Industry Observer

  • How A Little-used NREL Tool Could Have Saved the US Solar Industry $6 Billion
  • 2014-05-28 17:01:45
  • Solar industry leaders gathered this past week at the Dept. of Energy’s SunShot Grand Challenge Summit to take stock of progress in achieving the challenge’s main goal – solar at less than $0.40 per watt by 2020.

    The industry driving force, the almighty $/watt, has only half its outcome based on efficiency (the watt). The other half is all about the manufacturing cost (the dollar).

    As an industry, we have spent far too much time focused on efficiency, and need to start focusing more on the “$”.

    Efficiency reports are the most common update in solar, whereas manufacturing costs are almost never mentioned. Investors have funded almost every organization that has achieved a world record efficiency.

    But if one looks at the data, almost no solar company has tanked for failing to hit its efficiency target. On the other hand, they have almost all failed because they did not hit their manufacturing cost targets. If you get the data from Solyndra, NanoSolar, SoloPower, I think you will concur that the problem is the “$” not the watt.  So why doesn’t manufacturing cost get more attention in the United States? Maybe it’s a metric problem.

    Efficiency is easy to measure and verify. The DOE’s National Renewable Energy Laboratory (NREL) is a highly respected third party certification body whose efficiency measures no one questions.  But on costs – how do you verify that?  It is tough. During my time at a leading renewable energy venture capital firm, I looked at over 200 solar deals, and it was very difficult to assess manufacturing costs. When I did, costs were not what entrepreneurs said they were. In fact, you would not believe how far they were off. For instance, one company claimed they could hit $0.50/watt in four years. Four years and $150 million later, they arrived at a cost of closer to $7-8/watt.  How is this possible? Why didn’t the investors do due diligence? Good questions. But without a standard measurement method, and without a third-party verification, lots of crazy things can happen, and did, and do.

    As I have written previously, the U.S. government a few years back focused on thin film PV, specifically CIGS. In Q4 last year, CIGS surpassed mc-Si on efficiency in the R&D labs. If this occurs on the manufacturing floor, thin film may make an incredible comeback. Subsequent to that article, First Solar improved CdTe efficiency in R&D to match that of mc-Si and forecast it will pass them on the manufacturing floor. More evidence the government got it right on thin film. But they could have done more.

    One of NREL’s “diamonds in the rough” is a very rigorous cost model for solar panels. Used properly this model could have predicted the cost outcomes of most of the solar manufacturing start-ups before they happened. By my estimate, $6 billion could have been saved by avoiding the struggles of many U.S. solar manufacturers such as Solyndra, Nanosolar, etc.

    And I am not just speaking from the outside looking in. I have used almost exactly the same cost model to create two companies that both reached IPO. At Novellus, the cost model was used in the R&D stage to vet technology. I did not allow the engineers to cut metal until the model predicted we would have twice the throughput of Applied Material, cutting our customers cost in half. The result: an IPO and a $3 billion company value. Building on that success the same model was used at Mattson to develop the Aspen platform and revolutionize the photoresist stripping business. Result, another IPO.  Cost modeling works.

    And NREL has already done the heavy lifting for solar. Its model is great. Al Goodrich and Mike Woodhouse deserve credit for their dogged effort in creating the NREL model, and Director Dan Arvizu for funding the effort. Now we need to use it.

    We need to evolve NREL’s model into a commercially available source of third party verification on manufacturing costs. It was done in semiconductors. The model pioneered at Novellus was adopted about five years later by almost every company in semiconductors, and independent companies grew up around offering these services. Not only could billions be saved, but the more important result would be that capital could be diverted to technologies that can actually be commercially successful. The solar industry needs that right now as current manufacturers are having trouble with profitability. Who is going to fund the next doubling of capacity we will need in the next 4-5 years? Low cost manufacturing is needed to avoid an era of profitless prosperity in solar manufacturing. The NREL model can help us get there.

    Leave a message