Walk into 981 Bing Street in San Carlos on a weekday and what hits you first is the quiet. Not silence - there is the low hush of a deposition chamber humming through a thin-film run, the small click of a probe station, somebody arguing politely about charge transport - but a quiet you wouldn't expect from a company that has set itself the chore of beating silicon. Swift Solar is fifty-six people. The plan is bigger than the building.
The plan is this: take perovskite, the lab-darling semiconductor that physicists have been promising since 2009, and stack it on top of conventional silicon. Two cells, one panel, more electricity per square meter than silicon alone has ever delivered. Then build a factory in the United States to make it. Then sell it.
Solar people have heard variations of this pitch for years. Many companies have raised money on it. Most have not survived contact with the basic, irritating question of whether perovskite films will still work after a Phoenix summer. Swift Solar has answered it, so far, by being more patient than its category usually permits.
02 / The ProblemThe Ceiling Nobody Built but Everybody Hits
For all of solar's spectacular cost collapse - and it has been spectacular; a watt of solar costs roughly a hundredth of what it did in 1976 - the technology has been hauling around a physical limit. Single-junction silicon cells max out, in theory, at about 29% efficiency. In the field, the good ones manage around 22 to 24%. The bad ones, much less. This is a ceiling not because engineers are bad but because photons of the wrong wavelength simply do not give up their energy to silicon. They warm it up instead.
Perovskites - a class of crystal structure named, mildly, after a Russian mineralogist - absorb light in a different and complementary slice of the spectrum. Put one on top of a silicon cell and you get two cells stacked: a tandem. The math says you can push past 33%, and engineers at NREL and elsewhere have already demonstrated 30-plus in lab cells. The math also says that if you can deliver that gain at scale, you change the unit economics of every solar project on earth.
The catch, of course, has always been "at scale." Perovskite films are sensitive. They degrade in moisture, in heat, in light - which is unhelpful, given the job description. The history of perovskite startups is the history of people trying to manufacture an unstable semiconductor in a stable way. Almost nobody has made it to a real factory.
03 / The BetSix PhDs Walk Into a Spin-Out
Swift Solar was founded in 2017 by six people who had spent their twenties making perovskite work in laboratories at Stanford, MIT, Oxford and NREL. The CEO, Joel Jean, had just helped author the MIT Energy Initiative's Future of Solar Energy report - two years of his life spent staring at exactly the kind of slow, structural barriers that academic papers are unusually well-suited to identify and unusually badly suited to remove. He decided to remove them.
The rest of the founding team - Tomas Leijtens, Max Hoerantner, Giles Eperon, Kevin Bush, and Sam Stranks - shares a CV that is almost unfair. Four Oxford doctorates. A Stanford PhD. A postdoc with Michael McGehee, who is to perovskite research approximately what Linus Torvalds is to operating systems. Among them, more than 80,000 citations. Four Forbes 30 Under 30 fellowships. The kind of group you would assemble if your goal were to publish the definitive textbook on tandem cells, rather than - as became clear by year three - actually build them.
For the first seven years the work was the unglamorous part: small grants, government awards, painstaking stability research, the slow accumulation of a patent stack now numbering more than forty. The U.S. Department of Energy backed early development. Safar Partners and a clutch of angels covered the rest. The team did not make a lot of noise. Perovskite startups that make a lot of noise tend to be apologizing about it twelve months later.
04 / The ProductWhat You Actually Get When You Open the Box
Swift Solar's product is a tandem solar cell. In the version the company is racing toward at gigawatt scale, the bottom of the cell is a heterojunction silicon wafer - HJT, as the trade calls it - and the top is a perovskite layer deposited in a thin film. Light hits the perovskite first; the photons it can't absorb pass through to the silicon below. Two bites of the spectrum instead of one.
The company has also built lightweight, flexible perovskite modules for places that conventional glass-and-aluminum panels cannot reach: vehicles, telecom towers, satellites, structures that cannot carry the weight of silicon. This is not the volume play - that is the tandem on a rooftop or a utility-scale field - but it is the play that demonstrates what the chemistry can do.
Solar cell efficiency, in theory and in the field
05 / The ProofCustomers, Capital, and an Army Microgrid
In June 2024, Swift Solar closed a $27 million Series A, co-led by Eni Next - the venture arm of the Italian energy major - and Fontinalis Partners. Stanford University joined, alongside Good Growth Capital, BlueScopeX, HL Ventures, Toba Capital and several angels. The pitch was not subtle: this is the round that takes us out of the lab.
A year later, in October 2025, the U.S. Army field-tested Swift Solar's perovskite panels in a deployable microgrid demonstration. Soldiers do not care about Shockley-Queisser limits. They care about watts per kilogram and whether the thing still works in a sandstorm. That the panels were on the demo at all is a quiet but real proof point.
Then, in March 2026, came the move that turned a research-led startup into something else. Plot twist Swift Solar acquired the U.S. manufacturing assets and the global patent portfolio of Meyer Burger - the Swiss heterojunction pioneer that filed for bankruptcy in 2025. With it came Gunter Erfurt, Meyer Burger's former CEO, and Marcel Koenig, its former global head of R&D, both joining Swift Solar to run the HJT line. The deal gave Swift a vertically integrated path: own the silicon bottom cell, own the perovskite top cell, run them through one factory.
A startup buying a public company's industrial spine is rare. A startup buying it for the express purpose of stacking its own chemistry on top of it is rarer. Industry analysts called the acquisition a pivotal moment for U.S. solar manufacturing. They were not exaggerating.
06 / The MissionAn Energy Supply Chain That Is Not in China
Roughly 80% of the world's solar manufacturing capacity currently sits in one country, and that country is not the United States. The Inflation Reduction Act has thrown money at this asymmetry; companies on the receiving end have not always known what to do with it. Swift Solar is one of a small handful that does.
Build the cells in America. Build the modules in America. Stack the perovskite layer here, on equipment whose patents you now own. Sell the panels to utilities, to data centers, to commercial rooftops, to defense buyers who would rather not place their grid on imported hardware. The company's commercial roadmap reads less like a startup deck than like an industrial policy memo.
Who actually uses this?
Today, in small volumes: research partners, a U.S. Army demonstration program, telecom infrastructure pilots, and a slow accumulation of utility-scale customers placing early orders. Tomorrow, in the company's plan: residential rooftops, commercial and industrial sites, data center campuses with hyperscale power demands, vehicle integrators, and the slightly absurd-on-paper-but-actually-very-real market of satellite power systems, where weight is everything and silicon is a tax.
07 / Why It Matters TomorrowOr: Why You Should Probably Care
Solar's cost curve has already changed the world's energy mix. Efficiency has not, particularly, because silicon's ceiling is what it is. Tandems shift the curve again - more electricity per panel, per acre, per rooftop, per dollar of installation labor. If perovskite-on-silicon scales, every project economics model in the industry needs a redraw.
There are at least four companies in the world that could plausibly be the ones to do that scaling. Swift Solar is now the one in the United States with a research lineage that goes back to the field's earliest papers, a patent stack that goes more than forty deep, and a factory's worth of HJT equipment sitting in inventory. None of which guarantees the outcome. All of which buys real options.
Back to Bing Street. The quiet is the same. The deposition chamber is still humming. The probe station still clicks. The argument about charge transport has been settled by someone going to look at the data, which is the only way arguments about charge transport are ever settled. Fifty-six people, give or take. A factory still on paper. A perovskite layer ten microns thick that, if it behaves, will quietly rearrange the unit economics of a five-trillion-dollar industry.
It is hard to picture. So is most of what works.