How far can existing PV capacity go, in terms of keeping up with end-market growth?
This is one of the key issues in terms of capital expenditures and tool suppliers waiting for an uptick in bookings. It is also essential to understand in terms of forecasting end-market demand in 2015, at a time when trade uncertainty and the preference for project financing is keeping new ‘greenfield’ fab build at a minimum.
Effective capacity is something we have been discussing for some time in the PV industry, as this is vital to working out how much good capacity is around. The term never existed five years ago in solar, but was used in most other technology sectors in a related way.
Effective capacity is different from nameplate capacity, and across the various PV technologies and manufacturers. It also depends on the quality of the substrates going in, and this has barely been discussed in the industry. And it has been the biggest factor in effective capacity re-assessments over the past couple of years, in particular for the p-type multi cell/module lines.
The debottlenecking bug
However, another factor has started to come into the equation: debottlenecking. Of course, anyone running a fab debottlenecks equipment. But at a time when new greenfield fab building is so unattractive, the attention to debottlenecking is much more focused and way more productive.
To understand these issues, let’s quickly look at rated MW-peak module powers (at STC conditions) for the most popular solar panel today: the 60-cell, 6-inch p-type multi panel. Just a couple of years ago, panel ratings of 235-240W were commonplace. Today, this figure has grown to 255-260W. So, this is an 8% increase in rated panel capacity. Or put another way, the cell line that was rated at 30MW then is now effectively a 32-33MW line.
It might not sound huge, but if we have 40GW of p-type cell capacity floating around globally, this is now almost 44GW of capacity. At a time when every last MW of supply is being scrutinised, this becomes a really big deal.
But when we add in debottlenecking, the equation gets more compelling. For a p-type multi cell producer, if process flow changes can be avoided and your 30MW line of a couple of years ago is now 10% higher in capability, it’s a win-win in many regards. Add in a dual-print upgrade on the front and a PERC upgrade on the back, and 30MW is now 36MW. Again, without having to make any huge floor space changes.
How much debottlenecking is being done? And how much can be eked out?
Right now, this is far from an exact science for the p-type c-Si lines, but what is clear is that many of the tools installed and ramped up during the quick capacity ramp-up phase of 2010-2012 were rarely matched in terms of capability. Therefore, in a 30MW line, some tools may have been the gating factors in the line not actually being capable of the nameplate line spec, while some other tools may have been capable of 40MW throughput from the start.
The sleeping capacity question
Another misconception that keeps the focus on debottlenecking relates to replacement (not upgrade) tooling. With much of the p-type c-Si capacity in China and Taiwan, there is clearly a much greater emphasis on keeping existing equipment going, rather than replacing tools after 3 or 5 years with the latest revision.
While each of the above issues is having a large impact on effective capacity levels edging upwards, there is potentially one factor that is even more important to consider: tier 2 and tier 3 distressed capacity reutilisation by tier 1 c-Si cell producers.
Many people used to think that when uncompetitive c-Si cell makers left the industry (for whatever reason), then so did their equipment. It was not too crazy a hypothesis, and possibly one of the reasons for the lack of market penetration was not having the very latest production equipment? But this appears rather a naïve assumption now. There are a few reasons.
Most (but not all) tier-two and tier-three c-Si cell manufacturers in China ordered the same equipment as the leading tier-one cell makers, at about the same time. There has not been a technology buy cycle in the PV industry yet. Therefore, this equipment is largely fit-for-purpose if returned to the food chain for tier 1 module supply.
Furthermore, lower tier manufacturers often ran equipment at 40% or 50% nameplate utilisation, making their contribution to effective capacity of the industry as a whole much lower than had the same equipment been located with Yingli Green or Trina Solar, for example.
Now, if any of these tier-two or tier-three companies’ capacity gets acquired (or the companies/employees get folded within a tier-one company’s operating assets), then the industry’s effective capacity sees an immediate increase.
First Solar adding to the mix
It is not simply p-type multi c-Si improvements that are nudging the effective capacity levels higher. SunPower has its own unique route to adding to the capacity numbers of existing lines (but is still the only leading c-Si maker with real plans for new factories as the main route to grow shipments). And then there is First Solar.
First Solar has to do everything on its own, and now has an R&D department that any Fraunhofer or UNSW professor would be proud of. Getting CdTe (production) efficiencies to where they are today would simply have not happened, had First Solar not had the level of in-house R&D expertise it currently possesses. It is another reason why there is almost no chance any other CdTe manufacturer will end up challenging First Solar for many years to come.
But the interesting thing about First Solar, in common with the leading c-Si manufacturers, is the ‘no-new-fab’ tactic/strategy. In fact, First Solar barely has to build a new floor or wall anywhere in the world in the next three to five years, if all goes to plan from an efficiency and throughput standpoint. Indeed, in three to five years’ time, it is likely every single piece of manufacturing equipment on First Solar’s lines will be different from that seen today in its fabs.
There is more than just restating line capacity however. First Solar also has the scope to bring mothballed lines back, and this surely has to be the backup plan, or the quickest and lowest risk route to quickly growing projects business, depending on where each of the pipeline projects may come online.
Nobody should be that surprised if, two years down the line, CdTe panels are once again coming out with a ‘Made-in-Germany’ label attached. Manufacturing equates to domestic jobs, and eastern Germany has been ravaged by solar industry bankruptcies in the past few years. Regional packages will always be on the table.
And, more importantly, if First Solar wants to play in the post-FiT world of utility-scale growth in Europe and be a preferred bidder in energy auctions, then it is obvious that having local manufacturing creates a massive incentive to government selection. And lastly, equipment has not been sold off from the previous fab adjustments done by First Solar.
Europe back in the picture?
Everything goes in cycles, and c-Si manufacturing in Europe may well see a mini revival in the next three to five years also. This is not because of any perceived manufacturing excellence the likelihood of any grand gigawatt fab plan, but due to several other less glamorous and more pragmatic drivers.
Trade wars (or the expectation of more countries and variants) make European manufacturing one of the safer bets, compared to Southeast Asia or some of the emerging regions that are not short of risk factors. Europe gets you into European solar (and free from the EU mandates) and also gets you a green card to the US market. So long as you are making cells and modules, for now.
Europe still has a fair amount of cell capacity, despite all the auctions. And within southern Europe (where much of the government investments were done in the mid-2000s to address domestic markets), the economic climate is still languishing in post bank-crisis debt and unemployment. There is not a single region or authority that would not welcome Chinese investors, if it meant mothballed fabs could be brought back to life with job recreation.
The numbers are not in the gigawatts, but they would add to the mix, and would provide another small part of the overall effective capacity growth. The political value however is the big pull here.
During the forthcoming investor calls and releases from the public-listed PV manufacturers, don’t expect to hear much about greenfield fab capex. In fact, don’t expect to hear much about capex, other than in relation to bridge or construction financing related to downstream project activity.
Maybe, again, only SunPower will be talking about greenfield capex? We will find out soon.
For now, if you want to read about c-Si capex, you can hear regular updates from SolarCity or just spend a few days in upstate New York and you will not be long to hear about some plans for creating a US-based PV manufacturing hub (again).
But for the rest of the PV world, the small and very real upward adjustments of efficiency, throughput, debottlenecking and acquiring distressed capacity assets are flying below the radar, and don’t garner the headline press of a new GW-fab with advanced cell processing. But in terms of their immediate impact on the solar PV industry and keeping effective capacity moving forward quarter after quarter, they are (along with cost reduction) some of the most critical issues to manufacturing stability and cautiously restoring investor confidence in guiding future profits.