A cell that lasted minutes

In 2009, a chemist named Tsutomu Miyasaka at Toin University of Yokohama dipped a titanium dioxide electrode into a solution containing methylammonium lead iodide, assembled a small photovoltaic cell, and measured 3.8 percent efficiency. The cell degraded within minutes. It was, by every practical measure, useless. But the material sitting on that electrode was a perovskite crystal, and Miyasaka had just demonstrated that perovskites could convert sunlight to electricity.

The paper barely registered. Twenty-six scholarly works mentioned perovskite solar cells that year, according to OpenAlex. Most of the photovoltaics community was busy optimizing silicon.

Then a British physicist named Henry Snaith visited Japan. Snaith, then at the University of Oxford, had been working on dye-sensitized solar cells. Through a collaborator named Takurou Murakami, he met Miyasaka and saw the perovskite work. He brought the idea back to his lab in Oxford, swapped the liquid electrolyte for a solid-state hole conductor, and got above 6 percent efficiency on his first attempt. By the time his team published in Science in late 2012, the number was 10.9 percent.

That paper cracked the field open. In 2013, papers mentioning perovskite solar cells jumped from 30 to 104. By 2015 it was 1,406. By 2024: 6,016. The total body of perovskite solar literature since 2009 now exceeds 48,000 works. A single unstable cell on a Japanese lab bench had become one of the most intensely studied materials in energy science.

The patent landscape tells a different story

There are 347 granted US patents for perovskite solar cells. For a field with 48,000 papers, that is a strikingly thin patent layer.

The trajectory is still accelerating:

But 59 patents in a year against 6,000 papers tells you something about where perovskite solar actually is: still crossing the gap between laboratory demonstration and manufacturing reality. The science is moving fast. The engineering is catching up.

Who’s filing

This is where the story gets interesting. The top patent holders are not who you’d expect from the academic literature.

CATL — Contemporary Amperex Technology Limited, headquartered in Ningde, China — is the world’s largest manufacturer of lithium-ion batteries. It holds roughly 37 percent of the global EV battery market. In fiscal year 2025 the company reported 423.7 billion yuan in revenue and spent 22.1 billion yuan on R&D. It supplies Tesla, BMW, Mercedes-Benz, Volkswagen, and Ford.

CATL has 12 US patents (combining its mainland and Hong Kong entities) on perovskite solar cell fabrication. Read the abstracts and they are unmistakably specific: electrode layer stacking sequences, electron transport layers, passivation chemistry, perovskite film deposition methods. These are not exploratory filings. They describe manufacturing processes.

Why would a battery company patent solar cells? Because the fabrication is nearly identical. Lithium-ion battery electrodes are made by coating chemical slurries onto metal foils in controlled atmospheres, then curing them. Perovskite solar cells are made by coating precursor solutions onto substrates in controlled atmospheres, then annealing them. The equipment, the clean-room protocols, the supply-chain logistics for handling air-sensitive chemical films at scale — CATL already owns all of this. Nikkei Asia reported that over 100 Chinese companies, including BYD, are now racing into perovskite cells. The battery supply chain is one retooling away from becoming the solar supply chain.

Oxford’s long road from lab to factory

Henry Snaith’s 2012 Science paper didn’t just launch an academic field. It launched a company. Oxford Photovoltaics, spun out of the University of Oxford, has spent over a decade turning perovskite from a laboratory curiosity into a shippable product.

Their approach: don’t replace silicon. Put perovskite on top of it. A tandem cell where the perovskite layer captures blue and green light while the silicon layer captures red and infrared, yielding more electricity from the same sunlight. In September 2024, Oxford PV shipped its first commercial perovskite-on-silicon tandem modules from its factory in Brandenburg an der Havel, Germany, to a US utility-scale customer. The modules achieved 24.5 percent efficiency. PV Magazine reported the company targets 26 percent modules by 2026 and a gigawatt-scale manufacturing plant by 2027.

Nine US patents back that work. They describe the specific interfaces between the perovskite and silicon layers, the deposition sequences, and the encapsulation methods that keep the notoriously moisture-sensitive perovskite film stable for years outdoors. The engineering problems in those patents are exactly the problems Miyasaka’s 2009 cell couldn’t solve: degradation, scaling, and longevity.

The American bet

On the US side, two startups are building their own manufacturing paths. CubicPV, formed from the merger of 1366 Technologies and Hunt Perovskite Technologies, has raised over $100 million backed by Breakthrough Energy Ventures (Bill Gates’s climate fund). They achieved a 24 percent mini-module efficiency record verified by NREL and hold 10 US patents.

Swift Solar, founded by researchers from Stanford and MIT, closed a $27 million Series A in June 2024 and has raised $44 million total. In August 2025, Swift completed a demonstration of its US-manufactured perovskite tandem panels in partnership with the Department of Defense during the Cyber Fortress exercise in Virginia Beach. They hold four US patents and recently announced a partnership with American Tower Corporation for deployment on wireless infrastructure.

The National Renewable Energy Laboratory (NREL) in Golden, Colorado, sits at the center of much of this work. Alliance for Sustainable Energy, NREL’s management contractor, holds 13 US perovskite solar patents, the second-largest portfolio after Panasonic. NREL’s efficiency certification is the benchmark every perovskite team races against.

What Miyasaka started

There is a direct line from a leaky cell in Yokohama to a German factory shipping tandem panels to Texas. Miyasaka made the crystal. Snaith proved it could work in solid state. Oxford PV spent twelve years turning that proof into a product. And now the battery factories of Ningde and Shenzhen are filing patents to manufacture it at a scale no solar startup can match.

The 48,000 papers were the rehearsal. The 347 patents are the set list. The show is about to start.

Method note

Patent counts are from 9.3M US utility and design grants sourced from USPTO bulk grant XML, filtered by full-text keyword search for “perovskite” co-occurring with “solar” or “photovoltaic.” Assignee counts combine variant spellings and subsidiary entities for each organization. Literature counts are from OpenAlex (357M scholarly works), using the same keyword filter. Oxford PV commercial and production details sourced from PV Magazine and PV Tech (September 2024 and January 2026 reporting). CATL financial data from SolarQuarter (March 2026). Swift Solar details from pv magazine USA and PR Newswire (2024). CubicPV funding from PitchBook. The 2025 patent count is partial-year through late September. All patent-year counts reflect grant date, not filing date.