Energies Media
A glass panel rolls off a conveyor belt in Fremont, California, still warm from the line behind it.
It looks like any other solar module, dark and rectangular.
But under its surface sits a second, invisible layer of crystal that was never there before.
That layer is why a small production line in a converted Bay Area building now matters to an entire industry.
So how does stacking two materials into one panel actually squeeze more power out of the same patch of sunlight?
A standard silicon solar cell is tuned to catch a narrow slice of the sunlight spectrum well, and it lets the rest pass through as wasted heat. The new approach adds a second, extremely thin light absorbing layer made from a crystal structure called perovskite, stacked directly on top of ordinary silicon.
That top layer is built to catch the higher energy, bluer end of sunlight that silicon handles poorly, while the silicon underneath keeps doing what it already does best with the red end. Because the two layers split the job instead of competing for it, the combined cell converts more sunlight into electricity than either material could alone.
The company behind the Fremont line, Tandem PV, says its cells have reached nearly thirty percent efficiency in internal testing, well above the raw ceiling of a conventional silicon cell. Stack that kind of efficiency gain onto the same roof or the same acre of desert, and a developer gets more megawatts without buying more land or more glass.
The facility itself is unglamorous: a 65,000 square foot building that most drivers on the surrounding streets would mistake for a logistics warehouse.
Inside, the production line threads silicon wafers through a sequence of coating and deposition steps that lay down the perovskite layer in a film far thinner than a human hair. Workers in clean suits watch monitors tracking temperature and vapor levels, because the layer is sensitive to exactly the kind of humidity and heat swings a California summer brings through a loading dock door.
That sensitivity is the main reason tandem technology spent so long stuck in small batches in university labs rather than running continuously on a factory floor. Getting that process to hold steady, batch after batch, is the real achievement behind the opening.
Tandem PV’s chief executive, Scott Wharton, described the shift plainly, saying the factory marks “the shift from impressive R&D results to repeatable manufacturing at a commercially meaningful scale.” That line, more than any spec sheet, is the claim being tested now.
The company opened the demonstration factory and began producing panels in April this year, according to trade press that toured the site. In the 65,000-ft2 facility, Tandem PV is producing tandem perovskite-silicon solar panels, and the line can accommodate 40 MW of annual capacity.
That is a modest output next to the gigawatt scale lines running in Asia. But it is a real, measured production run rather than a laboratory sample or a press release about a future plan.
The company said it would begin sending shipments to support customer validation trials later this year, anticipating its first commercial panel sales from the facility in 2026. Those trials matter because they are where a developer decides whether a new panel design can survive two decades on a rooftop, not just a test bench.
Even inside the company, nobody claims the technology has crossed every hurdle. Perovskite material is famously sensitive to moisture and ultraviolet light over time, and the industry’s hardest unresolved question is whether a tandem panel can match silicon’s quarter century of proven field reliability.
Competing efforts are moving just as fast elsewhere, which keeps Fremont from claiming the field outright. A gigawatt scale line opened in China the previous year, and other manufacturers have licensed similar tandem technology from a British developer that shipped its own first commercial modules back in 2024.
So Fremont is not the first tandem factory on Earth, and it is not yet the biggest. What it demonstrates is a domestic, US based line that can run a tandem process start to finish without importing the finished cell.
Scale alone rarely tells the whole story in solar manufacturing, and this is a case where the smaller number carries outsized weight. A 40 megawatt line proves a manufacturing process works at a size that can be inspected, audited and copied, long before anyone commits capital to a gigawatt scale plant.
That same logic has played out elsewhere on this desk, including an AI platform clearing review before wider deployment, where a smaller qualifying step unlocked a much larger rollout. Manufacturing credibility, once earned at small scale success, tends to compound quickly once customers start asking for proof rather than promises, as trade coverage of the opening made clear.
The panels now moving through validation trials will spend the next year on test racks, exposed to real weather rather than a climate chamber. If they hold their efficiency, Fremont’s modest line, much like the sensor work behind tracking devices built for a different industry, will be remembered less for its size than for being the place a stubborn lab idea finally learned to behave on a factory floor.
Hugo is an engineer with strong technical expertise. Multilingual from an early age, his writing combines technical clarity with a strong interest in science and energy.
Hugo is an engineer with strong technical expertise. Multilingual from an early age, his writing combines technical clarity with a strong interest in science and energy.
Hugo is an engineer with strong technical expertise. Multilingual from an early age, his writing combines technical clarity with a strong interest in science and energy.

source