Business

Know the Business

First Solar is a per-watt commodity manufacturer that has been temporarily transformed into a high-return franchise by U.S. industrial policy. Its proprietary CdTe thin-film technology is genuinely differentiated, but the 30%+ operating margins of FY2024–FY2025 are inseparable from Section 45X production credits and tariffs that shut Chinese silicon out of the U.S. market. The market is most likely overestimating how durable that policy moat is, and underestimating how much of reported earnings is essentially a transferable tax-credit subsidy.

1. How This Business Actually Works

First Solar makes one product — a glass-encased CdTe solar module — and sells it by the watt to utility-scale developers. Volume, average selling price (ASP) per watt, and U.S. tax credits per watt are the only three levers that matter for gross profit.

The 45X advanced-manufacturing credit (set at 7¢ per watt at the module level, plus cell and wafer credits) flowed roughly $602M through cost of sales in 2025 and produced $1.34B of cash when First Solar sold the credits to third-party taxpayers. Strip the credit out and gross margin is still respectable at ~29%, but most of FY2025's incremental margin uplift versus the FY2022 trough comes from policy, not from the technology getting cheaper.

The customer base is concentrated and contracted: as of December 2025, 50.1 GW of backlog priced at $15.0B — an average of about $0.30 per watt through 2030 — versus global Chinese ASPs near $0.10/W. Two customers (Silicon Ranch, NextEra) each accounted for over 10% of net sales in 2025, and 96% of revenue came from U.S. projects. When BP/Lightsource walked away from contracts in 2025, the dispute was over $384M of termination payments — a reminder that booked backlog is only as solid as the buyer's project economics.

Incremental profit comes from three things in this order: (1) U.S. capacity utilization — fifth U.S. plant ramped in 2025, sixth coming H2 2026; (2) holding ASP-per-watt firm in dollar terms while Chinese ASPs decline; (3) collecting 45X on every U.S.-produced watt. The vertical CdTe process — glass in, finished module out in hours, no polysilicon — is what makes (3) possible at all, since 45X requires that the wafer/cell/module steps actually happen domestically.

2. The Playing Field

First Solar is the only solar manufacturer earning real money in 2025. Every Chinese-linked silicon peer is in operating-loss territory; the polysilicon producer (DQ) has gross margins below zero; the U.S./Israeli inverter peers (ENPH, SEDG) are surviving the residential demand reset, with only Enphase profitable.

The peer set tells you what "good" looks like in solar today: positioning matters more than scale. JinkoSolar generated nearly 2× First Solar's revenue and lost ~$617M; First Solar generated ~$1.5B of net income on smaller volume. The vertical Chinese-silicon model — DQ → JKS/LONGi → modules — is in a multi-year deflationary collapse, and the only profitable seats in solar in 2025 are (a) U.S.-domestic, IRA-eligible module manufacturing and (b) hardware with brand and software lock-in (ENPH). First Solar has the first; ENPH has the second; nobody else has either.

What the best Chinese peer (Jinko) does better: scale and per-watt cash cost. CdTe is not cheaper to produce on a fundamental basis. First Solar's pricing premium is a U.S.-policy premium, not a technology cost premium.

3. Is This Business Cyclical?

Brutally so, and the cycles are policy- and ASP-driven, not demand-driven. Demand for utility-scale solar has grown almost every year since 2005; First Solar's profit has not.

The cycle hits at three places: module ASP per watt (the headline price), capacity utilization (which determines absorption of fixed costs and the start-up expense line), and inventory writedowns (when ASPs fall faster than committed inventory clears). Working capital tightens hard — receivables aged out in 2025 even within the IRA boom because BP/Lightsource walked away from $385M of contracted purchases. The capital-markets piece matters less for First Solar specifically because the balance sheet has been net-cash since 2017; it matters enormously for the leveraged Chinese silicon peers.

4. The Metrics That Actually Matter

Five numbers explain almost everything about this stock. Conventional ratios — P/E, dividend yield, asset turnover — are nearly useless because they conflate pre- and post-IRA profit pools.

The metric you should not use: ASP in foreign currency or before tax credits. Reported GAAP numbers already include the credit; if you're comparing First Solar to a Chinese peer, you must add back the credit to compare module-level economics, and you must subtract it again to compare cash earnings.

5. What I'd Tell a Young Analyst

This is a policy-arbitrage business with a real but narrow technology edge. Don't model it as a tech compounder. Model it as a U.S. industrial manufacturer whose returns track three things: 45X eligibility, tariffs against Chinese and Southeast Asian modules, and the dollar ASP of long-dated U.S. utility contracts. Roughly 35–40% of FY2025 net income is the 45X credit; if Section 45X gets revoked or narrowed (the OBBBA already added foreign-entity-of-concern restrictions in 2025), the earnings power on this order book drops by a third and the multiple compresses with it.

What to watch quarterly: dollar ASP on new bookings versus the $0.30/W backlog average; backlog cancellations and price-adjustment clauses (23.2 GW of the book has technology-improvement-linked price adjustments worth up to $0.6B); Series 7 warranty experience after the 2024–25 manufacturing issues; and the cadence of 45X credit monetization, which is now the single largest cash-flow line. The thesis genuinely changes if (a) the OBBBA accelerates the ITC/PTC phase-out enough to depress U.S. utility-scale demand, (b) Chinese tellurium export controls bind on CdTe inputs, or (c) a peer manages to qualify for 45X at scale and competes for the same domestic backlog.

What the market may be missing in either direction: backlog visibility through 2030 at premium ASPs is real and largely durable as long as the existing tax credits stay; but 96% U.S. concentration and two-customer dependence are not the diversified, defensive profile that "renewable energy infrastructure" framing implies. Treat this as a high-margin cyclical with a five-year political tailwind, not a 20-year secular winner.