John Doerr was crying. The billionaire venture capitalist had come to the end of his now-famous March 8, 2007, TED talk on climate change and renewable energy, and his emotions were getting the better of him. Doerr had begun by describing how his teenage daughter told him that it was up to his generation to fix global warming, since they had caused it. After detailing how the public and private sectors had so far failed at this, Doerr, who made his fortune investing early in companies that became some of Silicon Valley’s biggest names—Netscape, Amazon.com, and Google, among others—exhorted the audience and his peers (largely one and the same) to band together and transform the nation’s energy supply. “I really, really hope we multiply all of our energy, all of our talent, and all of our influence to solve this problem,” he said, falling silent as he fought back tears. “Because if we do, I can look forward to the conversation I’m going to have with my daughter in 20 years.”

As usual, Doerr’s timing was perfect. Just weeks earlier, Al Gore’s An Inconvenient Truth had won an Oscar for best documentary. (Gore is now a partner in Doerr’s green tech team at the VC firm Kleiner Perkins Caufield & Byers.) Interest in climate change had never been higher. And as the economy recovered from the dual shocks of the Internet bubble and 9/11, Doerr’s fellow Silicon Valley VCs were already looking to clean technology as the next big thing. What followed was yet another Silicon Valley gold rush, as the firms on Sand Hill Road were pulled along by the promise of new fortunes and the hope that they would be the ones to wean America off of fossil fuels. The entrepreneurs and tech investors who had transformed media and communications were ready to make Silicon Valley the Saudi Arabia of clean energy.

Never mind the fact that green technology had been struggling to achieve critical mass for decades. “You had folks who came in with the hubris to say, ‘I know these guys have been working on this for 50 years,'” says Andrew Beebe, chief commercial officer for Suntech, the Chinese solar manufacturer. “‘But I’ve got $50 million and I can blow the doors off this thing.'”

In 2005, VC investment in clean tech measured in the hundreds of millions of dollars. The following year, it ballooned to $1.75 billion, according to the National Venture Capital Association. By 2008, the year after Doerr’s speech, it had leaped to $4.1 billion. And the federal government followed. Through a mix of loans, subsidies, and tax breaks, it directed roughly $44.5 billion into the sector between late 2009 and late 2011. Avarice, altruism, and policy had aligned to fuel a spectacular boom.

Anyone who has heard the name Solyndra knows how this all panned out. Due to a confluence of factors—including fluctuating silicon prices, newly cheap natural gas, the 2008 financial crisis, China’s ascendant solar industry, and certain technological realities—the clean-tech bubble has burst, leaving us with a traditional energy infrastructure still overwhelmingly reliant on fossil fuels. The fallout has hit almost every niche in the clean-tech sector—wind, biofuels, electric cars, and fuel cells—but none more dramatically than solar.

Doerr’s TED talk wasn’t the start of this VC-fueled drive for a new-energy economy. Rather, it was a product of a transformation that was sweeping Silicon Valley. Many of the investors and entrepreneurs who had ridden the Internet bubble to various levels of success had already started pouring money and ideas into clean tech.

One of the first to bet big was Martin Roscheisen. He sold his email-management firm eGroups to Yahoo for $450 million, and in 2002 he cofounded Nanosolar, a panel manufacturer. But that was just the beginning. Vinod Khosla, cofounder and former CEO of Sun Microsystems, moved his VC firm, Khosla Ventures, heavily into biofuels and other renewables. Beebe, cofounder of the dotcom-era darling Bigstep.com, a web-hosting company, helped start the solar panel maker Energy Innovations in 2003. Arno Harris, who had helped steer what he now calls “an Amazon-Kleiner Perkins online wine store that left a big hole in the ground,” worked with Beebe at a subsidiary of Energy Innovations before founding Recurrent Energy, a company that develops utility-scale solar projects, in 2006. And PayPal cofounder Elon Musk has put $96 million of his own money into the electric-car startup Tesla Motors and was joined by well-known VCs Steve Jurvetson and Nancy Pfund.

In 2008, by which time Kleiner Perkins had allocated more than $300 million to clean tech, the firm launched a $500 million growth fund that it said was “intended to help speed mass-market adoption of solutions to the world’s climate crisis.” Doerr, who told Forbes that curbing climate change was “the largest economic opportunity of the 21st century, and a moral imperative,” helped direct money to everything from solar to smart meters.

These investors were drawn to clean tech by the same factors that had led them to the web, says Ricardo Reyes, vice president of communications at Tesla Motors. “You look at all disruptive technology in general, and there are some things that are common across the board,” Reyes says. “A new technology is introduced in a staid industry where things are being done in a sort of cookie- cutter way.” Just as the Internet transformed the media landscape and iTunes killed the record store, Silicon Valley electric car factories and solar companies were going to remake the energy sector. That was the theory, anyway.

The major energy bills that passed in 2005 and 2007—which provided tax credits and loan guarantees for clean tech—gave investors further confidence. Venture capital in solar alone rose from $32 million in 2004 to nearly $1.85 billion in 2008. Investment in battery tech rose more than 30-fold during the same period.

Other clean-energy sectors were thriving as well, buoyed not only by VC money but by the fact that the average price of electricity, which had been stable for years, shot up 35 percent between 2002 and 2008. At the end of 2006, the total capacity of all the wind turbines installed in the US was 11,468 megawatts, enough to power 3.2 million homes. By 2010, it was nearly four times that much. “As more entrepreneurs and innovators saw there was capital available in the clean-energy sector, you saw more folks looking into developing solutions and business around that,” says Joshua Freed, vice president for clean energy at the think tank Third Way. “There was a virtuous circle of capital moving to clean energy, and entrepreneurs moving to clean energy because there was a capital.”

One of these was Chris Gronet, a Stanford PhD in semiconductor processing who had been general manager of the thermal processing group at Applied Materials, a firm that provides equipment and software to semiconductor and solar companies. He had come up with a design for a revolutionary new solar module (a module is a light-gathering photovoltaic cell with all the attendant structural hardware and circuitry) that he believed would be vastly more efficient than the flat-panel modules that had dominated the market for more than three decades.

Conventional photovoltaics are tricky things to install. Under the best conditions—when their surfaces are clean and aimed directly into the sun—they generally operate at no better than 20 percent efficiency, meaning that they convert just a fifth of the energy striking them into electricity. But an immobile flat surface faces the sun head-on for only a brief period each day, at best. And simple dust can reduce the efficiency by 5 to 10 percent. Furthermore, flat panels’ vulnerability to wind poses numerous structural challenges—from mounting hardware to rooftop integrity. Solar firms routinely employ aeronautical engineers to deal with this issue, and VCs looking to get into the sector sometimes brought these experts on board to help judge whether a startup’s product could withstand intense wind patterns.

Gronet’s design called for a grate made of rows of cylindrical cells rather than a single panel of flat cells. The sun tracking across a cylinder will always be shining directly on part of it. That meant Gronet’s modules could be mounted parallel to a roof and out of the wind, rather than angled up into it. As an added bonus, the tubular cells would gather not just direct sunlight but also ambient light reflected off of the rooftops on which they were mounted.

At around this time, investors were searching for an alternative to the crystalline silicon used in photovoltaics, which was skyrocketing in price. As more and more manufacturers had been getting into making solar panels, increased demand had driven the price of processed silicon from around $50 per kilogram in 2004 to well above $300 by 2008. When the higher production costs were factored in, the price of electricity from solar firms was 17 to 23 cents per kilowatt-hour, even after subsidies. That was about twice the average price of conventionally produced electricity at the time.

Gronet’s design called for a mix of copper, indium, gallium, and selenium, or CIGS, instead of crystalline silicon. Though slightly less efficient than silicon in direct sunlight, CIGS performs better under cloud cover and in variable light. The technology had been around for several years but was too expensive to be practical. That changed as soon as silicon climbed above $200 per kilogram. Suddenly CIGS could compete. With his cylindrical module and exotic coating, Gronet had a model for transforming the solar industry. He incorporated his company in 2005, first calling it Gronet Technologies but quickly changing the name to Solyndra.

Gronet and his chief financial officer, Jonathan Michael, set out to raise capital for a factory. By 2007, they had $99 million from sources including RockPort Capital Partners and Argonaut Private Equity and were busy renovating an old Hitachi building in Fremont, California. In 2008, Virgin Green Fund, an investment arm of British business icon Richard Branson, chose Solyndra as the only solar company that it would put money into, out of more than 100 that applied for funding. By the end of that year, Solyndra had raised $600 million, boasted more than 500 employees, and had two major orders—$325 million from Sacramento-based Solar Power and $681 million from a German company called Phoenix Solar. “Everyone was pretty optimistic,” recalls Lindsey Eastburn, who was designing factory-automation software for Solyndra. “We were making product, and we were selling it.”

Just as Solyndra was starting to take off and needed more money for expansion, the venture capital climate began to cool. The 2008 financial collapse erased a quarter of the gains VC firms had made between 2003 and 2007, and the sudden paucity of capital—combined with the difficulty of taking smaller companies public—hit renewable startups particularly hard. Venture investments in clean tech fell from $4.1 billion in 2008 to $2.5 billion in 2009.

There was an additional factor at work: impatience. Venture capitalists tend to work on three- to five-year horizons. As they were quickly finding out, energy companies don’t operate on those timelines. Consider a recent analysis by Matthew Nordan, a venture capitalist who specializes in energy and environmental technology. Of all the energy startups that received their first VC funds between 1995 and 2007, only 1.8 percent achieved what he calls “unambiguous success,” meaning an initial public offering on a major exchange. The average time from founding to IPO was 8.3 years. “If you’re signing up to build a clean-tech winner,” Nordan wrote in a blog post, “reserve a decade of your life.”

The truth is that starting a company on the supply side of the energy business requires an investment in heavy industry that the VC firms didn’t fully reckon with. The only way to find out if a new idea in this sector will work at scale is to build a factory and see what happens. Ethan Zindler, head of policy analysis for Bloomberg New Energy Finance, says the VC community simply assumed that the formula for success in the Internet world would translate to the clean-tech arena. “What a lot of them didn’t bargain for, and, frankly, didn’t really understand,” he says, “is that it’s almost never going to be five guys in a garage. You need a heck of a lot of money to prove that you can do your technology at scale.”

Luckily for the clean-tech industry, a much larger investor stepped in to replace the retreating VCs—the federal government.

In 2005, Congress created a federal loan guarantee program as part of the Energy Policy Act, which initially was authorized at $4 billion. Though ostensibly set up to promote nonpolluting energy sources, it was, like most federal slush funds, created by a politician (in this case, former Republican New Mexico senator Pete Domenici) to help a specific industry (in this case, nuclear energy). But the expected nuclear renaissance never happened; the private market was unwilling to finance plants that cost billions to build, created toxic waste, and ran into all the NIMBY hurdles that come with nuclear energy. So the door was open for applications from other clean-energy sectors.

While solar projects would ultimately receive more than three-quarters of the program’s financial support, the list of recipients included everything from a wind farm in Oregon to a cellulosic ethanol plant in Kansas. But by the time Bush left office, not a penny had been distributed. Most of the applications, including one from Solyndra, were still wending their way through the approval rounds at the Department of Energy. There were only 16 employees tasked with sorting through the applications and relevant data, and the loan program was more of a theoretical construct than an engine of economic activity.

Then Obama took office, and the loan program suddenly had an administration committed to using federal dollars to stimulate what it referred to repeatedly as “the clean-energy economy.” For Democrats, the concept of clean energy hit every button there was to push: It addressed the looming problem of climate change, offered a domestic source of electricity and fuel, and promised new jobs in a shaky economy.

The Department of Energy, which for decades had focused on managing nuclear waste and weapons and doling out subsidies to the fossil fuel industry, had a new leader—Steven Chu, a renowned physicist and Nobel laureate—and a fresh mandate.

The money the federal government delivered dwarfed what VCs had put into clean energy. The loan guarantee program alone provided a little more than $16 billion for 28 projects. The government pumped an additional $12.1 billion into the sector through tax credits. All told, federal subsidies for renewable energy nearly tripled between 2007 and 2010, rising from $5.1 billion to $14.7 billion. The federal largesse also made clean tech look like a safer bet to the VC world, whose investments rebounded after the 2009 dip.

Solyndra’s $535 million loan guarantee closed in September 2009. The firm had no problem putting the funds to use, starting construction on a second factory, expanding its workforce to 1,100 employees, and paying millions for a custom machine designed to put the finishing touches on the cells at a rate of 60 per minute. As part of an ongoing “Main Street tour” highlighting the nation’s manufacturing prowess, Obama scheduled an appearance at the Solyndra factory in May 2010. After a tour of the facilities, the president gave a speech on the factory floor in which he called Solyndra “an engine of economic growth.” “The future is here,” he added.

By fall of 2010, Solyndra had scuttled plans for a $300 million IPO and was still waiting to hear back on an additional $469 million loan application, filed just days after the first loan was approved, to help finance its second factory. While the company’s solar modules were working as planned, Solyndra needed to increase its production capacity to get per-unit costs down. The custom machine had turned out to be a dud. Despite months of work by a team of engineers sent over from the Dutch company that had built the two-story-tall behemoth, it was struggling to reach its expected output. When all the costs were factored in, a Solyndra module cost at least 30 percent more per watt than a traditional photovoltaic, and the gap was growing. Unless Solyndra got faster and cheaper, there was no way it would be able to compete.

Given the concerns about Solyndra’s financial viability, the company agreed with DOE officials to drop the request for a second loan. Yet in early 2011, despite further warnings about Solyndra’s cash-flow issues, the DOE agreed to restructure the original loan, with a provision guaranteeing that private investors, not the federal government, would be repaid first in the case of a default. It was a decision that the Obama administration’s critics would make much of within a matter of months.

Solyndra’s failure wasn’t just the result of manufacturing problems. It was also a product of a broad shift that was happening in the US energy sector. The financial models that had justified the massive investments in clean-energy sources were built on assumptions that the price of fossil fuels, in particular natural gas, would continue to rise. But those models began to fall apart as a natural gas boom transformed the energy landscape.

As with the Internet bubble, and the more recent housing bubble, there were signs of trouble. In fact, in the weeks and days leading up to Obama’s visit to the Solyndra plant, officials at the Office of Management and Budget were issuing warnings. “I am increasingly worried that this visit could prove embarrassing to the Administration in the not too distant future,” wrote one OMB official.

In fact, though Solyndra CEO Brian Harrison painted a rosy picture for lawmakers in July 2011—boasting that revenue “grew from $6 million in 2008 to $100 million in 2009 to $140 million in 2010” and would nearly double in 2011—the truth was laid out in an internal White House memo obtained by The Washington Post after Solyndra filed for bankruptcy. The August 2011 memo, written days before Solyndra went bankrupt, stated simply that “the company has had 0 percent sales growth since [fall] 2009.”

Perhaps the biggest force working against not just Solyndra but clean energy in general is this: Because natural gas has gotten so cheap, there is no longer a financial incentive to go with renewables. Technical advances in natural gas extraction from shale—including the controversial practice of hydraulic fracturing, or fracking—have opened up reserves so massive that the US has surpassed Russia as the world’s largest natural gas supplier.

The price of natural gas peaked at nearly $13 per thousand cubic feet in 2008. It now stands at around $3. A decade ago, shale gas accounted for less than 2 percent of America’s natural gas supply; it is now approaching one-third, and industry officials predict that the total reserves will last a century. Because 24 percent of electricity comes from power plants that run on natural gas, that has helped keep costs down to just 10 cents per kilowatt-hour—and from a source that creates only half the CO2 pollution of coal. Put all that together and you’ve undone some of the financial models that say it makes sense to shift to wind and solar. And in a time of economic uncertainty, the relatively modest carbon footprint of natural gas gets close enough on the environmental front for a lot of people to feel just fine turning up the air-conditioning.

Solyndra’s Epic Missteps

From Chinese competition to the color of customers’ roofs, the solar manufacturer made assumptions that proved disastrously wrong.—R.S.

Ramp-Up Costs

Gearing up to manufacture a new consumer product is notoriously expensive. In the energy sector, the costs can be crushing, as Solyndra found out: It spent at least $87 million to outfit its first factory and get to market, $290 million in research and development, and $733 million on just the first phase of its second factory, which was necessary to manufacture at the required scale. Per watt, Solyndra’s projected prices were up to double what consumers can now pay for conventional solar power.

Silicon Prices

Traditional solar panels are made from silicon. Solyndra’s next-gen design used CIGS—a combination of copper, indium, gallium, and selenium. When Solyndra launched, processed silicon was selling at historic highs, which made CIGS a cheaper option. But silicon producers overreacted to the price run-up and flooded the market. Prices dropped by as much as 90 percent and stayed there. Solyndra’s business model was based on a price advantage for CIGS that no longer existed.

Shale Gas Output

In 2001, shale gas accounted for less than 2 percent of US natural gas output. Today, thanks to advances in horizontal drilling and the effective though highly controversial technique of hydraulic fracturing, or fracking, it accounts for 30 percent. Meanwhile the price of natural gas has fallen by 77 percent since 2008, and the cost of producing electricity in gas plants is down 40 percent since then. Renewables simply can’t compete.

Chinese Supply

In 2010, China established a $30 billion line of credit for the nation’s solar industry as part of a strategy to bolster domestic production. The result: Chinese firms went from making just 6 percent of the world’s solar cells in 2005 to manufacturing more than half of them today. The US share has plummeted from 40 percent to 7 percent. Solyndra and other manufacturers were simply price out of the market.

Rooftop Colors

Solyndra’s model assumed that its cylindrical cells would generate 15 percent more energy per square foot than flat crystalline-silicon cells. This math assumed that the cells would be installed on white roofs, where their sides and bottoms would absorb reflected light. The company hoped to forge partnerships with roofing companies to facilitate this—and to open new sales channels—but was unable to do so in sufficient numbers.

Another blow to the domestic clean-tech industry was a glut of processed silicon that sent prices back down below $30 a kilogram. That price, combined with the technological simplicity of manufacturing conventional solar panels, opened the door to relatively unsophisticated operators. For example, in 2007, a Chinese textile manufacturer approached Arno Harris, CEO of utility developer Recurrent Energy, to see if he’d be interested in buying solar panels that they hoped to begin making. When the bar to entry is so low that textile makers can churn out solar modules, Solyndra’s expensive CIGS-coated cylinders and other next-gen renewable technologies simply can’t compete.

There was another factor driving down the cost of conventional photovoltaics. In recent years, China has worked aggressively to develop its domestic solar production capacity. National banks have given credit lines that dwarf the federal loans US firms enjoyed; local and provincial governments have provided tax incentives as well as land at below-market rates; and the national government recently established a so-called feed-in tariff, which compels utilities to buy electricity from solar developers at above-market rates to offset their production costs.

Understandably, American firms have struggled to remain competitive. In 1995, more than 40 percent of all silicon-based solar modules worldwide were made in the US; now it’s 6 percent. In less than two years, at least eight solar plants have closed or downsized, eliminating nearly 3,000 American manufacturing jobs, including the 1,100 employees who saw their jobs disappear with Solyndra’s spectacular September 2011 bankruptcy. China now accounts for more than half of global photovoltaic output, and Chinese-made modules are up to 20 percent cheaper than American ones.

Wind has also taken a hit. Not only can the turbines not match the current costs of gas-fired plants, the flood of cheap Chinese solar panels can make them less attractive as a green option, too. The pace of new wind-turbine installations in the US has declined by more than half since 2008. This past October, Cliff Stearns, the Republican chair of the House Energy and Commerce Oversight and Investigations Subcommittee, admitted to NPR what had by then become obvious: “We can’t compete with China to make solar panels and wind turbines.”

The boom has gone bust.

And yet, clean tech is far from dead. Certain companies and technologies will emerge from the ruins not only to survive but to thrive, just like they did after the bursting of the Internet bubble.

Electric cars seem like a relatively safe bet, spurred by both rising oil prices and federal rules requiring greater fuel efficiency. Additionally, as it has with solar, China has aggressively pushed into the competitive battery industry. As a result, prices for the lithium-ion battery modules in electric cars—which can cost more than some gas-powered cars—are coming down. Tesla started out making 600 sports cars a year, priced at $109,000 each; in 2012 it will begin selling the Model S, a full-size sedan that goes from zero to 60 in six seconds and costs just under $50,000 (once you kick in a $7,500 federal tax credit). Within five years, the company says, it will be producing 100,000 cars annually and charging just $30,000 apiece. The company’s stock took a hit in early December, after Morgan Stanley cut its price target—citing concerns about the broader EV market—but it was still up for the year, even after the drop.

Meanwhile, the low silicon prices and cheap Chinese photovoltaics that undermined next-gen clean tech have proven a boon to distributed-generation businesses—the firms that install solar systems to power individual homes and offices. These companies are thriving because they came up with a new financing model that makes installing standard flat-panel solar generators truly affordable.

A decade ago, a rooftop solar array for a 3,000-square-foot home would have cost the owner about $45,000. The price can now be less than $20,000. That’s not cheap, but instead of having to pay it up front, homeowners can now work with companies like San Mateo, California-based SolarCity and Oakland-based Sungevity and lease the systems for $119 a month—less than a lot of conventional electricity bills. John Stanton, head of federal affairs for SolarCity—which recently closed a $350 million deal with Bank of America to install panels that will provide power for up to 120,000 military families—likens it to leasing Xerox machines to offices. “It’s taking a 60-year-old business-equipment model and bringing it into the solar industry,” he explains.

That leasing model, combined with a number of software advancements, has transformed the rooftop solar business. It used to take months to close a residential sale; now these companies can use a combination of remote mapping and mathematical calculations to help determine exactly how many solar panels an individual home would need and how they should be positioned. The whole process can be completed in a matter of weeks.

In at least one respect, these companies rely on a very old-fashioned boost: federal and state subsidies and tax breaks. When they install a solar system on someone’s roof, they take all the government sweeteners that accompany the installation, which helps these firms offer their systems at lower prices. “Between 40 and 50 percent of the system is covered up front,” says Danny Kennedy, founder of Sungevity. “The customer is getting an incredible value proposition: ‘I’m going to save money from day one.’ That’s a hell of a thing. For no investment, I’m going to save money.”

But there is an investor: the taxpayer. Government coffers have been compensating for a number of market challenges solar faces, including the incumbency advantage of the fossil fuel industry and private investors’ distaste for capital-intensive enterprises that will take years to deliver a return. And in 2012, the solar industry may face a sudden reduction in these subsidies, as the post-Solyndra political climate grows less and less receptive to investments in clean energy. Despite the fact that renewable energy received only a quarter of the subsidies that fossil-fuel-based electricity received between 2002 and 2007, it’s wind and solar that are on the chopping block.

Even solar’s biggest allies on Capitol Hill—people like Edward J. Markey, a top Democrat on the House Energy and Commerce Committee—fear the industry’s oil and gas foes may have gotten the upper hand now that the clean-tech bubble has burst. “We are not Panglossian about what lies ahead,” Markey says. “The fossil fuel industry and its allies in Congress clearly see the solar and wind industries as a threat and will try to kill these industries as they have for the preceding two generations. They want this to be a five-year aberrational period.”

In other words, John Doerr may once again have a good reason to shed a tear.

Juliet Eilperin (@eilperin) is the national environmental reporter for The Washington Post and the author of Demon Fish: Travels Through the Hidden World of Sharks.