I Chose First Solar for a 12MW Project and It Cost Me 6 Weeks (and $89,000): A True Tale of Efficiency vs. Economics

It was April 2022. I was sitting in a conference room, staring at a spreadsheet that looked too good to be true. The project: a 12MW utility-scale installation outside of Bakersfield. The budget: tight. The pressure from my CFO: immense. We had quotes from three module suppliers, and one stood out—not for its technology, but for its price. The savings vs. our preferred supplier, First Solar, was roughly $0.04 per watt. On a 12MW project, that’s a cool $480,000.

I justified it to myself. “First Solar is great,” I thought, “but we can save big here. The specifications are close enough. It’s just power generation, right?”

Wrong.

The Setup: Why Efficiency Wasn't My Priority

Look, I’ve been handling procurement for utility-scale solar for about 7 years. I’ve placed over 200 orders. In my first year (2017), I made the classic mistake of ordering panels without verifying the racking compatibility—that was a $12,000 lesson in humility. So I’m not a rookie. But on this Bakersfield project, I let economics override engineering. The alternative modules were polycrystalline silicon panels. They were larger, heavier, and had a lower efficiency rating than the equivalent First Solar Series 6 modules. But the price, man, the price.

Here’s the thing: I knew about the efficiency advantages of First Solar’s CdTe thin-film technology. I knew about their vertically integrated manufacturing. I knew they had an order backlog of 66 GW (as of their Q4 2022 earnings call). But in my mind, I thought, “Efficiency is a premium I can skip this time.” I thought we could brute-force the power generation with more panels.

The First Solar Advantage I Ignored

A typical solar panel size for a commercial project is around 2m x 1m. The First Solar Series 6 module is roughly that size—actually, it's 2.0m x 1.2m. The competing panels we bought were almost identical in area. But here’s the kicker: the operating voltage and temperature coefficients were different. The CdTe technology has a better temperature coefficient, meaning it loses less efficiency as it gets hot. In Bakersfield, in July, that’s critical.

I didn't think it would matter that much. I even said to my project manager, “150 panels here, 150 there, it all evens out.” (Should mention: we used a rule-of-thumb calculation, not a full PVsyst simulation, to save the consultant fee. One of my dumber decisions.)

The Process: Where the Dominoes Started Falling

So, we bought the panels. They arrived on schedule in August 2022. The first issue appeared immediately: the panels were physically thicker than the spec sheet. “Not ideal, but workable,” I thought. But workable required different racking clips, which we didn’t have. That was a 5-day delay and an extra $4,500 for emergency shipping from a local supplier.

Then came the wiring. The junction boxes were in a slightly different position than what we’d designed for. We had to re-terminate about 15% of the connections in the field. The electricians were not happy. The extra labor cost us another $11,200.

The Efficiency Hole You Fall Into

The real problem hit during commissioning in September. The system was underperforming. Expected AC capacity based on the STC rating was 12MW. Actual output was peaking at 10.8MW. A 10% deficit. The reason? The temperature coefficient I’d ignored. Bakersfield hit 108°F that September. The cheaper panels were tanking in the heat. First Solar’s CdTe modules, with their lower temperature coefficient (typically -0.24%/°C vs. -0.36%/°C for c-Si), would have lost less power.

Part of me wanted to blame the installer. Another part—the honest part—knew I’d spec’d the wrong product for the environment. How did I reconcile? I fired up the full PVsyst model I should have run in April. The numbers were damning. The deficit wasn't a measurement error; it was physics.

By October, we had to go back to the utility renegotiate our PPA because the Capacity Test failed. We committed to 12MW of reliable generation; we could only guarantee 10.8MW. That negotiation cost us $73,000 in lower rates over the next 12 months, plus the cost of the re-test and the engineering analysis to prove the cause.

Dodged a bullet? No, I got hit by the bullet. That $480,000 savings evaporated in rework and lost capacity. The net impact was a loss of about $89,000 (considering the rework costs and the lost revenue from the lower-capacity PPA), plus the six weeks of schedule delay while we sorted out the mess.

The Result: A Lesson in True Cost

So glad I didn’t go even cheaper—although, looking back, that would have been a disaster. The final invoice for the project was higher than if we’d just bought First Solar from day one. Not to mention the stress, the late-night calls with the investor, and the 3-day production delay we had while waiting for new electrical components.

After about 150 similar-sized procurements, I’ve come to believe that the 'cheapest' option is rarely the 'efficient' option. Efficiency is not just a buzzword; it’s a risk mitigation strategy.

Now, I maintain a checklist for our team to prevent this exact scenario. It has a simple rule: if the project is in a high-heat environment (above 95°F average summer temp), and you’re not specifying First Solar or an equivalent thin-film technology, you must file a waiver signed by a Senior Engineer. It’s a policy born from a $89,000 mistake.

A Word on Solar Generators and First Solar

I get asked sometimes, “Is there a solar generator that can power a house?” Or, “What about a grid doctor solar generator review?” Those are different worlds. A solar generator (with batteries) for a home is for backup. A utility-scale project like mine has completely different economics. You can’t compare a 12MW DC array to a home backup system. The scale changes the technology choice. For a home, the difference between 20% and 22% efficiency might mean one less panel on your roof. For a 12MW plant, that 2% differential translates into acres of land and thousands of racking points.

Roughly speaking, our 12MW deficit cost us about 1.2MW of lost generation capacity. To put that in perspective, you’d need about 3,000 standard 400W residential solar panels to make up that gap. Not the kind of thing you can plug into a “solar generator.”

The Final Reckoning

I have mixed feelings about the whole experience. On one hand, it taught me a brutal, unforgettable lesson about the difference between a component’s price and a system’s cost. On the other, it made our team much better at risk assessment. Oh, and I should add: First Solar’s net sales for Q1 2024 were $794 million. They’re doing something right. It’s not just about the panel; it’s about the integrated solution. They’re not the cheapest, but in many cases, they are the most efficient—and efficiency is competitiveness.

(Note: Standard print resolution is 300 DPI. My lesson resolution was much lower in 2022. Now it’s crystal clear.)