First Solar, Inc.

On reported numbers First Solar looks cheap: 7.7x FY2027E consensus EPS and 6.5x guided EBITDA for a company with net cash. Our P/E basis is FY2026E and FY2027E consensus EPS as of October 7. The last three lines of the table qualify that: the credit is about six times the gross profit the business earns without it, and the backlog behind the earnings is shrinking.

First Solar has round-tripped twice this year. From $195.57 on April 6 it ran to a $318.25 closing high on June 3 ($320.95 intraday the same day), helped by a GLJ Research upgrade to Buy with a $315 target on May 27; the stock rose 10.9% the next day. By July 29 it was back at $199.24. The Section 232 proclamation, signed August 6, lifted it to $250.05 on August 7. Then it gave everything back: on September 24 it fell 10.3% to $172.16 as solar stocks sold off on higher borrowing costs, and on October 1 it set a $172.11 closing low ($168.60 intraday).

At $179.80 the stock is 12.9% below its 50-day average ($206.46), 19.5% below its 200-day ($223.45) and down 31.2% this year. The trend is down on every horizon we track. The rate move matters most here: the 10-year yield went from 4.49% on the day of the peak to 5.31% on October 5, and utility-scale solar is one of the most financing-sensitive buyers of anything.

Section 45X pays US manufacturers a fixed amount per watt made and sold. First Solar says a module fully produced in the US qualifies for up to about $0.17/W. In 2026 that comes to $2.10B–$2.19B, more than the company's entire expected operating income. The credit turns into cash: First Solar has sold credits to third parties at about 95% of face value, and the Treasury has paid it directly. It is in statute through 2032 and steps down from 2030. But it pays the same whether a module sells for $0.30 or $0.36 a watt, so it says nothing about the business underneath.

Valued as what it is, a government payment that depends on volume, the remaining 2026–2032 credit stream is worth about $5.07B for 2026–2028 and $4.35B for 2029–2032. That uses an 8% discount rate, the 2030–2032 step-down and the roughly 4.5% haircut First Solar has accepted on credit sales. Add $1.26B of credits already earned but uncollected at June 30 (up from $625.2M at year-end; the 10-Q discloses no sale of 2026 credits) and $1.7B of net cash. That is ~$12.33B, or ~$114 a share. It is a soft floor: the ~$40 a share in 2029–2032 credits depends on sales that are mostly not yet booked.

The remaining ~$65 a share is the manufacturer. Its backlog fell from 68.5 GW at the end of 2024 to 45.1 GW at June 30, partly by choice: in 2025 it recorded 8.3 GW of debookings, mostly by terminating contracts after customer breaches, against 7.4 GW of gross bookings. The price did not move. The backlog has carried about $0.30/W all the way down. New US bookings came in at $0.364/W (reported in February), about $0.35/W (April) and about $0.36/W (July). Every gigawatt that rolls off at $0.30 and is replaced at $0.36 adds about $60M of revenue with little added cost.

So what does $179.80 assume? Putting a 12x multiple on unsubsidized after-tax operating profit, the market is paying for about 5.5¢ per watt of ex-credit operating margin by 2029. Guidance implies about −⁠1.5¢/W for 2026, so that is a swing of roughly 7¢. Repricing the US book from $0.30 to $0.36/W is worth about 5¢/W across all volume. CuRe technology adjusters (up to $600M of potential revenue, mostly in 2027–2028) and the end of this year's underutilization ($115M–$135M) and start-up costs ($90M–$100M) add more. Rising costs take some back: management described a "pretty challenging rising commodity cost environment" and said shipping from Ohio to the West Coast now costs about as much as shipping from Asia. The swing is achievable, but the current price already assumes it.

First Solar makes one thing, cadmium-telluride thin-film modules, and sells them mostly to US utility-scale developers. It reports a single segment; 95.7% of 2025 net sales ($5.2B) came from the United States. In Q2 it produced 4.3 GW and sold 3.7 GW. Revenue is lumpy because shipments and contract-termination payments are lumpy: Q2 net sales of $1.06B were down from $1.10B a year earlier, mainly because the year-ago quarter included termination revenue.

The competition is crystalline silicon, most of it Chinese-owned or Chinese-supplied. First Solar's advantage in the US comes from policy and from what its modules are made of. About 41 GW of its 45 GW backlog carries a domestic-content requirement. Its modules contain no polysilicon, the material targeted by the new Section 232 tariff, and management argues that the foreign-entity rules now restricting credits for Chinese-linked supply favor it for projects completing in 2029–2030. It is also enforcing TOPCon cell patents it acquired with TetraSun in 2013. After the Section 232 action it withdrew its ITC Section 337 complaint (September 16) to pursue district-court cases against Canadian Solar and JinkoSolar affiliates, and on October 1 it sued JA Solar and American Panel Solutions in Delaware.

Profitable US solar peers trade at richer multiples: Nextracker at 18.2x and Enphase at 15.4x FY2027E consensus EPS, against 7.7x for First Solar. The discount reflects how much of First Solar's profit is the credit.

We discarded a standard DCF built on reported free cash flow. Free cash flow at First Solar follows the timing of credit sales, not the business: about $2.14B of inflow in H2 2025 and an outflow of $640M in H1 2026. Instead we work backward from the price, separating the credit and solving for the ex-credit margin the remainder implies.

The grid values the core at 2029 ex-credit operating margin (rows, 17.5 GW, 21% tax) times a multiple of after-tax operating profit (columns), discounted 2.25 years at 12%, plus the ~$114 floor. Our base case of 6¢/W at 12x gives $186. That margin assumes repricing to about $0.36/W, CuRe adjusters and no repeat of this year's start-up costs, partly offset by cost inflation. Note that $23.33 of FY2027E consensus EPS at the same $186 is 8.0x.

Probability-weighted value: $183 (+2% vs the reference price).

Chief Executive Mark Widmar and Chief Financial Officer Alexander Bradley keep a conservative balance sheet. At June 30 First Solar held $1.69B of cash and equivalents, $38.7M of marketable securities and $214.3M of restricted marketable securities, against $37.6M of current debt, having prepaid the last $328.2M of its India credit facility in May. In February it signed a five-year $1.5B unsecured revolving credit facility with a $1.0B accordion. Customer deposits are a second cushion: current deferred revenue was $1.18B. There is no buyback or dividend. Cash goes to capacity (2026 capex guidance $0.8B–$1.0B; first-half spending went mainly to South Carolina), perovskite development and working capital. Net cash fell from $2.4B at year-end to $1.7B at June 30. In our view, reaching the $1.7B–$2.3B year-end guidance requires selling or collecting this year's credits.

Insiders sold about $15.5M of stock since early April and bought none. About $14.4M of that came under Rule 10b5-1 trading plans; the other $1.1M was stock the company sold on executives' behalf to cover taxes when restricted stock vested. Widmar's plan dates to November 2025, well before the spring rally. Three of the six largest sellers adopted their plans in May 2026, during it. The table shows those six at Form 4 transaction prices, including their tax-withholding sales.

The top risk is the one the market watches least. The credit floor is only as solid as the volume behind it, and the volume after 2028 is not yet sold.

Options price a move of about ±9.6% into the October 30 expiry, which captures the expected October 29 report. Call OI exceeds put OI (383,100 calls against 296,892 puts, a put/call ratio of 0.77), so positioning is not bearish. Short interest at 10.4% of float is elevated for a profitable large cap but has not squeezed. Gamma is concentrated at $180, which should dampen moves near the current price. Below $170 the open interest turns put-heavy, so a break of the October 1 low could speed up.

The Street is bullish and drifting lower. The consensus mean target is $267.31 (median $267.50, range $197–$324) as of October 7, with 44 Buy, 23 Hold and 7 Sell ratings. FactSet's poll showed a $280.74 mean on September 30. Our $186 target is 30% below the consensus mean. At $267.31 the stock would trade at 11.5x FY2027E consensus EPS, a multiple that treats the credit as permanent earnings when it steps down from 2030.

What would move us to BUY. Net bookings that refill 2029–2030 at $0.36/W or better (the roughly 4 GW pipeline converting), a profitable restart of the 1.8 GW of idle Asian capacity under the Section 232 regime, or 2026 credits sold at 95% or better so year-end net cash lands in the upper half of guidance. Any two would lift our margin assumption to 8¢/W and the target to about $210.

What would move us to SELL. Renewed net debookings or new US pricing below $0.33/W; a legislative move to cut or accelerate the phase-down of 45X; or a 10-year yield held above 5.5% long enough to push project starts beyond 2030. Each of these hits the credit floor, not just the multiple.

HOLD, 12-month target $186 (+3% versus the $179.80 close on October 6). First Solar is not as cheap as its P/E suggests, and not as broken as its chart suggests. About two-thirds of the price is a government receivable, and the rest already assumes the backlog reprices. The probability-weighted value is within a few dollars of the price. For a momentum investor there is no trend to own here. The next checkpoints are the ITC vote on October 14 and the Q3 bookings number expected October 29: if the roughly 4 GW pipeline converts at $0.36/W, the largely unbooked 2029–2032 volume starts to fill, and so does the floor.