding their time. As hard as it is to get the New Generation message across to these software guys, it's been even harder to get the ear of the auto industry. So the company has had to take an unusual route, circling offshore to foreign customers in hopes of working its way back to the Promised Land of Detroit. Tempered by years of hard knocks, they still have the dream that made them devote their careers to the low pay and uncertain future of running a small start-up business. They've come this far. If they can make it one more year, they think, they'll be in good shape. Just one more year.

The auto industry is almost irresistible to certain people looking to make their mark in business. It's so pervasive, so big. Americans bought 16.7 million cars and trucks last year. According to government figures, we now have 230 million vehicles registered in a nation of 290 million people. Think of that. It's a parallel population of machines, each weighing a ton or more and consuming petroleum products by the barrel, waiting for the command to drive to the mall. Cars literally carry the consumer economy.

For most of the past century, cars have been where ordinary people get into the guts of technology. Americans work on their cars in the driveway, changing oil, hopping them up to make them go faster. Historian Stephen Ambrose credited the Yankee compulsion to tinker under the hood with helping defeat the Nazis in Europe -- GIs fiddled with their tanks and adapted to problems, while the Germans struggled with their beautifully engineered machines whenever they broke down.

For inventors, automobiles are a practical object of desire. Invent something for the space program, and you're a dreamer; invent something to use in cars, and pretty soon you start seeing dollar signs. If your gadget goes onto even one car in 10 and makes you a dollar of profit on each, boom, you're an instant millionaire. Except it almost never works out that way.

When cars first became commercial products a century ago, scores of companies sprang up to make them. One by one, those disappeared or got swallowed by the competition. Today the Big Three are so big -- General Motors does more business in three months than Lockheed Martin, the world's largest defense contractor, does in a year -- that they're slow to embrace change. Radical new technologies would upset the way the U.S. companies make money. Electric cars, for instance, have very few moving parts, which is bad news for the huge business in parts and maintenance.

Introducing technology to the auto market is different from making, say, a new type of stealth fighter plane, which is going to be bought only in small numbers by a government with bottomless resources and trained armies of technicians. Cars have to be available to every Joe Schmoe with a checkbook, and, at the same time, have to be immune to liability lawyers. You can't just go sticking gee-whiz gadgets on them.

The guys at New Generation didn't see those hurdles at first, because they didn't set out to invent something that would make them rich. They were just out to win races.

In the early 1990s, engineering students at GW wanted to enter the solar car races that had recently become an international phenomenon. There was a big one every three years in Australia, and the U.S. Department of Energy, along with GM, had begun sponsoring biennial solar races in the United States, as well. Universities saw them as a way to inject some pizzazz into engineering programs; the government used them to promote solar power; and GM was looking to both associate its name with cutting-edge technology and help groom future auto industry scientists.

Bedewi was in a unique position to help the GW students put together an entry. The engineering professor had founded the National Crash Analysis Center at the university's Ashburn campus, under a contract with the Department of Transportation. Bedewi ran the center, which used computer modeling to predict how cars and trucks would behave in crashes and then actually crashed them to validate the results.

Detroit, needless to say, was interested in the professor's work. Bedewi used his contacts in the auto industry to get Louis Ross, then vice chairman of Ford Motor Co., to take a look at GW's fledgling solar car effort. Ross pledged support, and the team was up and running.

Solar car teams are a massive undertaking, requiring students to raise hundreds of thousands of dollars and spend countless hours working on the complex design, assembly and testing of their vehicles. One of the earliest to sign up for the GW effort was a long-haired engineering student named Eric Takamura.

Takamura was exactly the kind of student the solar projects were designed to reach. He had never been thrilled by class work, which seemed dry and pointless. As a kid growing up in Hawaii, Takamura idolized his uncle, a French engineer who roamed the globe building things in far-flung parts of Africa and Indonesia. The work -- stringing power lines, building dams -- was just an excuse for venturing to the edges of the known world. He was Indiana Jones with a protractor.

The solar car project gave Takamura a hint of that kind of adventure. It let him learn with his hands and created a little exotic world on the fringes of everyday university life. The school rented office space to the team in the basement of the student center, in a former parking garage. Takamura spent almost all his time there. He missed classes, or showed up late and slept through them. One time his mother came to visit and found him in the basement office, sleeping in a box full of styrofoam packing peanuts.

The team spent eight months building its three-wheeled car, a low-slung craft that looked like a long, flat raindrop. In the 1993 Sunrayce, a 1,000-mile trek from Dallas to Minneapolis, GW's team members did well for newcomers, finishing fourth out of 34 competitors. But the car was plagued by problems large and small -- tires that blew out, a solar array that didn't give it full power.

Bedewi enlisted the aid of Joel Jermakian, a scientist at NASA's Goddard space center in Greenbelt who had consulted for previous teams at the University of Maryland. After the Sunrayce, Jermakian -- who at NASA designed solar panels for space vehicles -- helped the GW team rework its solar array.

Now they were aiming for an even bigger race, the World Solar Challenge, run across 1,800 miles of the Australian Outback. This time, the team was determined to be prepared. Members analyzed every system on the car, and they tested it repeatedly on a practice course in southern Maryland. Their confidence began to soar. They could measure themselves against what the competition had done in the Sunrayce; GW was now better than that. Australia, the most prestigious race on the solar car circuit, was starting to look like a sure thing.

A few weeks before race time, they hired professional packers to load the car and a backup solar array into separate wooden crates for shipment. When the fragile load reached Australia, Takamura and two teammates were there to meet the plane. A forklift raised both crates from the plane, lost balance and tipped them forward. The crates came down on the tips of the forklift, which speared the car through its center like a bug on a pin. The backup solar array fell about 12 feet and smashed.

All that work, splintered in an instant.

As angry as Takamura felt, it's the nature of an engineer to regard life as a series of problems that can be solved through calculation and action. He got on his cell phone and called Jermakian. With four weeks remaining before the race, every team member who could make it flew to Australia and began rebuilding the car.

The GW team managed to make it to the starting line, but the disaster cost it any real chance at the title. It finished ninth out of 51 competitors. Still, the crash turned out to be a fortunate event. It was the key to everything that came later.

When GW had to rebuild so quickly, it had gotten help from the local Northern Territory University and its solar car team leader, Dean Patterson. Takamura and his crew got a good look at an experimental type of electric motor Patterson had designed for the NTU solar car. It fit into the wheel hub, eliminating the need for a complicated transmission system between motor and wheel. It was like the difference between pedaling a bicycle, which uses a chain to make the rear wheel turn as you work the pedals, and pedaling a unicycle, which puts the force directly into turning the wheel. Getting rid of the "chain" made the solar car simpler and more efficient. What's more, the motor seemed to use energy more efficiently than the off-the-shelf motor GW had bought for its car, which analysis showed was the least-efficient component of the vehicle.

The following year, when Patterson was visiting the United States, he stopped by GW. Patterson agreed to license his motor design to the university's solar team. The team spent months making refinements and, by the 1995 Sunrayce, had produced a better version of the Aussie motor -- winding the coils differently and making other improvements that increased efficiency. More problems with the solar array kept GW from winning that race, but the team won an innovation award for its propulsion system. A few months later, with the solar problems ironed out, GW traveled to Japan and won the top prize at the World Solar-Car Rallye.

Suddenly, the GW motor was the hottest thing in the solar car world. Other teams lined up for a look, and many wanted to know if they could buy one. The motor the team had created, based on the Australian model, used a rare design called "axial flux." Most motors are configured somewhat like a hot dog. The meat consists of magnets mounted on a rotor. The bun is called the stator and is lined with electrified coils. When electric current passes through the coils, it creates a magnetic field that repels the magnets on the rotor, making the rotor turn.

The axial flux motor is more like a pair of pancakes, with the rotor and stator shaped like disks and set next to one another. The gap between the two disks determines how powerfully the rotor turns, a force called torque. On a vehicle, you need a lot of torque to start rolling or to get up a hill, but you don't need much to coast along at a constant speed. That's why ordinary cars start in high-torque first gear and cruise in fifth.

Eventually, the engineers from GW figured out how to change the size of the gap while the motor was operating to respond to different demands. For high-torque acceleration, the coils are very close to the magnets on the rotor; for cruising speed, they back off. That helps the motor use energy efficiently, and allows a smaller motor to achieve a wider range of performance. Combined with a computerized controller, the motor has converted up to 95 percent of available energy into output, compared with about 85 percent efficiency for a typical electric motor and between 30 and 35 percent for a gasoline engine.

Bedewi, Jermakian and another adviser realized they had something with commercial potential. In 1995, they formed New Generation Motors Corp. to begin making the motors for other solar car competitors. Today, virtually all the top U.S. university teams use New Generation products, including last year's U.S. champion, the University of Missouri-Rolla.

But from the very beginning, Bedewi, in particular, was thinking beyond the solar car market. The motors could work with any electric power source, and could be made big enough to run a family car or even a city bus. Bedewi was thinking of Detroit.

In 1997, he took a sabbatical from the university to devote himself full time to the company. The path seemed to be opening up. Ford's Ross, one of the most powerful men in the auto industry, was advising Bedewi and encouraging him to aim high. California had caused a storm among carmakers by demanding that 10 percent of all vehicles sold there have zero emissions, forcing the production of electric cars. The U.S. government was funding electric car research in Detroit. It was a guaranteed market. And the Internet boom was taking off. Technology was in the news; companies were flush with cash and an optimism that everything was possible. Bedewi wrote a business plan that projected revenue of $500 million a year within a decade. He envisioned New Generation electric cars rolling off an assembly line in Northern Virginia.

It looked like a sure thing.

Posters from the solar car races line Jermakian's office at New Generation. Over his desk is a picture of Albert Einstein with the quotation, "In the middle of difficulty lies opportunity." Behind mounds of papers, with Diet Dr. Pepper and a Hewlett-Packard calculator at the ready, Jermakian calls up a program on his Dell workstation. Colorful three-dimensional drawings flash across his screen, swoops and pie shapes that simulate the inside of a working electric motor. Cascades of yellow and green represent the flow of electricity. The colors change as Jermakian changes small details of the motor's design, looking for ways to make it more efficient.

It's late afternoon, but there's a midnight hush at New Generation, with the hum of computer fans the only noticeable sound. An engineer sits in a cubicle outside Jermakian's office; a technician works quietly at a bench in the workshop. The company employs nine full-timers these days. At its peak, a few years ago, 35 people worked there. New Generation rents space in one of those countless strip-mall-style office parks near Dulles Airport in Ashburn -- servants' quarters compared with the deluxe plantations of AOL, MCI and the like nearby. Next door is a lawn care service.

The offices are small and cluttered, their walls plastered with engineering diagrams. Bookshelves bulge with reference manuals and notebooks of technical data. A small kitchen pantry contains a sink, a microwave, a coffee machine, a dorm-size refrigerator and the main server for the computer system. Down a short hallway from the offices, the workshop is a wide, high-ceilinged garage space with two roll-up bay doors leading outside. Colored power cables snake across the scuffed concrete floor, threading through a maze of equipment and workbenches. Many of the work areas have ancient wooden tabletops, hash-marked from years of use. Hulking machines for shaping metal sit silent, but small piles of silvery shavings show they've been running recently. Motor parts -- coils of wire, aluminum housings -- are scattered throughout, and tucked among the equipment are half-built scooters and golf-cart-size vehicles used for testing motor designs.

It's a place that seems indifferent to outsiders. Anyone who got sucked into a campus organization in college -- a student newspaper, a radio station, a theater troupe -- will recognize it for what it is: the lair of a bunch of engineering and science majors who never wanted to give up the lab from freshman year.

Only four of the original GW solar car team members remain at New Generation, but the bonds they formed during the races are what shape the company. Jermakian, at 41 one of the oldest people on staff, guards the idealism that pulled them all together in the first place. He's chief technology officer, the guy whose job is to think big thoughts and keep working at the motors to make them better. When he speaks, he tends to cross his arms, lean his head back and shut his eyes, as if plumbing deep for an answer. Unlike most of his colleagues, Jermakian is a physicist by training, not an engineer, and his more abstract style is a contrast to Takamura's clipped and direct manner. Takamura, 34, runs day-to-day operations and has to think about both the science they're trying to do and the budget that will, somehow, keep it going.

Takamura's effort to get outside investors to put money into the company bothers Jermakian a bit. He knows the stakes -- they need to get through the coming year until revenue from their biggest venture starts flowing. But every outsider who writes a check will chip away at the company's freedom to chart its own course. Jermakian would rather chase federal grants for support, and tonight he'll take home an application for a government research contract that could bring in $200,000. He's put too much into New Generation to hand it over to someone else now. Jermakian quit his career at NASA to help start the company. The years of 80-hour work weeks contributed to the breakup of his marriage, and Jermakian has gone without pay for months at a time, living off savings and credit cards. If he had known what it would be like, he might not have done it. But every time he doubts he can keep going, he remembers Takamura in the Australian Outback at the end of each grueling leg of the solar car race, patiently remounting tire after tire in the back of a pickup, his face and arms black with insects.

This is not just a job. Jermakian says the technology can improve the world by making nonpolluting cars that use less energy. Toyota's gas-electric hybrid, the Prius -- which Jermakian is on a waiting list to buy to replace his 1995 Saab -- would be a far better car with a New Generation motor, he believes. "We have the most efficient drive system in the world," he says. If more people had access to the technology, "it would be life-changing for everybody in the company and hopefully would change the way everybody thinks about these things."

Takamura has no less faith in the technology or in New Generation's ability to chart its own course. But he knows survival involves compromise.

It is a business, after all. It has to make money.

While Jermakian works on his motor simulation program, Takamura sits in his own office on the opposite side of the building, filling out forms for the same venture capital group whose chat session he grudgingly attended several days before. He has to write a description of the company to try to interest investors, and the catch is that it has to fit into a strict format. Every section is limited to a certain number of characters -- that's letters, spaces and punctuation marks. It's better than cocktail party chitchat, but still. For "Company Background" -- the nine years of work and worry, sunny days spent under fluorescent lights, nights away from family -- he gets 882 characters. For "Technologies/Special Know-How" -- the years of refining the motors, the experiences of the solar car races, the two patents issued and five patents pending -- 315 characters. It's like writing haiku, with the future at stake.

Jermakian would be hard put to distill everything into such stark terms, even if he wanted to. But Takamura gamely pecks away at his keyboard. Since the earliest undergraduate days of the solar car team, and since joining New Generation in 1996, Takamura has been the guy who volunteers for the grubbiest jobs. At the company, someone has to juggle everyone's schedules, coordinate purchasing, look for ways to raise money. It's usually him. As his responsibilities have increased, his pay has actually declined, thanks to the company's dwindling cash flow. He often uses his own credit cards to buy office supplies or plane tickets, then gets reimbursed when New Generation has the money. Until his son was born two years ago, Takamura regularly spent up to 80 hours a week at the office. Now, with his wife often out of town doing the software consulting work that supports them, he sticks to a schedule nearly as grueling as a cross-country solar car race.

Takamura leaves home in the District every morning at 7:30, if the nanny shows up on time, then works at the office until about 4:30 so he can be back to let the nanny leave by 6. He makes dinner for his son, and the two hang out for a few hours. Takamura then stays up into the night finishing projects for work. If he's writing proposals for supplying motors to other companies, or going back through motor design changes, he might toil away until 5 a.m.

His wife, Holly, who went to GW with him, says he doesn't complain much about the grind. "He's always really liked what he does. There are not a lot of people who can say that." When they married in 2000 after many years together, Holly insisted that he cut back on his hours at the office so they could have some time together. But she has accepted the company's central role in his life. He'll never give it up. "Eric's gonna be with it 'til they run out of money," she says, "And then maybe another six months."

Now, left hand on chin, right knee bouncing under the desk, Takamura sits stumped by the last entry on the company profile. A pack of Camel Lights and a green lighter wait next to the keyboard. "Use of Funds," 126 characters. If investors give them money, they'll hire more people, buy equipment, find new markets, develop new uses for the motors. He's gotten the explanation down to 128 characters, but how to get two more letters out?

There it is: "sales and marketing" becomes "sales & marketing." Done.

Takamura sees this as a good exercise. He has to be clear-eyed about the company to help it survive. Jermakian's vision for what they can do with the technology is important, but by itself won't get them anywhere.

Vision can be a dangerous thing in the auto industry. It tends to blind people to the cruel realities of the market. Vision leads to grand failures such as Preston Tucker's 1940s effort to build a stylish car with unheard-of safety features, which drew attacks from industry and government, and John Z. DeLorean's stainless-steel gull-winged cars, which live on primarily in those "Back to the Future" movies. The world is littered with martyrs to the car culture.

In Columbia, on the other side of Washington from New Generation Motors, Bob Beaumont knows the nameless frustration of being absolutely right and having the world turn away. In the 1960s, Beaumont was a chain-smoking, Elvis-haired car dealer in upstate New York when he had an epiphany: Gasoline engines were ruining American cities. What everyone needed, he decided, was a two-seat electric commuter car. In the early 1970s, he set up a company in Florida to manufacture a glorified golf cart called the CitiCar, which could go about 40 miles at 30 mph on a single charge. For a while, it looked like Beaumont was on to something. He got coverage in Popular Mechanics and the Wall Street Journal, and in three years, he sold more than 2,500 CitiCars.

In 1975, though, Consumer Reports printed a scathing article about the CitiCar's safety, and Beaumont says sales dried up almost overnight. He retired to Columbia and opened a used-car dealership.

In the early 1990s, Beaumont got his old team back together. Like Jermakian and Bedewi at New Generation, Beaumont was watching California's push for electric vehicles and decided that the world was finally catching up to him. He went back to Florida, this time to build a sexy little electric runabout called the Tropica, with a list price of $17,000. The car was so eye-catching that actor Don Johnson bought one and used it in his "Nash Bridges" television series. But the company had financing problems and disagreements about management, and it folded in 1996 after producing only 14 Tropicas. One of them sits in Beaumont's Columbia garage, its white body streaked with dust and its seats hidden under a pile of old lawn chair cushions.

Beaumont, 71, insists he isn't bitter. Suffering from emphysema and other ailments, he says he's made his last run at the auto industry and is content to spend his days reading books. But he still can't believe it didn't work. "I was so positive, and still am -- you could move millions of people this way with no pollution. Cheaply." He laughs. "Here I am, one guy, trying to convince the world to do this. What a schmuck . . . There's an opportunity to make millions of these things. Why they don't do it, I don't understand."

One person who can tell him why is Larry Burns, who as head of research and development for General Motors is the technology gatekeeper for the biggest automaker on the planet. "I've seen examples where there's an idea and a promise and a potential and an entrepreneurial spirit, and it's really just tough for the companies to find the market," Burns says.

He has 600 automotive scientists at labs around the world, plus an "intelligence" squad of about 30 people who spend all their time scouting fresh ideas from small companies or backyard geniuses. One outsider that managed to break into GM with a new idea is right here in Washington: XM Radio, the satellite radio service that's now on more than 1 million GM cars and trucks.

Sometimes GM will take an ownership stake in little-guy ventures, if the research is aimed at a specific area the corporation wants to pursue. For instance, it owns 20 percent of Quantum Technologies, a small California company that makes compressed gas storage systems that could be used for hydrogen-powered vehicles.

But for GM to be interested in something, the product has to appeal to the corporation's quest for that basic quality that makes the auto industry so alluring in the first place: volume. GM isn't looking for technologies that might sell a few thousand cars or trucks. Even the success of the Prius, which is selling about 50,000 units a year, doesn't do much for GM, which has delayed rolling out its own hybrids in a bid to develop a bigger market. GM only cares about selling millions of vehicles per year. Toyota, the most profitable automaker, spent years losing money developing the Prius and selling it at a loss; GM and the other American manufacturers don't have that kind of money.

Burns says most small entrepreneurs don't survive long enough to get their ideas to the point of mass production. First, you have to have money to sustain you while you're developing the technology, because you're not yet producing anything that can pay the bills. Then you have to figure out how to turn a hand-crafted laboratory device into something that can be mass-produced safely and cheaply, and invest in the equipment to make it. Finally, you have to figure out how to market the product, which takes yet more money.

New Generation has raised just over $4 million in its nine years, all cobbled together from family members, friends and associates still patiently waiting for a return on their investments. The company has won a little more than twice that amount in grants and research contracts over the years from other corporations or the government. That's pocket change in terms of the 1990s Internet boom; a fashion Web site called Boo.com, for instance, burned through something like $23 million a month during its brief existence several years ago.

But New Generation never bothered with the frippery of the Internet age, the flashy Web sites or publicity stunts. Instead it tried to pour everything into its technology, in the belief that the motors were so good and the market was so clear that success was just a matter of staying on track. Early on, the company had promising avenues into Ford and with a company called Global Electric Motorcars, or GEM, which makes "neighborhood electric vehicles" -- snazzy golf carts for gated communities or large estates.

But the dream of big success shattered almost as quickly as that first solar car in Australia.

First, New Generation's contact at Ford, Lou Ross, died unexpectedly in 1998, depriving the company of its primary patron in the auto industry. Then California, under relentless pressure from automakers that did not want to bother with building a few thousand electric vehicles, backed off its zero-emissions mandate. Government and industry research switched almost completely away from electric-powered cars and into the pursuit of hydrogen fuel cells -- a technology that could be decades away from practical use.

GEM was bought by DaimlerChrysler and quit working with New Generation. Then the Internet bubble burst, and terrorists attacked on September 11, 2001, and the economy went sour, and all the money and confidence flowing into new technologies of all sorts simply dried up.

Bedewi, 40, gave up on his teaching sabbatical in 2001 and went back to GW full time. He's still chairman of New Generation, and through direct investment and debt has about a $500,000 stake in the company. But he now spends most of his time in the classroom and running the crash analysis center. He and the others haven't given up on the dream, though, just scaled it back. New Generation has found an unusual alternate route that, if it works, might let the company achieve something almost as rare as big success: survival.

While the United States is the world's biggest market for automobiles, there are far bigger potential markets out there that the industry is determined to tap. China is tops, but India, with its 1 billion-plus population and burgeoning middle class, is not far behind. In some ways, India is better suited than the West to experimenting with transportation technology. There is no ingrained car culture, no tradition of unlimited personal mobility.

Years of using primitive gas and diesel engines have left India's cities choking on pollution, so the government has required that public transportation in cities switch to using alternate fuels. Natural gas is popular at the moment, but rechargeable battery-powered vehicles are well-suited to the type of driving done in India -- limited speed, an average range of 40 miles a day or less and lots of stop-and-start traffic. The energy generated by braking, which on a gas-powered car is lost as heat, can be captured to help recharge a battery.

India's Bajaj Auto Ltd., one of the world's biggest producers of scooters and three-wheeled autorickshaws for crowded city streets, went looking in the late 1990s for a company that could help it convert part of its product line to electric propulsion. It teamed up with New Generation through the U.S. Agency for International Development.

New Generation agreed to develop small, cheap motors for autorickshaws. Over seven years, Bajaj invested some $3 million into new machines and tooling so its factory could build the motors. New Generation stripped itself down to the bare minimum number of employees to concentrate almost solely on the autorickshaw job, hiring Sethi, 32, last year and grooming him to move to India to be a full-time liaison with Bajaj.

"Their technology is unique," says Tapan Basu, Bajaj's head of production. "We're always looking out there. We haven't been able to duplicate this with anybody else . . . We had an exciting chance of working with an advanced technology that was not sewn up by the Big Three."

The partnership is close to a turning point. A test fleet of five electric auto-rickshaws carries visitors within the Taj Mahal complex, which is off-limits to internal combustion engines. If the first 60 production models sell well later this year, for roughly $2,500 apiece, Bajaj will build 1,200 next year and aim for as many as 12,000 in 2006. That year, New Generation expects to become profitable for the first time.

For New Generation, everything hinges on surviving until production starts next spring. Because the Bajaj effort has been so all-consuming, the company doesn't have much else to sustain it, and the financial tanks are running dry. The company has considered firing almost all remaining employees and putting the company in "hibernation," but it wants to be ready to take off on other projects as soon as the Bajaj revenue starts flowing. That's why the push is on to get investors or grants for the coming year.

If New Generation can get past the production start-up, wild possibilities open up. Bajaj could decide to put electric motors on two-wheeled scooters; it builds 1 million of those a year. New Generation also has the right to buy its own time at the Bajaj factory, and could use it to build motors for other clients in other locations. With a product geared for mass production and the means to crank it out on a large scale, New Generation could finally be on a path that leads, realistically, back to Detroit.

As the company has worked its way to this position, though, the rest of the world has not stood still. Bodine Electric Co., based in Chicago, now produces a competing pancake-style motor that made a strong showing at last year's American Solar Challenge cross-country race. And not 10 minutes from New Generation's offices in Ashburn, another company has emerged that could pose a serious threat -- WaveCrest Laboratories, which also makes electric motors that fit into the hub of a wheel, and also claims to have technology far more efficient than that of the typical motor.

It's hard to compare the two companies' products, because both WaveCrest's and New Generation's are still in the laboratory stage. WaveCrest's advances are largely based on the software used to control the motor, instead of on the configuration of the motor itself. The company has taken a different approach to the market, as well. With the generous backing of a single angel investor -- Alan Anderson, who the company says has a background in software and biotech ventures -- WaveCrest has taken a route more similar to the classic Internet startup, spending millions to establish a name as well as a technology. Retired Gen. Wesley Clark served on the board of directors until embarking on his failed presidential bid. Former Chrysler product development chief Richard Schaum, a well-known figure in Detroit, is now a WaveCrest vice president.

The company makes no bones about its goals: It wants to take Detroit by storm. Last year it opened an office there. To help attract visitors, the company brought a futuristic Smart car from Europe -- a two-seat commuter roadster -- and replaced its engine with electric wheel-hub motors to show off the technology's capabilities.

At 120 employees and climbing, WaveCrest has a 30,000-square-foot headquarters building with landscaping, a receptionist and security badges for visitors. Its laboratories are big-league all the way, with pristine metal-topped worktables, battleship-gray floors and clean white walls. The engineers even wear crisp white lab coats.

Where New Generation's one successful commercial product -- solar car motors -- is sold to a small group of in-the-know academics, WaveCrest recently went commercial with a tough-looking motorized mountain bike. At $2,400 each, the bikes have begun attracting notice from hobbyists, police departments and the military. They're great calling cards for a company looking for attention.

It's impossible to say whether WaveCrest will be any more successful than New Generation at breaking into the industry in Detroit, or even in achieving that lesser goal, survival. The two companies have checked one another out, even talked of partnership. But WaveCrest president and chief executive Roy Barbee dismisses New Generation now.

"I think they've come upon fairly hard times," he says. "I would say we are ascendant. I'll leave it to you to decide if that's what they are."

It's a Friday night three weeks after that first venture capital get-together. Takamura is at his Capitol Hill rowhouse, cooking pasta with sausage marinara as his wife and 2-year-old son sit at the kitchen island.

This is rare family time. Holly Takamura has been spending weekdays in Florida on a consulting job with her software company. Little Kalani, with a mop of dark hair, sits close to his mom as they draw pictures with crayons. Once upon a time, Friday night would have meant going out to clubs with friends. Holly Takamura still makes it out once in a while, and winds up telling her husband what he missed. Eric Takamura is too absorbed in his work. A year ago, he cut off his shoulder-length black hair. It was time to face up to what his life has become.

Takamura has at least tasted the kind of life he wanted, the French-uncle life. He goes to India a few times a year. He's been to Egypt, trying to work out a deal to put motors on buses. And New Generation itself definitely counts as uncharted territory.

Word has come back from the venture capital group. New Generation has been invited to return for the next phase of competition for funding. Tonight, after Kalani goes to bed, Takamura will stay up late filling out the final application. In a few weeks, he'll make his formal pitch to the potential investors, then wait to see if anyone steps forward with an offer of money.

Jermakian, too, has gotten good news. He won the government research contract for $200,000. Things seem to be looking up.

Still, it doesn't surprise Takamura to hear that Barbee believes WaveCrest is ascendant and New Generation is faltering. He shrugs. "If you're talking about who has the most people working for them or the best facilities, then, yeah, they're ascendant," he says. "But in terms of technology" -- he frowns and shakes his head -- "ours is better."

After all this time, he still believes. Maybe he has to, because that's the only reward for so much effort. At one point, he figured he could have made the same amount of money by working the same number of hours at McDonald's.

"If things keep going, this will have a dramatic impact on the road for people to come -- like the little one here," he says, gesturing at Kalani. "There's satisfaction in that. But I'm not sure how satisfying it would be if it failed."

Greg Schneider writes about the auto industry for The Post's Business section. He will be fielding questions and comments about this article Monday at 1 p.m. on www.washingtonpost.com/liveonline.