The 5-minute business brief. Join 50k + readers
Home – News – Solar Panels’ 25% Upgrade Only Pays Where Space Is Expensive
Tandem panels are being priced as a multiple of an ordinary panel. Which ordinary panel decides whether the technology works at all.
Tandem solar panels pull roughly 25% more electricity out of the same patch of sunlight, and Qcells, First Solar and Tandem PV all say they are close to selling them. The catch sits in a single line of The New York Times report on the race: by some estimates, tandem panels cost three to three and a half times as much as a conventional panel. Run that through the cost sheet of an American solar farm and the break-even lands at about 1.8 times. The gap between those two numbers is the whole story.
The coverage frames tandem as a technology race, with the United States finally holding a lead over China. That frame skips the question every buyer asks first. Extra efficiency is only worth what it saves you, and what it saves is space: land, racking, cables, crews. So the value of a tandem panel depends on how expensive a square meter is where it gets installed, and on whose cost structure the panel was built.
Here is what the arithmetic says about who wins that race, and why the answer is not the one in the headline.
US manufacturers told the Times they are on the edge of commercializing tandem panels, which stack a second light-absorbing layer on top of the usual silicon so the device can capture more of the spectrum. The added energy, roughly a quarter more than most panels on the market today, would let developers produce the same power on less land or less roof.
Qcells, the US arm of South Korea’s Hanwha, became the first company to obtain both UL and IEC 61215 certification for a perovskite-silicon tandem in July, with TÜV Rheinland verifying modules built on its pilot line in Bitterfeld-Wolfen, Germany. Tandem PV of Fremont, California, says a year of field testing shows degradation under 1% a year. First Solar says it has put more than $2 billion into thin-film research, the material other manufacturers would likely need to master for tandem layers.
The pitch leans on national stakes. Mike Carr of the Solar Energy Manufacturers for America Coalition told the Times Chinese tandem technology is no better than the American version and possibly behind it. China supplies about 80% of the solar panel market directly or through components, and AES chief executive Andrés Gluski put the scale bluntly: China “could probably supply the world twice over.”
The obstacles are just as clear. Tandem is expensive, power producers will not buy until they believe the panels age as well as silicon, and Wayne Li of the Electric Power Research Institute told the Times that manufacturers are reluctant to share their test data.
Tandem is not new physics. Scientists built early versions decades ago. The hard part has been survival: perovskite top layers have historically broken down under heat, humidity and ultraviolet light, which is why Qcells has spent the past two years firing ice balls at panels and baking them in test chambers rather than shipping them.
The “first to market” claim also needs a footnote. Oxford PV, a University of Oxford spin-off, shipped what it called the world’s first commercial perovskite tandem panels to an unnamed US utility-scale customer in September 2024. Those 72-cell modules ran at 24.5% efficiency and came off a pilot line in Brandenburg an der Havel, Germany.
The scoreboard on efficiency does not support the claim that China trails, either. LONGi announced a 35.5% crystalline silicon-perovskite cell in July 2026, certified by the European Solar Test Installation, and reports certified tandem modules at 31.4% and 29.4%. Trinasolar built an 808-watt tandem module in March 2025 and says it held 481 tandem patent applications at the time, which it ranked first in the world.
Then there is the American cost base. The Solar Energy Industries Association and Wood Mackenzie put utility-scale system prices at $0.95 per watt for fixed-tilt and $1.06 for single-axis tracking in Q2 2026, with utility modules averaging $0.33 per watt. China’s benchmark export price for TOPCon modules was about $0.108 per watt in late August, per OPIS. An American developer pays roughly three times the world price for the same commodity panel.
That threefold gap is the hidden variable in every tandem headline.
Each US player is taking a different route to the same product.
Qcells is the silicon route. Its tandem puts a perovskite layer on top of its own Q.ANTUM silicon cell. That makes the US story depend on Cartersville, Georgia, where Hanwha has invested $2.5 billion in an ingot-to-module campus. Silicon cell production there began in June, with the 3.5 GW cell line expected at full output in the third quarter. The road was not smooth: in November 2025, Qcells furloughed 1,000 workers after US Customs detained imported cells under the Uyghur Forced Labor Prevention Act, and normal production resumed only in March.
First Solar is the non-silicon route. In February it signed a non-exclusive license to Oxford PV’s perovskite patents for US manufacturing, a deal that explicitly excludes crystalline silicon. That exclusion tells you the plan: perovskite paired with First Solar’s own thin film, with no silicon wafer anywhere in the stack.
Tandem PV is the venture route. It raised $50 million in March 2025 in a round led by Eclipse, with Constellation Energy participating, bringing its total to $83 million, and its chief executive Scott Wharton told the Times tandem will dominate the industry by the mid-2030s.
None of those plans answers the question the cost sheet asks. So here it is in one picture.
The math is short. A tandem panel delivers 1.25 times the power of a conventional panel of the same size, so a panel that costs three times as much costs 2.4 times as much per watt. On a $0.33 American panel, that is $0.79 per watt.
What does the extra efficiency buy back? Everything on a solar farm that scales with area. On a $1.06 tracking project, $0.73 per watt, or 69% of the cost, is not the module. Assume, generously, that all of it shrinks with the land a project occupies. A panel with 25% more output needs 20% less of everything per watt, which saves about $0.15. Break-even is $0.33 plus $0.15: roughly $0.48 per watt, or 1.8 times the price of a conventional panel.
Use a narrower assumption, where only racking hardware and installation labor shrink with area (42% of installed cost in the National Renewable Energy Laboratory’s global breakdown), and break-even falls to about 1.6 times. Either way, a tandem priced at three times a US panel overshoots break-even by roughly two thirds.
Now redo it against China’s panel. Three times $0.108, divided by 1.25, is about $0.26 per watt. That is cheaper per watt than the conventional panel American developers buy today.
So a threefold price is either a deal-killer or a bargain, depending on whose panel it multiplies. The Times did not say.
Efficiency is a substitute for square meters, so price it in square meters. A more efficient panel does not make cheaper electrons on its own. It makes fewer racks, shorter cable runs and smaller land leases. That means its value rises with the cost of everything around the panel. Rooftops, where the area is fixed and installation labor is expensive, and land-constrained utility sites are the first buyers, not cheap desert acreage. Any founder selling a “more performance per unit” product should ask the same question: performance per unit is worth exactly what the unit’s surroundings cost.
The prize and the cost advantage live in different countries. IRENA put the average installed cost of a US solar project at $1,058 per kW in 2024, against $591 in China. Redo the break-even with Chinese numbers (2024 project costs and today’s module price, so read direction rather than decimals) and a tandem panel can cost about $0.20 per watt, or roughly 2.4 times a Chinese panel. A Chinese-built tandem at three times the local price needs a cost cut of about 21% to break even at home. An American-built tandem at three times the US price needs about 40%. The US market pays the most for efficiency. China is closer to delivering it profitably. What decides who serves American demand is trade policy: an August 6 proclamation added a 15% tariff and indicative minimum import prices on the solar chain, including $0.38 per watt for modules, according to SEIA. Where a product is made has already become the product itself for American battery plants selling into storage, and solar is heading the same way.
The patent holder may out-earn the factories. Oxford PV shipped first, then turned its head start into licenses. Trinasolar holds an exclusive license to its perovskite patents for China, with the right to sublicense. First Solar holds a non-exclusive one for the US. In a technology that every major manufacturer expects to adopt, a portfolio that sits on both sides of the trade wall collects a toll from whoever wins. Being first to ship was the credential. Licensing is the business.
The biggest risk to tandem economics is not price. It is time.
Conventional panels come with warranties that assume very slow aging: First Solar warrants 0.3% a year for its Series 7, and n-type silicon panels typically promise around 0.4%. Tandem PV’s sub-1% figure is reassuring for a new chemistry and expensive for a buyer. Take the ceiling of that claim at face value over 30 years and a tandem panel averages about 85.5% of its starting output, against about 94.2% for a 0.4% panel. The 25% energy advantage shrinks to roughly 13% over the life of the project, and the break-even multiple falls from about 1.8 to roughly 1.4. That is the gap Qcells’ Fabian Fertig means when he says the company has to convince banks.
The bottom cell is a second risk. A perovskite-on-silicon tandem still needs a silicon cell underneath, and the United States had only about 3.2 GW of cell capacity against 70 GW of module assembly at the end of March 2026, per SEIA figures. Cartersville helps, but it is one site, and last winter showed how a customs hold on imported cells can idle a factory for four months. Panels assembled from someone else’s cells are the same gap that Tesla’s Buffalo panel line still has to close.
Third, the learning curve. Solar has gotten cheap by running the same production steps billions of times. LONGi and Trinasolar already operate the largest silicon lines on earth and are adding tandem layers to them. If cost follows volume, the efficiency lead matters less than the volume lead, and the volume sits in China. It is the same pattern that left the US routing around China’s battery processing instead of out-building it.
Fourth, demand. SEIA reports that the Q2 surge in utility-scale installations came from developers racing safe-harbored projects into service before the Section 48E and 45Y tax credit window closes, and the Times notes federal permitting for new solar farms has become harder. A premium product launching into a market where the subsidy runway is shortening has less room to charge for novelty. The demand side is not gone, though: solar and batteries made up 70% of capacity added to US grids in the first half of 2026, per Wood Mackenzie.
Here is the strongest case against this article’s own arithmetic. The three-times estimate is an early-production number, and early numbers fall fast. If tandem layers add a modest step to an existing silicon line, as LONGi and Qcells are designing them to, the premium could compress well below 1.8 times within a few product cycles. Meanwhile land near transmission and data centers is getting pricier, which raises the value of every watt per square meter. Both trends move break-even in tandem’s favor.
What is a tandem solar panel? A panel with two light-absorbing layers instead of one. The top layer, usually perovskite, captures high-energy light, and the bottom layer, usually silicon, captures what passes through. Together they convert more of the spectrum into electricity than either can alone.
Is the US actually ahead of China on tandem? On certification, Qcells has a real first with its UL and IEC approvals. On cell efficiency, LONGi holds the record at 35.5%. On commercial shipments, a UK company with a German factory got there first. “Ahead” depends on which race you pick.
Will tandem raise solar’s capacity factor from 24% to 31%? Executives told the Times tandem could exceed 31%, against a 24% average for US solar farms per the EIA. Be careful with that comparison. Capacity factor divides output by a plant’s rated size, and a more efficient panel raises its rating along with its output. Scaling 24% by the extra 25% to 29% of output gets you right to 30% to 31%, which suggests the number describes more energy per acre, not a plant that runs more of the time. Energy quoted in one unit and bought in another is a familiar trap, and it is the same one that decides which long-duration batteries clear. Tandem will not make solar work at night.
Who buys tandem panels first? Buyers paying the most for space. Rooftop owners, land-constrained developers, and any project where racking, labor and land dominate the budget. EPRI says some energy companies have already bought early units so they are not left behind.
Does First Solar win either way? It has the only large-scale US thin-film business, a $2 billion research bet and an Oxford PV license that leaves the silicon route to others. If tandem ends up perovskite on thin film, First Solar starts with a head start. If it ends up perovskite on silicon, it does not.
The tandem story is being sold as a physics race, and physics races make good headlines. But the physics is mostly settled. Every serious manufacturer, in every country, is building some version of the same two-layer panel. The open question is economic, and it fits in one ratio: can a tandem panel get to about 1.8 times the price of an ordinary panel before the market stops paying a premium for the privilege of using less land?
That reframes who is actually competing. The United States has the most valuable market for efficiency in the world, because its land, labor and steel make every square meter expensive and its trade walls keep cheap panels out. China has the cheapest path to the product, because it already builds the bottom cell at a scale nobody else touches. The American opportunity is real, but it is a demand-side advantage protected by policy, not a technology lead.
The company that turned an early lead into a durable business so far is the one that shipped a small batch first and then licensed its patents to both sides of the wall.
Watch one number. It is not efficiency records, which China keeps setting. It is the first published price of a bankable tandem module with a 30-year warranty, sold in the United States. When that price lands under twice the cost of the panel next to it, the race is over, and the winner will be whoever made the most of them.
More About Me
The 2026 midterms may be the first US election meaningfully shaped by an answer engine. […]
Every marketplace you use runs on one number. Not users, not gross merchandise value, not […]
Apple filed what it calls “shocking evidence” against a former engineer. Read it against the […]
A $1.4 billion valuation, thousands of agents, and a published rate card that quietly says […]
The Canadian factory everyone films is not where the jackets come from. It is where […]
The Indian startup just raised $37 million to reach “cost parity with trucking.” Its own […]
Bezos spent 26 years as Blue Origin’s only customer for equity. The first outside round […]
Washington just waved through a media giant that would put CNN and CBS News under […]
Learn
Company
Business Model Examples and Types Using the Business Model Canvas for Detailed Analysis
Learn
Company
RECEIVE OUR UPDATES
The Biz Model Club
Get daily, no-fluff insights on the latest business models, startup strategies, and trends delivered straight to your inbox.
Get daily, no-fluff insights on the latest business models, startup strategies, and trends delivered straight to your inbox.
