SunPower Corporation (SPWR 2.21%) is one of the oldest companies in the solar industry, meaning it's survived a number of ups and downs to get to this point. But in 2016 the company faced a crossroads thanks to rapidly falling prices for solar panels. Most importantly, its mid-level efficiency E-Series solar panels, which accounted for most of its manufacturing capacity, were no longer competitive, particularly in the highly competitive utility scale market. And there were no signs the competitive landscape would improve. 

What management decided to do was shut down Fab 2, the company's oldest operating manufacturing plant, and buy a company that had technology that would allow it to produce a new low cost product using commodity solar cells for use in the utility market. It's a risky strategy, with less differentiation than SunPower's past products. Here's how management is hoping the strategy plays out. 

Image source: SunPower.

Why investing in solar manufacturing is so risky

To understand the pickle SunPower is in, we have to look at why solar manufacturing has been a terrible business for over a decade now. The core problem is that a company will spend millions, or sometimes billions, of dollars to build out capacity, only to see the prices of the solar products it's producing fall every year the plant is open. Products may be competitive the day a factory is opened and margins may be strong, but by year 4 or 5 the plant is usually not competitive with newer technology and the plant is losing money. 

This risk can be seen in SunPower's own manufacturing plants. Fab 1 opened in 2004, only to be closed in 2012 because it wasn't competitive with newer technology. Fab 2 opened in 2007, and its closure was announced late last year. The trend here is that a manufacturing plant has less than 10 years of useful life.

The financial impact of the closures can be significant. Management said the recently completed Fab 4 would cost $185 million to $230 million to build and may not have long to make a financial return on that investment. To put this investment into perspective, the 350 MW capacity means the capital expenditures for this latest investment were about $0.65 per watt. That's a big investment in a world where commodity solar panels are selling for $0.40 per watt and gross margins are under 20%. Making a solid return on investment is far from certain with those prices.

How SunPower plans to lower its own risks

The big risk I'm highlighting is the capital invested in manufacturing capacity, a big part of the cost structure for any manufacturer. Any company that can lower the upfront capital risk, or offload it to someone else, will lower its risk overall. 

Image source: SunPower.

Given the market dynamics, SunPower decided to buy Cogenra in 2016, a company that had developed technology that stacked cells like shingles into a module. This created a low cost product that could use commodity cells as an input, but made a module that was slightly more efficient than traditional module construction. 

SunPower planned to use the new product, dubbed P-Series, as a module for its Oasis utility scale solar product. E-Series was no longer competitive in Oasis, so this will allow SunPower to stay in the utility market with a slightly differentiated but still cost competitive product. 

What could make this strategy worthwhile is SunPower's capital costs. Management said capital costs for P-Series production are just $0.05 cents per watt, and capacity can be built in just 6 months. That's a fraction of previous generation products, and will allow SunPower to bid in utility auctions around the world and then plan out manufacturing capacity to meet its own demand years in advance.

To look at this another way, SunPower may be able to lock in the margin necessary to pay for its own capital equipment before the equipment is even built. We can see this in the auction the company won in Mexico last year. The power plant won't be constructed until 2018, giving plenty of time to build out P-Series capacity before modules need to start rolling off the assembly line. 

With P-Series, SunPower is hoping that instead of spending hundreds of millions of dollars to build plants that will be obsolete in 8-10 years, it will build plants that have a payoff already contracted when they're built. The risk it used to take building out capacity would be reduced dramatically. 

How the utility solar strategy could go wrong

SunPower's entire strategy to move into P-Series is built on the idea that solar cells will be readily available and cheap to buy on the open market. That's proven to be a good bet in the past, but if solar demand grows in the next few years and demand exceeds supply we could see suppliers just self-consume cells in their own modules. Building out the module component of manufacturing is relatively cheap, and if it adds a little value it could be worthwhile for cell producers to move into the module business. 

I think the risk that cells won't be readily available is low, but it's a risk worth noting. 

How to tell if SunPower's strategy is paying off

What investors will want to watch in 2017 is whether or not SunPower is winning solar energy auctions around the world, or partnering with developers that will build them. The Oasis platform allows SunPower to evaluate sites quickly, and the power plant design should lower costs to a level that makes them competitive with the commodity market. That should translate into winning contracts to build power plants in 2018 and beyond. 

If SunPower isn't able to win contracts, it will be a sign that competitors are either bidding at a level they can't make money at or that SunPower's Oasis/P-Series product isn't as competitive as planned. 

No matter how it plays out, SunPower is taking a much lower risk with P-Series than it did in building Fabs 1-4. Maybe that's a sign the company is tired of fighting against the commodity suppliers that have caused pricing pressure for more than a decade now. And having commodity cell prices as a tailwind rather than a headwind may be a good place to be in 2017.