According to CLSA, sources Phonton.
69GW supply vs. 25GW demand
Solar cell makers claim to have shipped 37 GW of product in 2011 vs. our demand estimate of 24 GW, plan to raise capacity to 69 GW in 2013 with expected production of 53 GW vs. our demand estimate of 25 GW. These kind of numbers suggeest the price along the chain will be able to remain irrational much longer than most companies can remain solvent.
SOL is in much bleaker picture. I estimate 18c loss in Q4.
Poly: My guess is that $27 kg production cost at Q4 through 4000MT capacity of $30/cost and 6000MT new capacity of $25/cost vs. ASP $24. loss of $3 per kg on GAAP basis. But in cash flow basis, it is positive. Also, after ramping up, phase 2 could be better than $25 toward low $20.
Wafer: 30c as guided from 15c processing cost. Given spot poly $24, there is 2c profit at gross level. But eventually this 2c profit might be gone given over-capacity situation at wafer and GCL ramping up quasi-mono. (320MW output in Q4)
Module: 120MW with tolled cell, 67c processing including cell-tolled. 60MW fully-integrated with 66c processing. ASP 76c.
SG&A: The company attempts to clarify it as overall opex quarterly at $25 millions. I guess it $6.5 million selling exp., $10 mil G&A, $8.5 millions R&D.
6% gross margin at wafer sales, 12% at module sales. SOL needs to ramp up module sales more than 180MW level.
Three good houses in bad neighborhood. I am not familiar with GTAT. But roughly in a order of preference, I would say CSIQ, GTAT, then SOL. I posted quite bearish view from Tony Hung. My view isn't as bearish as he is. But he is the guy in the fronteer. So, you better discount what I am going to say following for your decision since admittedly, very likely I am as confused as you. To be perfectly candid, I am reluctant to share my view, since there are always some uncertain factors might force me to eat my own foot.
Assuming poly ASP hits $24 by Q4, GCL will have to make their decision on how much they want to fill 65K MT poly capacity and 8GW wafer capacity with the condition not hurting their ASP too badly. No matter they could have poly cost of $16 or $18, they still enjoy hefty margin on poly sales. They will have 0.5% incremental margin or 4.3% incremental margin by taking extra step to make wafer to sell it. Sounds not too enticing to pursue wafer-making, but it is an issue of strategic deployment. First, it doesn't hurt them to get extra margin from wafer. Second, that leaves a pressure to their competitors namley LDK and SOL.
CSIQ will ship 1800-2000MW this year probably 80% filled by GCL. This is a sizable customer in comparison with GCL 8MW wafer capacity. CSIQ is not only significant but also a worry-free customer for GCL. Since CSIQ already achieved a tier 1.5 status. So, GCL doesn't need to provide extra backing for their warranty. This stands a vast advantage in landing better deal than others to get a good wafer ASP from GCL. Of course, JASO and TSL are in similar positions. In conclusion, CSIQ is highly likely able to get 25.5c wafer from GCL.
From this 25.5c, with a (cell+module) processing cost of 38c (discounted from their goal of 35c), 11% opex, CSIQ could make 11c in Q4 if ASP is at 78c.
This does not count TRP projects to be harvested in Q4 and earlier 2013. In the past, CSIQ stated a system ASP of high-$2 toward low-$4 could bring them extra 10% margin above module-making. TRP projects has 84MW pegging at system ASP near $5 (I am writing from memory, please check the numbers). Even you assume that Canada module fab needing extra 20-25c to make module, there is windfall profit left to enrich Q4 EPS.
1) 11c is too thin to defend the unforeseeable factors. CSIQ historically disappointed investors for never-able-to-deplete inventory.
2) CSIQ's strength is in Japan, weakness in China. They didnt have China market in the past.
3) Reading the stock movement, investors seem not to give much credit on projects' profit for its one-time nature. FSLR supposes to have $30 value for only considering the discount cash value of their projects pipeline. But the stock is traded at near $20.
4) On a relative competitiveness stance, I have given CSIQ a prefered status by endorsing their cell+module processing to 38c while others' YE goal implies a maybe around 42c (C+M) processing. If you digest from tony's statement, others very likly may run neck-in-neck with CSIQ on that processing cost efficiency. The reasons caused me to put CSIQ in a more favorable procesing cost may be just the difference between Shawn's more aggressive goal and other's less aggressive stance.
You mean fill them in with the same as in 2011?
Yes that is the plan, if I fail to get the data.
Another semantically based error. 80% of names listed not 80% of capacity. That makes more sense, right?
so there are around 20% companies which are actively putting news and data on their sites,
those you listed will be in the group.
For pratical reasons, I would suggest filling 2012 cell capacities blank for most of taiwan makers, LDK, Hareon.
I couldn't find fault with major capacities you listed. CETC, Prarie, BYD (http://bydit.com/doce/products/Solar%20Energy/). DMEGC only got UL certificate 6-months ago.
Few explanations, the very first link is to the SPVI capacity, this is a very first table I ever created and it has been updated since last year. The fact that those companies do not have update for 2012 (the cell) is that they are not providing one. I have made reasonable attempts to contact every one of them on the list. What you see updated is a result of the e-mail, phone contact or public announcement. This link is a dynamic table if I come across data, I update, I always review financials looking for it. This is why predominately that list is for public companies.
The sum of link 1 (which is I consider in my writing a tier 1 and 2) and 2(published yesterday) is the true data I was able to capture from the net. The risk with the group identified yesterday, is whether the statement is real. Some of the news release about the company date to 2007. One more thing which for me is a relief, 80% on the second link are enterprises dating to 2006 to 2007 area, with equipment which is rather outdated. Not seeing any recent updates on the website , may mean they do not update the website only, but that usually spells trouble for company's state of immediate business. The other 20% are new enterprise (after 2008) but they lack detail and all of them sport artist's rendering of their factory, which gives me idea that they are not what they proclaim to be. When I started this exercise there were well over 300 businesses after I took participants form the first list off. What you see is a what is left, namely because I did not find any live links or second source of support (beside ENF) to the existence of the company. This is why second pass 167 became what it is now as well. ENF had production data, which was not linked to a live site or did not have confirmation elsewhere. Some of the live links led to non existing domains etc. One was even a malware site.
Whenever I found someone listed, I added them to on of the indices I have on the site just to see if I can get more data from my reviews of financial documents.
Just for my sake of the custodian of records I would like to see a quick consolidation :)
If I were stp, yge, tsl I would guide for 70-75 US$ cents as well and scare everyone away. then I come out with a surprise in q3, q4 and say we sell at 85 US$ cents. stockholders are then happy and competition was succesfully scared away...