Perovskite.
Records don't sell panels.

The efficiency record doesn't matter much anymore.

In July, LONGi certified a 35.5% perovskite-on-silicon solar cell, the best of its kind ever made, against a bit over 20% for a standard silicon panel. The buyers who'll decide whether perovskite ever goes mainstream barely reacted. They want to know something no lab can tell them yet: will it still work in 2051?

That question has dogged perovskite from the start. Perovskites are a family of crystal materials that turn light into electricity cheaply and very well. Layer a thin film of one on top of an ordinary silicon cell and you get a "tandem" that captures more of the spectrum than silicon alone, which is how those 30%-plus numbers happen. But perovskites break down when heat, moisture, and light get to them, and solar panels are sold with 25-year warranties.

Now customers are writing durability into contracts. When Huaneng, one of China's big state-owned power generators, ran the country's first commercial tender for tandem modules in March, it demanded international safety certification and a 25-year warranty along with the efficiency.

The trouble is that the companies best placed to build this at scale are in no shape to bet big. LONGi and JinkoSolar both lost money in the first half of 2026, China's solar installations fell by about two-thirds, and their stocks have been hammered.

So perovskite stocks in 2027 are a race to sell a warranty someone will believe, run by companies that can barely afford the factory. Here's how the list breaks down: the Chinese record holders, the ones chasing bankability, the ones putting perovskite where silicon can't go, and the private companies trying to get to market first.

China's solar giants own the lab records. What they don't have is spare cash.

LONGi Green Energy SHA: 601012

LONGi Green Energy has set the perovskite-silicon tandem record again and again, most recently 35.5% in July. Nobody has a better claim to the technology.

It also lost CNY 3.7 billion in the first half, its revenue fell 18%, and the stock is down almost 40% this year. Its next step toward tandem production is a planned 100 MW pilot line costing about CNY 200 million, a rounding error for a company that makes tens of gigawatts of silicon panels. LONGi has the science. The question is whether the solar price war leaves it room to industrialize it before someone else does.

Trina Solar SHA: 688599

Trina Solar has done the thing that matters more than a record: it sold some. In June it supplied its first commercial tandem modules to a customer in New Zealand. It also holds the record for the most powerful tandem panel, 907 watts from a full-size module.

Trina licensed Oxford PV's perovskite patents for China in 2025 and is aiming for gigawatt-scale tandem capacity by 2029. Its losses narrowed sharply in the first half, which makes it one of the healthier Chinese makers to carry that plan.

JinkoSolar NYSE: JKS

JinkoSolar is the world's biggest panel shipper, and most of what it ships uses TOPCon, the most common type of silicon cell. Its tandem work puts perovskite on top of TOPCon, reaching a certified 34.82% this year.

That matters because it's the most direct upgrade path for the largest pile of existing factories in the industry. The stock is down about 65% this year after another losing half. JinkoSolar is a deep-value bet that the price war ends and the tandem upgrade comes after.

These companies are working on the part buyers care about: proving it lasts and backing it with a balance sheet. For perovskite stocks, that's what bankability means.

Hanwha Solutions KRX: 009830

Hanwha Solutions, the parent of Qcells, has a quieter claim than the efficiency records. In July its tandem technology became the first to be certified to both the international IEC 61215 standard and the U.S. UL 61215 standard, the durability tests that lenders and buyers look for.

The tandem cells come from a pilot line in Germany, and Qcells has targeted commercialization in 2027. Its stock is up about 35% this year, and its new factory in Georgia makes silicon cells in the U.S. If tandem becomes bankable, Qcells is the company with both the certificate and the U.S. footprint.

GCL Technology HKEX: 3800

GCL Technology is mostly a polysilicon maker, a business that has been losing money for two years. It's here for its roughly 40% stake in GCL Optoelectronics, the Kunshan company that's closest to making perovskite modules at scale.

GCL Optoelectronics opened a gigawatt-scale perovskite base in 2025 and is ramping its first 500 MW line. In March it won Huaneng's tandem tender, the one with the 25-year warranty requirement, and it's targeting 50 to 70 MW of shipments this year. Its founder says the modules cost about twice what silicon does at this scale and should approach silicon's cost at 3 GW.

GCL Tech has said it's preparing a Hong Kong IPO for the unit. As of early October, nothing had been filed with the exchange. Until it is, buying GCL Tech means buying the perovskite stake with a lot of polysilicon attached.

First Solar NASDAQ: FSLR

First Solar is America's biggest solar manufacturer, and it doesn't use silicon at all. It makes thin-film panels, which gives it a manufacturing base that's closer to how perovskite is made than any silicon factory.

It has a $13.6 billion order backlog to fund the research. In February it signed a U.S. license to Oxford PV's perovskite patents and is developing perovskite at its Ohio line. If perovskite gets bankable in America, First Solar is the company with the money, the factories and the legal right to build it.

Perovskite can be printed onto flexible film. That opens places a rigid glass panel can't reach, and Japan is making it a national project. These two companies are leading it, and they're the perovskite stocks furthest from the Chinese price war.

Sekisui Chemical TSE: 4204

Sekisui Chemical makes a flexible perovskite film that can stick to walls, curved roofs and structures too weak to carry a heavy panel. It started selling it in March, first to government pilot projects and Tokyo Metro sites.

It's building a 100 MW plant in Sakai, with mass production starting in its 2027 fiscal year and about ¥90 billion committed, backed by a Japanese government subsidy. The business is still losing money. But Sekisui is one of the first big companies to sell perovskite as a product in its own right, not as a layer on silicon.

Kaneka TSE: 4118

Kaneka, a Japanese chemicals maker with a long history in high-efficiency solar, is doing what more companies should: proving it in the field first. Its tandem panels are installed on Saitama City Hall in a yearlong demonstration that runs until March 2027.

Commercial sales are planned for its 2028 fiscal year. That's a slower timeline than the Chinese makers promise, and a more honest one.

Oxford PV (Private)

Oxford PV says it was the first company to ship commercial perovskite-silicon tandem panels, and its patent portfolio may be worth more than its factory. Trina licensed the patents for China in 2025, and First Solar licensed them for the U.S. this February.

That makes Oxford PV a toll collector. Whoever wins tandem in the two biggest solar markets may be paying Oxford PV to do it.

Caelux (Private)

Caelux doesn't make a whole panel. It makes a perovskite-coated glass layer that goes on top of an existing silicon module and adds power. In April it signed a five-year, 3 GW deal to supply Solx, which makes panels in Puerto Rico, with volumes starting next year.

That's the lowest-friction way to get perovskite onto roofs: no new silicon factory, just a better piece of glass.

Tandem PV (Private)

Tandem PV opened a 40 MW factory in Fremont, California, in April. It's tiny, and that's on purpose. The company is shipping trial panels to customers this year and building up durability data before it scales, with high-volume manufacturing planned for 2028. It has raised $87 million, including Energy Department money.

Swift Solar (Private)

Swift Solar started as a perovskite company, then in March bought the technology and patents of Meyer Burger, the bankrupt European solar maker. The plan is to build a U.S. factory for high-efficiency silicon cells first and add a perovskite layer later. It's the clearest example of the new playbook: earn money on silicon now, upgrade to tandem when buyers trust it.

The scarcest thing in perovskite solar is time in the sun.

Banks will finance tandem panels at scale when they've seen how they hold up outdoors over years, not just in accelerated lab tests. There's no shortcut for that. Every month a panel sits on a roof in real weather is a month of data nobody else can copy.

That's why the small projects matter more to perovskite stocks than they look. Trina's panels in New Zealand, Huaneng's demonstration plants, Sekisui's film on Tokyo Metro sites and Kaneka's panels on Saitama City Hall are all starting clocks right now.

By 2028, the most valuable thing a perovskite company can show a buyer won't be a certificate or a record. It'll be the oldest set of panels still performing in the field. The companies that got hardware outdoors earliest, even in tiny amounts, will have the head start nobody else can buy.