Technical note

First Solar vs. the $/Watt Trap: A $1.8M Procurement Lesson

2026-11-06 · Kofi Asante

If you're still comparing utility-scale solar modules primarily by $/Watt, you're making the same mistake I made in 2019. That year, my team lost roughly $1.8M on a single 120 MW project—not because we bought bad modules, but because we bought cheap ones and paid for it everywhere else.

Quick heads-up: if you landed here searching for "when is the first solar eclipse of 2026," that's a different kind of solar (February 17, 2026—an annular eclipse, mostly visible near Antarctica). This article is about the kind that shows up on an EPC budget sheet.

I've been handling procurement and project delivery for utility-scale PV projects for about seven years. In that time, I've personally made four procurement decisions that I now document as cautionary tales for new hires. The 2019 project was the worst. It's also the reason I keep a TCO checklist taped to my monitor.

Here's the short version of my argument: for utility-scale solar, $/Watt is the least useful metric in the entire procurement process. Not useless—just the wrong starting point. Total cost of ownership (TCO) is what actually determines whether a project makes money over 25 years. And most procurement teams I've worked with still don't calculate it.

The $1.8M Comparison That Changed My Approach

In early 2019, I was managing procurement for a 120 MW project in the Southwest U.S. We had two module options on the final table:

  • Option A: A tier-1 crystalline silicon (c-Si) supplier, quoted at $0.21/W.
  • Option B: First Solar Series 6 thin-film modules, quoted at $0.24/W.

On paper, Option A saved about $360,000 on the module line item. I went with it. My VP approved it in under an hour.

Then the project started, and the savings evaporated.

The c-Si modules were heavier—roughly 15% more per panel. That meant more steel in the racking and slower installation. We budgeted 3.2 labor-hours per kW installed. Actual was closer to 3.9. That overrun alone ate the $360K.

But the bigger cost showed up in financing and operations. The performance model I'd built assumed a certain annual degradation curve. Actual output in year one came in about 2% below projections—partly temperature effects, partly my own modeling error. Over 25 years, that gap compounds into a meaningful LCOE hit. (And yes, I've since apologized to our asset management team. Repeatedly.)

When I later compared that project side by side with a nearly identical 110 MW project that used First Solar modules—same racking vendor, similar climate, same O&M contractor—the difference was clear. The thin-film project had lower balance-of-system costs, faster installation, and better actual output in high-heat conditions. The higher sticker price turned out to be the cheaper option by a wide margin.

That's when I finally understood why the TCO framework exists. I'd been optimizing the wrong number.

What Actually Goes Into Solar TCO (It's Not Just the Module)

If you've ever priced out a Level 2 EV charging station in Virginia or specced an emergency power inverter for a commercial site, you already know the pattern: the box on the quote isn't the whole cost. Solar modules are the same, just with more zeros.

Here's the TCO checklist I use now. It's not exhaustive, but it covers the items that actually move the needle on a utility-scale project:

  1. Module price — the $/Watt number everyone fixates on.
  2. Balance of system (BOS) — racking, cabling, inverters, and the labor to install them. Module weight, dimensions, and connector type all affect this line.
  3. Installation labor — measured in hours per kW, not dollars per panel. Heavier modules and slower crews change this number fast.
  4. Degradation and energy yield — not just nameplate efficiency, but the actual annual output curve over 25–30 years. A 0.1%/year difference in degradation rate is real money at utility scale.
  5. Financing cost — lenders price risk. A module with a well-documented degradation curve and a large installed base can lower your cost of debt. First Solar's 2023 10-K (available as a PDF on their investor relations site) lists a 66 GW backlog—that kind of number matters to lenders evaluating bankability.
  6. O&M — cleaning frequency, soiling losses, and replacement rates all vary by technology.
  7. Time-to-energization — every week of delay costs money. I've seen projects burn $40K–$60K per week waiting on the wrong module spec.

Notice what's missing from that list: a single "lowest price wins" column.

Also—if a vendor's pitch leans hard on "eco-friendly" claims, it's worth checking them against the FTC Green Guides (16 CFR Part 260). Terms like "recyclable" have specific definitions, and unsubstantiated environmental claims are an enforcement priority. I'm not saying anyone in the solar industry has a problem with this. I'm saying I read the Green Guides once and it changed how I read spec sheets.

The Time-Pressure Problem

I get it—procurement deadlines are real. In 2022, I had 48 hours to lock a module order before a price increase kicked in. Normally I'd run a full TCO comparison. Instead, I went with our existing supplier based on gut feel and a spreadsheet I threw together in an afternoon.

In hindsight, I should have pushed back on the timeline. The "price increase" turned out to be a two-week window that extended anyway, and I locked in a module spec that added about $200K in racking costs because of a dimensional quirk I hadn't modeled.

That's the trap. Urgency makes $/Watt feel like the only number you have time to check. But the cost of a bad decision scales with the size of the project, not the speed of the decision.

An Analogy That Stuck With Me

My brother-in-law is a mechanic. He once told me that disconnecting a car battery to reset the check engine light takes about 10–30 minutes. It works—sometimes. But it also wipes the ECU's learned adaptive data, which can make the car run rough for a few drive cycles and occasionally fail an emissions test. The "fast" fix creates three slower problems.

That's exactly what buying on $/Watt does to a solar project. You save a few hundred thousand dollars upfront and spend the next two years dealing with BOS overruns, financing questions about your degradation assumptions, and irate asset managers wondering why year-one production is under budget.

I'd rather spend an extra two days on a TCO model than two years explaining why I didn't.

"But Thin-Film Efficiency Is Lower—Doesn't That Matter?"

Yes, it matters. To be fair, crystalline silicon modules generally have higher conversion efficiency per panel—meaning more watts per square meter. If you're constrained on land or roof area, that's a real advantage, and I'd look closely at c-Si in those cases.

But for utility-scale ground-mount projects—the kind I work on—land is usually the least constrained variable. What matters is energy produced per dollar invested over the project life, not watts per panel. First Solar's Series 7 modules, for example, are up to about 545W with roughly 19.7% efficiency. That's competitive, though not record-breaking. The bigger lever is the low degradation rate (around 0.5% per year) and the performance in high-heat conditions.

So no, thin-film isn't universally "better." I'm not arguing that. I'm arguing that the right technology depends on your project's actual cost drivers, not a single spec-sheet number.

One Caveat Before You Run Off and Rebuild Your Spreadsheet

My experience is based on roughly 30 utility-scale projects in the U.S., mostly in the Southwest and Southeast, and mostly in the 80–300 MW range. If you're doing rooftop commercial, community solar, or projects outside the U.S., the TCO math may look different—land costs, labor rates, and interconnection rules all change the equation.

I've also only worked with a handful of module suppliers in depth. I can't speak to every technology on the market, and I'm not claiming First Solar solves every problem. What I can say is that the framework—calculating total cost instead of sticker price—has saved my team more money than any single vendor relationship.

Bottom Line: Stop Comparing Watts. Start Comparing Lifecycle Cost.

The next time you're staring at a module comparison spreadsheet, don't sort by $/Watt. Sort by TCO. Include the racking, the labor, the degradation curve, the financing implications, and the timeline risk. Maybe even build in a contingency line for "things I haven't thought of yet." (Mine is 8%. It used to be 3%. I learned.)

If you don't do it, your finance team will eventually do it for you—after the project is already built. And I can tell you from experience: that conversation is a lot less fun than the one where you show up with a TCO model and a defensible recommendation.

Take it from someone who's already made the $1.8M mistake so you don't have to.

Previous: Commercial Emergency Backup Power: A Scenario-Based Buying GuideNext: Vivint Solar Ratings and Pros/Cons for Commercial Buyers: A Cost Controller’s Shortlist