A 70-Year History of the Solar Industry

The efficiency of the first practical solar cell was about 6%. That was a figure measured in a laboratory. Commercial modules mounted on rooftops today sit around 22~24%. That is roughly a fourfold increase. Yet over that same period, the price plummeted from hundreds of dollars per watt to below $0.1. Even in nominal dollars without adjusting for inflation, that is a drop of more than 1,000 times. Efficiency alone cannot explain this gap.

The rest came from elsewhere. As wafers became thinner, they used far less silicon, and a dedicated grade of polysilicon for solar emerged—one that did not need to be as pure as the semiconductor grade. Before 2000, 90% of polysilicon was produced for semiconductors, but by 2012, 90% went toward solar. On top of that, factories grew larger, specialized automation equipment was introduced, and manufacturing yields climbed.

There is a regularity to this decline. Every time cumulative production doubled, module prices dropped by about 20%. Richard Swanson, the founder of SunPower, outlined this pattern in a 2006 paper, and a 2012 article in The Economist dubbed it 'Swanson's Law.' He did not name it after himself.

The crucial detail is the condition behind this law. Prices do not fall simply 'with the passage of time'—they fall 'when you make a lot.' That means for prices to drop by half, someone has to step up beforehand, buy them at expensive rates, and double that cumulative production. That 'someone' is the real protagonist of this story.

Of course, this is not a law of physics, but an observation of what has happened so far. If raw material prices spike suddenly, the curve can temporarily reverse. In fact, that happens once on a massive scale, and that episode will appear later in our story.

Before we begin, let's clear up one distinction. Solar photovoltaics (PV) and solar thermal are different technologies. Solar PV converts light directly into electricity by leveraging the property of semiconductors where electrons are released when light hits them. Solar thermal simply uses sunlight to heat water or oil. Because photos of them on rooftops look alike, the two are still often mixed up today, and our story features one of the most famous examples of that confusion.

The fact that light can generate electricity has actually been known for a very long time. In 1839, Edmond Becquerel of France—who was reportedly just nineteen years old—discovered that shining light on an electrode submerged in an electrolyte caused an electric current to flow. It was the very first observation of the photovoltaic effect.

The first solid-state solar cell appeared in 1883. American inventor Charles Fritts made it by coating a selenium plate with an extremely thin layer of gold, and in the mid-1880s, he installed it on a rooftop in New York. Its efficiency was less than 1%. In 1905, Albert Einstein published a paper explaining the photoelectric effect using the light quantum hypothesis. It was this paper, not his theory of relativity, that earned him the 1921 Nobel Prize in Physics. Still, Einstein did not invent the solar cell. The photoelectric effect and the photovoltaic effect are neighboring phenomena, not the exact same thing.

The architecture of every silicon solar cell we use today originated at Bell Labs in the early 1940s. While examining a silicon ingot, Russell Ohl discovered a naturally formed p-n junction along a boundary created by impurities. When light hit it, a voltage appeared. The efficiency was only around 1%, but the underlying structure was the same as today's.

Then in April 1954, Daryl Chapin, Calvin Fuller, and Gerald Pearson of the same laboratory unveiled a silicon solar cell with an efficiency of about 6%. The New York Times ran the story on its front page the next day. But the task originally assigned to Chapin was not "solar power generation." It was to find a way to keep people from having to hike up to remote telephone repeaters just to replace their batteries.

Two years later, Chapin ran the numbers. Powering an average American household with these cells, he calculated, would cost about $1.5 million. In mid-1950s dollars. The invention was a success, but there was no one to buy it. There was hardly anyone on Earth willing to pay hundreds of dollars per watt for electricity.

On March 17, 1958, the U.S. Navy's Vanguard 1 satellite was launched carrying six silicon cells. Together, they generated about 1W. While the chemical battery on board died after just 19 days, the transmitter powered by the solar cells kept running until 1964. Vanguard 1 remains the oldest human-made object still orbiting the Earth today.

Satellites were customers who did not care about price. After all, you cannot fly up into orbit to swap out batteries. Between 1958 and 1969, the U.S. government spent about $50 million on solar energy. In 2023 dollars, that is about $500 million. Yet virtually the only customer was the space program. Even up to 1974, 94% of the solar market was satellites.

The breakthrough that brought solar down to Earth came from an unexpected place. Chemist Elliot Berman believed that if he could bring the price down from $100 per watt to $20, solar would find uses on the ground. So he took off-grade wafers discarded by semiconductor factories and used cheap packaging materials instead of space-grade enclosures. In 1973, his company rolled out modules with a manufacturing cost of $10 per watt and a retail price of $20. All figures are based on nominal dollars at the time.

The funding came from Exxon. Its first customers were also offshore drilling platforms, navigation buoys, mountaintop telecom repeaters, and remote water pumps. It was oil and maritime infrastructure that opened solar's first terrestrial market. These were places where the cost of running power lines was higher than the cost of the panels themselves.

The government changed its approach too. From 1975 to 1981, the U.S. JPL purchased modules directly under an initiative called "Block Buy." Across five rounds, it bought about 400kW, which represented about 17% of the global market at the time. That is less than what the world installs in an hour or two today. That is how small the market was. What mattered was not the volume, but the requirements. With each round, they raised the required performance standards, establishing the durability standards that allow today's modules to hold up for decades. It was a model where the government became the first customer to make products better.

In 1977, the United States established the Solar Energy Research Institute in Golden, Colorado. The following year, the Energy Tax Act introduced deductions covering 30% of the first $2,000 and 20% of the next $8,000 for residential installation costs, along with 10% for commercial systems. This became the prototype for solar tax credits in the United States today.

In 1979, President Carter installed 32 panels on the roof of the White House West Wing. It is a scene that has been cited for over 40 years as a symbol of the solar era. However, those 32 panels were not solar photovoltaic panels, but solar thermal collectors that heated water for the cafeteria and laundry. They were not built to generate electricity. Simply because the rooftop photo looked similar, the industry's most famous photograph has been remembered as a picture of an entirely different technology.

The direction soon reversed. In 1981, the Reagan administration proposed cutting the Department of Energy's solar budget by 60–70%, and total renewable energy budget cuts for fiscal years 1982–1986 were tallied at about $1.73 billion. In 1985, they allowed the Carter-era residential tax credits to simply expire. When the roof was resurfaced in 1986, those 32 panels were removed and never put back. After sitting in a warehouse, the panels were acquired in 1991 by Unity College in Maine, and one of them is now in the Smithsonian.

Yet the statement that 'Reagan killed American solar' is only half true. Around the same time, oil prices plummeted, completely erasing the economic viability of alternative energy. Capital drained away with it. In 1984, Exxon shut down its solar subsidiary, and in 1990, ARCO sold ARCO Solar—once the world's largest solar manufacturer—to Siemens for $36 million. Even after peaking with $40 million in revenue in 1988, it had failed to turn a profit. Then in the 2000s, cheap shale gas emerged as an alternative power source, stripping away the justification once again.

In short, what the United States missed out on was not the result of a single political decision. It was a 30-year process where policy disruptions, the withdrawal of oil capital, and the arrival of cheap alternative energy all coincided. The United States invented the cell and established durability standards, but failed to create a market that would actually keep buying them.

In 1974, right after the first oil shock, Japan's Ministry of International Trade and Industry launched the 'Sunshine Project.' For a country without natural resources, an oil crisis was an immediate national emergency. So they designed the initiative with the premise that it was fine if it took 25 years. From 1974 to 1981, more than 25% of this project's budget went to solar energy. The justification of 'energy security' forged during this period became the foundational logic of global solar policy for the next half century.

The problem was that solar electricity back then was nowhere near cheap enough to feed into the power grid. So Japan pivoted. In 1980, it established NEDO and began putting amorphous silicon cells into pocket calculators and wristwatches. While these cells had low efficiency, they used as little as one-hundredth the silicon, and indoor ambient lighting alone was more than enough to run a single calculator. By 1982, 80% of Japan's solar production was going toward consumer goods like these.

It was calculators, not power plants, that kept the factories running. Back when solar panels could not sell as devices for generating electricity, Japan sold products where you never had to replace the batteries. If you recall the rule from the previous chapter, this, too, was a way to build up cumulative production.

Climbing onto rooftops came after that. In 1994, Japan introduced subsidies for residential solar power. At first, the government covered about 50% of installation costs, lowering that share with each passing year. From 539 installations in the first year, annual installations climbed to 54,475 in 2004. By the time the program wrapped up in 2005, about $1.1 billion had gone into about 300,000 households. Over that same span, Sharp's production multiplied 200-fold, and in 2003, it became the world's largest solar manufacturer, holding about 30% of the global market. Japan was once number one in the world.

Then, in 2005, Japan ended the subsidy. The assessment was that prices had dropped as much as targeted. Once demand vanished, the top spot disappeared with it. Annual installations fell from 260MW in 2005 to 180MW in 2007, and Sharp's market share sank to around the 10% level in 2007. Japan would later revive generous feed-in tariffs at 42 yen per kWh in July 2012 following the Fukushima disaster, but by that point, the global market already belonged to factories in other countries.

Germany is by no means an ideal country for solar power. Its solar irradiance is far lower than that of Spain or California. Yet in 2004, Germany alone accounted for about 60% of the global solar installation market. The reason was not the weather, but the policy framework.

It did not take off right from the start. In September 1990, the Federal Ministry of Research and Technology launched the '1,000 Roofs Program' to place panels on 1,000 residential rooftops and collect real-world operating data. In 1991, the Electricity Feed-in Act mandated utilities to purchase renewable electricity at 90% of the average retail rate. But that price fell far short of the cost of solar power, so deployment barely happened. That led the city of Aachen to try a different experiment: a tariff designed to cover costs, purchasing electricity at about 20 times the retail rate at the time. This 'Aachen Model' became the design blueprint for the legislation that followed.

In 1999, the '100,000 Roofs Program' offered loans at interest rates 4.5 percentage points below market rates, featuring a 10-year repayment term and a 2-year grace period. By its conclusion in 2003, financing had gone out to over 300 MWp of installations.

Then on April 1, 2000, the Renewable Energy Sources Act (EEG) went into effect. It guaranteed a fixed purchase price for solar electricity for 20 years: 50.6 euro cents per kWh for rooftop systems and 45.7 euro cents per kWh for ground-mounted systems. That was more than double the retail electricity price at the time. It was a law paired with the nuclear phaseout agreed upon in the same year. Grounded in the logic of preventing fossil fuels from expanding as nuclear plants were turned off, it intertwined energy security with the goal of fostering a new industry. Lawmakers including Hermann Scheer of the Social Democratic Party and Hans-Josef Fell of the Green Party championed the bill.

The key was not the size of the tariff, but the length of the commitment. With a 20-year fixed purchase guaranteed by law, banks began lending money against rooftops. The funding mechanism was also unusual: it came not from the national budget, but from a surcharge added to electricity bills. Thanks to this, it was not swayed by annual budget reviews, and instead, the cost was paid by the people who used electricity. In 2013, this surcharge accounted for about 18% of German household electricity bills. Estimates of the total amount committed under these 20-year contracts vary widely. While some estimates reach hundreds of billions of euros, the figures differ across institutions because they represent aggregate future commitments rather than cash already paid out.

When the 2004 amendment raised tariffs and removed the total capacity cap, German installations that year surged to four times the previous year. From this point on, solar factories around the world expanded their capacity with their eyes on the German market alone. Some countries copied this approach to the letter. In 2007, Spain introduced a very generous feed-in tariff by royal decree, and the following year, installations surged by about 400%, with subsidy payouts jumping from 194 million euros to 990 million euros. Startled, Spain slashed its support, but the problem was how it cut it: the reductions were applied retroactively to installations that were already up and running. The market froze immediately, followed by a flurry of international arbitrations filed by foreign investors. During the same period, Germany lowered tariffs only for new installations while honoring existing contracts. What set the two countries apart was not the size of the subsidy, but the credibility of the promise.

In 2001, Shi Zhengrong returned home after completing his doctorate in solar cells in Australia and founded a company backed by capital from the Wuxi municipal government and state-owned enterprises. That was Wuxi Suntech. The company set up a 10MW production line in 2002 and a 30MW line in 2003. Even then, China accounted for less than 2% of global solar cell production.

The German market exploded the following year. Riding that surge in demand, Suntech went public on the New York Stock Exchange in December 2005, raising about $400 million. It was the first Chinese private enterprise to list on the New York exchange without prior operating revenue in the United States. That same year, Japan ended its residential solar subsidies. The year one side closed its doors, the other opened them.

In 2008, several major shifts collided at once. As solar demand ate into silicon supplies reserved for semiconductors, the spot price of polysilicon soared to $475 per kg in April. It was a brief moment when the learning curve ran in reverse. That year, Germany's Q-Cells became the world's No. 1 cell manufacturer, and German production overtook Japan's. At the same time, China became the world's largest producer for the first time. Fully 98% of Chinese-made cells were exported. China had become the world's leading producer without having a domestic market of its own.

What happened next was decisive. Spurred by sky-high prices, companies rushed to build polysilicon plants, and the number of manufacturers jumped from 7 in 2008 to 70 in 2010. By July 2009, the spot price plunged to $55 per kg. That was an 88% crash in just 15 months. In this brutal downturn, Chinese state support demonstrated its real power. It took a fundamentally different form. China's support was not 'so much per panel,' but 'credit to build factories.' In September 2010, JA Solar signed a financing agreement with China Development Bank for up to 30 billion yuan. Across the 2010s, the financing provided by this single bank to Chinese renewable energy firms is estimated at over 200 billion yuan, or about $28 billion. On top of that came local government support in land, electricity tariffs, and tax relief, and in 2009, the 'Golden Sun' program was launched, subsidizing installation costs upfront.

These are the documented facts. On the other hand, the exact total of these subsidies and whether they breached international trade rules remain contested to this day. When the United States brought a challenge before the World Trade Organization, the subsidy China actually terminated was a special fund for manufacturing wind power equipment; most solar issues were dealt with through unilateral tariff investigations by individual nations.

Looking purely at the economic structure, this is how it played out: electricity bills paid by German consumers created global demand, and the factories that expanded to capture that demand were overwhelmingly in China. The United States, Japan, and Germany paid the cost of the learning curve, and China was the one that slid down it the fastest. Yet Germany did not simply lose out. What those utility bills ultimately paid for was not Chinese factories, but the collapse in the cost of solar energy itself—a balance sheet we will examine later on.

European manufacturing collapsed first. In December 2011, Germany's Solon went bankrupt, followed in April 2012 by Q-Cells, which had been the world's number one just four years earlier (it was later acquired by Hanwha to become Hanwha Q CELLS). In 2013, Conergy went under, and in 2017, SolarWorld, the last major cell manufacturer, collapsed. SolarWorld was the company that had bought the old ARCO Solar plant in 2006. German solar employment fell by more than 70% from its peak. Germany created the world's largest market and nurtured the industry, but couldn't keep that industry at home.

The United States was shaken too. In August 2011, Solyndra went bankrupt. It came just two years after receiving a $535 million loan guarantee from the Department of Energy. As the price of Chinese crystalline silicon fell by half, the company's defining distinction—cylindrical thin film that used less silicon—lost all its purpose.

This was when tariffs entered the picture. In 2012, the United States imposed anti-dumping and countervailing duties on Chinese cells, and the EU launched an investigation in September of that year. In June 2013, the EU imposed provisional tariffs, patching things up that July with a minimum import price agreement. The condition was not selling below 0.56 euros per watt and not exceeding 7 GW per year. But in 2018, the EU brought all of these measures to an end. There was hardly any manufacturing left in Europe to protect, and installers and developers wanted cheap modules.

China's response was swift. As export routes narrowed, the State Council opened up the domestic market in 2013 by setting a target of a cumulative 35 GW by 2015. They pivoted toward their own rooftops and deserts. When the United States raised tariffs, production shifted to Malaysia, Vietnam, Thailand, and Cambodia, and in April 2025, the U.S. treated this as circumvention and finalized duties once again. An anti-dumping duty of 125.37% and a countervailing duty of 3,403.96% slapped on Cambodia made big headlines, but these were the maximum rates applied to companies that did not cooperate with the investigation. Not all imports were subject to these rates.

At the very front of the supply chain, a different kind of problem broke out. Solar manufacturing begins with extracting metallurgical-grade silicon from quartz, and this initial stage was concentrated in the Xinjiang Uyghur Autonomous Region. By around 2021, 40 to 45% of the world's solar-grade polysilicon—the next step in the process—also came from this region. In June 2021, U.S. Customs issued a Withhold Release Order against the region's largest producer, and the Uyghur Forced Labor Prevention Act, signed in December of that year, went into effect on June 21, 2022. Any goods produced wholly or in part in Xinjiang were presumed to be made with forced labor and barred from import. Module containers were detained en masse at U.S. ports, and Chinese industry responded by shifting production to places like Inner Mongolia and Sichuan. Because this is where an industry built for the climate collides head-on with human rights issues, neither side lends itself to an easy conclusion.

The United States did not rely on tariffs alone. In August 2022, the Inflation Reduction Act began offering tax credits of 7 cents per watt for domestically produced cells and 17 cents per watt for modules. At a time when module prices were in the 10-cent range per watt, this structure added 17 cents just for the single step of module assembly. It showed just how expensive bringing manufacturing back really was. Yet in July 2025, residential solar tax credits were set to wind down by the end of that year. They had originally been scheduled to remain at 30% through 2032. It was a replay, after more than 40 years, of the expiration of tax credits from the Carter era. In the meantime, as of the first quarter of 2024, the average installation cost for U.S. residential solar was $3.15 per watt, of which modules accounted for only $0.40. The rest went to soft costs like sales, permitting, and labor. Even if modules became free, rooftop solar on American homes wouldn't get significantly cheaper.

Tariffs also have the opposite effect. They keep module prices sold inside the United States higher than global market rates. That means the cost paid to protect manufacturing is shared by the people putting panels on their roofs. Bringing manufacturing back and making installation cheap may not point in the same direction—and that is where the United States stands right now.

In 2024, new global solar installations reached about 600 GW. Cumulative capacity surpassed 2 TW, with China alone installing 329 GW. Module prices dropped to around $0.09 per watt. The global average levelized cost of electricity (LCOE) for utility-scale solar fell to $0.043 per kWh, a 90% drop from $0.417 in 2010.

Yet in that same year, the top 10 module manufacturers shipped 502 GW and still lost a combined total of about $4 billion. From China's top five companies alone, 87,000 people left, and more than 40 companies exited the market. In the first half of 2025, China's top four companies lost about $1.54 billion as well. Even the players who took the entire market aren't making any money.

The reason lies in the rule we saw earlier. In an industry where prices drop predictably, today's cutting-edge facility becomes tomorrow's liability in just a few years. That is why the market leader kept changing. Sharp in 2003, Q-Cells in 2008, and Suntech in 2009 each collapsed or were pushed aside within a few years. Today's top companies stand on the very same rule.

When talking about market share, you have to specify which stage you mean. China accounts for over 80% at the module stage, but when it comes to wafers, China produced 97% of the world's 804 GW production in 2024. For polysilicon, it holds about 94% based on production capacity. That is why people say, 'Even if you build module factories in other countries, you can't escape China.'

So what did Germany gain? They lost their factories. In return, Germany added 16.2 GW in 2025 to reach a cumulative 116.8 GW, and solar accounted for 17.0% of German electricity generation, surpassing lignite for the first time. In 2000, when the EEG took effect, this share was less than 0.1%. What Germany bought was not factories, but a world where cheap solar power exists, and they actually got it. Of course, the reality remains that German electricity consumers footed the bill. In July 2022, Germany abolished the EEG surcharge and shifted the funding to the national budget. After 22 years, the principle of 'electricity consumers pay' came to an end.

In 2025, solar power broke records once again. With new installations reaching 647GW and annual generation growing 30% to 2,778TWh, it surpassed wind power for the first time in history. Renewables as a whole accounted for 33.8% of global electricity generation, overtaking coal (33.0%) for the first time, and solar alone met 75% of the growth in global electricity demand that year. Solar power's share of global electricity stood at about 8% on an annual basis, compared to 6.9% in 2024.

Yet in Germany that same year, wholesale electricity prices turned negative for 573 hours—an all-time high. These are hours when feeding electricity into the grid actually means paying money to do so. Depending on the accounting criteria, it is also reported as 453 hours. In California in 2024 alone, 3.4 million MWh was curtailed and discarded, with solar accounting for 93% of it. The statement that solar is cheap and the statement that solar is in surplus are both true at the same time.

So the current bottleneck isn't the cost of generation; it's storage, the grid, and the hours of the day. This area is already shifting as well. In 2025, battery costs fell 45% and deployments surged 46% to 250GWh, with about 14% of newly added solar electricity passing through storage.

When talking about efficiency, three distinctions must be made. In April 2025, LONGi set a world record of 34.85% with a perovskite-silicon tandem cell, but this was a lab cell the size of a fingernail. During the same period, commercial modules installed on rooftops were around 22–24%. The true mainstream on commercial production lines is TOPCon, which accounted for about 80% of new cell production in 2025. The remaining challenge for perovskite isn't efficiency—it's lifespan.

There is no need to try to predict the future. Instead, let's map out what to keep an eye on. First is storage: how much further the cost of shifting surplus daytime electricity to the evening falls will determine the limit of how much more solar can grow. Second is the upstream supply chain: while module factories are going up in many countries, wafers and polysilicon remain heavily concentrated in one place. Third is losses: how long each company can endure current deficits will dictate who survives. Fourth is policy consistency: throughout this story, the places where markets collapsed were not where support simply tapered off, but where promises were broken.

One last thought. There is no clear-cut winner in this story. The United States invented the cell and set the standards, then lost the market; Japan nurtured the industry, then lost its top spot after scaling back demand; Germany built the market, then lost its factories; and China captured the entire market, only to bleed red ink. Yet in the meantime, something that once cost hundreds of dollars per watt fell below $0.10. Those who paid the bill and those who reaped the spoils were different, but the ultimate result—a world where cheap solar power exists—ended up belonging to no single nation.