The Solar Squeeze Sits in the Cell

Building a solar module factory in America is now the easy part. The scarce thing is the cell, the silicon or thin-film wafer that does the actual work, and the next round of clean-energy politics will be fought over who makes it.

SEIA's September 2026 factsheet, relayed by TaiyangNews, puts operational US module capacity at 75.3 GW, with 14.4 GW more under construction. Cell capacity looks different: 10.6 GW operating and 19.1 GW under construction. Before federal manufacturing incentives arrived, SEIA's dashboard says, the country had 7 GW a year of module capacity. Assembly came home first because it is cheap, quick and tolerant of imported parts. Cells are capital-hungry, and that is where the strategic exposure sits.

Capital finds the chokepoint

SEG Solar's latest announcement shows where the money is moving. The company said on October 29 that it has secured land near Houston for a $950 million cell plant at its Tomball campus, designed to turn out 4 gigawatts of heterojunction cells a year, according to Oil & Gas 360. Production would start in two phases, targeted for 2028 and 2029, with more than 700 jobs promised. The site sits beside the 4 GW Tomball #1 module line and the 4.6 GW Tomball #2 plant now under construction.

The pitch is domestic content. SEG says modules built from its own cells would help developers and asset owners meet domestic-content requirements, Big Energy reported. "American-made cells, American-made modules, one company accountable." CEO Jim Wood said. Oil & Gas 360 adds a caveat worth keeping: the cell plant would supply components for finished panels, not add another 4 GW of module output. Big Energy notes that SEG disclosed no detailed construction schedule and named no public funding.

First Solar, meanwhile, is playing a different hand. It opened its Gaffney, South Carolina, plant, its sixth US factory, with Commerce Secretary Howard Lutnick and Governor Henry McMaster at the ribbon, pv magazine reported. The first phase is running. The second, due by mid-2027, lifts the $330 million site from 1.7 GW to 3.4 GW. First Solar credits rising domestic demand after federal trade policy and the One Big Beautiful Bill Act, signed in July 2025. "Every module that leaves this facility is made without dependence on Chinese crystalline silicon supply chains." CEO Mark Widmar said. Yet PV Tech notes that Gaffney finishes its Series 6 Plus modules from thin-film cells made at the company's overseas sites. Even the most self-reliant US manufacturer still imports the cell stage.

Trade policy has chosen its target

Imports explain the urgency. Wood Mackenzie says India, Indonesia and Laos, together with Malaysia, Thailand and Vietnam, supplied 78% of US cell imports last year. The administration's answer is a 15% tariff plus minimum import prices under Section 232 covering polysilicon, wafers, cells and modules, with indicative floors of $0.22/W for cells and $0.38/W for modules. TaiyangNews reports the floors take effect December 4, 2026.

That structure favors the vertically integrated. Wood Mackenzie, per TaiyangNews, believes it could strengthen fully integrated domestic suppliers while making stand-alone cell and wafer projects harder to finance. This is the real test of the thesis: a floor price protects a factory's revenue, but only if the factory gets built. SEG's plan is a bet that integration, not a bare cell line, is the survivable model.

The counter-argument: demand is the weak link

The obvious objection is that factories mean little without buyers, and the buyers are being squeezed elsewhere. TaiyangNews reports that US installations reached 11.4 GW DC in the second quarter, up 45% year on year, driven by utility-scale developers rushing the July 4, 2026 safe-harbor deadline, per SEIA and Wood Mackenzie. They expect annual additions to hover around 44 GW DC after lifting their 2026-2031 outlook by only 1.2%. Permitting problems, they say, will hurt about 30% of the early-stage pipeline.

Wind shows how quickly politics can strand supply. Bloomberg reported, via Energyconnects, that BloombergNEF sees about 7 gigawatts of US onshore wind this year, down from 17 gigawatts in 2020. Absent big policy changes, it expects no offshore farms to start operating beyond those already under construction. The administration asked courts to let it reconsider three Biden-era offshore projects, then cleared five wind developments in one week. Capstone analyst Paul Greenough called the approvals "This is a bargaining chip." Senator Sheldon Whitehouse, a chief negotiator of the permitting bill, said: "This is how you blow up a permitting deal." The compromise would bar federal interference with projects that hold all their approvals.

Markets are not giving credit yet. Market data shows the Invesco Solar ETF, TAN, closing May at $73.93 and October at $41.18. First Solar's FSLR was last seen at $176.89 at Thursday's close.

Still, I think the read holds. The permitting fights and the demand wobble are cyclical and political; they can reverse with a deal or an election. The cell shortage is structural. Reuters reported, via Dailysabah, that IRENA's chief expects the world to add less clean capacity in 2026 for the first time in five years, blaming inadequate transmission and trade barriers. That follows a record 692 gigawatts of clean power added in 2025, 15.5% growth to 5,149 GW. China turned away enough clean energy in the first half of 2026 to power Mexico for a year, while networks in Australia and Japan also lagged their buildouts. Scarcity, in short, is moving from the panel to the wire and the wafer. A country that wants domestic content has to make the part everyone else makes cheaper.

If that is right, the winners will be companies that own the cell and can survive a floor price, not those that assemble imported ones. SEG will break ground on a lot in Tomball, where the module lines already stand waiting for something local to feed them.