First Solar's (FSLR) fiscal 4Q11 and 2011 results came with a couple of unexpected surprises — including a big nasty one related to performance issues — and industry watchers are raising voices for what could be a radical shift in the company's business.

Here's a bullet-point summary of FSLR's fiscal 4Q11 and full-year 2011:

  • Revenues: $660.4 million (-34 percent Q/Q, +8 percent Y/Y). The Q/Q loss was blamed primarily on timing of revenue recognition on the systems business and softer module sales.
  • Gross margins: 20.9 percent (was 37.7 percent in 3Q11, 48.7 percent in 4Q10)
  • Net loss: $413.1 million (vs. profits of $196.5 million in 3Q11 and $155.9 million in 4Q10), EPS (-4.78). (Barrons makes the unfortunate but accurate observation that FSLR's 4Q11 losses equal 20 percent of the company's market value.)
  • "Core" cost/watt lowered to $0.71 (including $0.01 for warranty costs), the first stepdown in more year after being stagnant at $0.73. Total cost/W produced inched down a point to $0.73.
  • Conversion efficiency raised to 12.2 percent, up from 11.8 percent.
  • Megawatts (MW) produced dipped 2 percent to 539.7 MW.
  • For full-year fiscal 2011, FSLR reported revenues of $2.8 billion and a net loss (EPS) of ($0.46).

Included in the 4Q11 net loss was $485.3 million in charges that were announced in the December quarter. Within that was a big surprise: nearly $164 million in warranty-related charges for replacement modules remediation/compensation to customers. That one-percentage-point increase (going back to a 2008-2009 excursion) attempts to compensate for increased failure rates foreseen for hotter climates that the company is targeting, noted CFO Mark Widmar. (More on that below...)

Here's what FSLR is now guiding for 2012 (vs. early-December expectations in parentheses):

  • Revenues revised downward: $3.5 billion to $3.8 billion (vs. $3.7 billion to $4.0 billion), and operating cash flow down to $800 million to $900 million (was $900 million to $1.1 billion)
  • Reiterated with no change: Operating income $425 million to $450 million; earnings per share (EPS) $3.75-$4.25; and capex $375 million to $425 million.
  • Reiterated 2012 installations: 1.2 gigawatts (GW) DC installed, with 95 percent under contract, though only 300 to 500 MW DC in modules (down from 720 MW DC forecast).
  • Production outlooks lowered: 1.5 to 1.8 GW (from 2.0 GW), and 60 to 70 percent utilization (from 80 percent).
  • Average module costs $0.74, up from $0.72, including $0.07 impact of underutilization.
  • Average module efficiency raised a point to 12.7 percent.

The company had been expected to reveal more details about how it plans to shift away from today's major subsidized markets (what it called a frustrating "whack-a-mole" game) and focus more on stable, emerging markets unencumbered by the volatility of subsidization. FSLR now says those details will be forthcoming in the next couple of months.

Among the things it did reveal about 2012, however, was that it's idling four lines in its German plant "for up to six months," and it will stick to its earlier plan to postpone its Mesa, Ariz., facility and "discontinue work" on its proposed plant in Vietnam. Asked during the results conference call why the company doesn't just shut down manufacturing capacity outright vs. temporary idling, Widmar said it "makes the most sense" given current market expectations, though he kept that as a possibility should visibility or market recovery erode getting closer to 2013.

Considering the flux in Germany's solar market, chairman/interim CEO Michael Ahearn was asked about the US utility-scale market, and admitted that it's "not nonexistent but sporadic and not at particularly high levels for the next several years."

Analysts' take: Mostly bad with some good

Jesse Pichel with Jefferies called it "a near-kitchen-sink quarter" encompassing lots of bad news of missed estimates and some lowered key forecasts, though improvements on the balance-of-systems side are encouraging. (Satya Kumar with Credit Suisse, though, thinks the system EPS guidance "was too conservative to begin with.")

Tim Arcuri respectfully calls the results "better than [they] could have been," pointing to the company's ability to maintain EPS guidance and reduce core balance-of-system costs even as revenues are seen falling (blaming Germany).

Maxim's Aaron Chew takes issue (once again) with FSLR's liquidity, which dropped from $184 million to $124 million in 4Q11, as "an overlooked risk" and "an overarching concern." Given the anemic module business, the company might resort to acquiring projects outside its own pipeline, he suggests. "Our bias has shifted to the downside," he said, saying the company's future "lies in utility-scale development not module sales."

And on the far bearish end (where it's been for a while), Cantor Fitzgerald called the results "awful" and said it has "no confidence in the company's forward guidance" — and warns that "2013 will be worse than 2012."

Performance problems looming

Those warranty problems — which have amounted to an overall cumulative charges of $253 million — are a very big deal. Avian Securities analyst Mark Bachman, quoted by Barrons, points out those charges are "about 10 times what they said they were going to be when they first reported the issue."

Credit Suisse analyst Satya Kumar goes even further, hammering home the importance of product quality as "the most significant metric" for solar panels, "and doubly so for thin-film panels" whose 20-year performance track record lags behind crystalline silicon (c-Si) panels. Field reliability of thin-film panels is less proven, and high temperature degradation of cadmium-telluride (CdTe) panels is known and understood (he cites an NREL study). It's the most important metric for a company's long-term survivability, he writes.

Kumar worries that "this may not be the last time we hear of the warranty related issues" for FSLR, and points out that SunPower's panels have lower thermal degradation (due to use of n-type silicon), while degradation for P-type silicon used commonly in c-Si panels is "probably lower than thin-film" as well. As Deutsche Bank's Vishal Shah points out, the areas of growth FSLR is targeting for current and future projects — the US, and increasingly nonsubsidized markets in India and Africa and Asia — are exactly those high-temperature areas where this degradation is becoming a problem.

Time for a radical idea?

During the results call, Ahearn underscored what the company sees as its value proposition: designing and building solar PV systems, with an advantage partly from its own modules but also through doing installation directly including turnkey systems. "We're seeing this as a holistic offering," he said.

But analysts increasingly aren't seeing it that way. Arcuri takes a jab at the company's "captive lifeline — er, we mean pipeline — [which] continues to mask massive losses in its core merchant module business." He also interestingly notes that FSLR's projected 400 MW module sales for 2012 is less than CIGS player Solar Frontier.

Assuming FSLR's lower-cost benefits don't greatly outweigh efficiency performance benefits (and its ballooning reputation problem with warranty issues), maybe it's time for a radical proposal: FSLR should "become more technology 'agnostic' and open up the possibility to use c-Si or even CIGS products in its captive development business," Arcuri suggests. The company "now admits that selling CdTe modules as components is no longer the company strategy, and certainly not profitable," adds Pichel.

FSLR isn't quite ready to make that leap just yet, though, explained Ahearn during the call:

"Our modules cost less and perform better than crystalline silicon, so it wouldn't make any sense for us to use crystalline silicon modules. I mean, a significant part of our competitive advantage is in our manufacturing costs, and where we intend to be 3 years from now would be substantially below even the cash costs of a silicon module to our best estimate. So that really wouldn't be part of our game plan. The ability to integrate modules into an engineered system that optimizes all-in performance is something a module manufacturer like us is uniquely capable of doing. And to be able to wrap that with a data set and a monitoring capability and provide assurance to a utility that the manufacturer stands behind the entire result, that's pretty significant. And I don't see us ever being able to do that with some third-party product."

Nor is FSLR eager to go beyond constructing solar PV projects to actually running them, even though owning solar projects is proving to be where the money is (and not in selling modules). "We're not currently thinking that we would actually own and operate the assets," said Ahearn, but "we might be interested in doing [that], if not directly, with a partner on a given market."

 

Update 3/19: Thanks to Tilly210 for the reminder... here's what I learned from FSLR's CFO Fred Meyer, who contacted us (and other media and analysts) to help untangle these numbers which he claims were "mooshed together" unfairly:

  • First, the manufacturing excursion, which occurred from June 2008-July 2009. (He never revealed exactly what the glitch was or how it was fixed...but about 4-8 percent of production was determined to have a sudden power drop in the field because of it.) FSLR says it has processed 95+ percent total claims submitted under the life of its remediation program; total remediation is now ~$215 million (with $146M "above and beyond our standard warranty"), with the remainder of claims as yet undetermined if remediation will be required, which could add another $44M. About 40 percent of claims were processed during 4Q11 so that's why the numbers were so eye-popping now. (He also noted it's tough/impractical to identify and remove individual modules if you're not also the site owner/operator, so many more modules were removed in aggregate, or in some smaller sites replaced entirely, than were strictly necessary.)

    Note that there might be a new wrinkle to this issue: a two law firms are now looking into, among other things, whether not previously disclosing the impact of the "manufacturing excursion" violated federal securities laws.

  • Second is the increase in its warranty accrual, hedging risk against increased warranty returns from hotter climates where it is going after more growth (southwest US, Asia, Africa, etc.). Return rates from these places is below FSLR's prior rate but somewhat higher than from temperate climates, Meyer said, so the company is raising its accrual rate proportionally. That comes out to an extra $37.8M to handle additional claims that might occur. FSLR also is setting aside an additional percentage point (100 basis points) of sales for warranty accrual.
  • We asked for, but didn't get, Kumar's NREL report; we'll ask again. NREL has various reports about this topic, which is well known if not fully understood — an NREL rep pointed us to several papers from the most recent EU PVSEC about this very topic, from silicon to thin-film and even concentrated solar (CSP). He also acknowledged that thin-film is known to degrade faster than silicon, and in such climates maybe it's a little faster than thought -- instead of a fraction up to a full percent degradation per year, maybe it'll be maybe 1-2 percent.
  • Everyone seems to admit that it's tough to determine how modules ultimately perform in the field, until you actually have the field data — even for silicon it's not known how all the materials behave under all light and humidity conditions, the NREL rep said. FSLR's been out for years collecting lots of data on its own panels across many sites and regions... would there be value in providing some of those insights to others (say, for a fee)? "That's an insightful thought," Meyer said cryptically.