Seeking Alpha

EysteinH's  Instablog

EysteinH
Send Message
Norwegian. Master degree in psychology. Currently studying law and working part time. Interested in stock Analyst work.
View EysteinH's Instablogs on:
  • Solar Cost Comparison: Production Cost And Added The Operational Expense Cost (Opex)

    For this article we will look at: Renewable Energy Corporation (REC) a Norwegian solar stock. Yingli Solar (NYSE:YGE) , Trina Solar (NYSE:TSL), China Sunergy (NASDAQ:CSUN), and Hanwha SolarOne (NASDAQ:HSOL).

    Debt of REC from Q42011 to Q12012:
    Interest bearing debt (short+long) q4 2011: 1032 mil $ (6300 mil NOK) Q12012: 999 mil $
    Cash and cash equivalents q4: 261 mil $ q1: 308 mil $
    Less debt, more cash.

    Debt of Yingli from Q42011 to Q12012:
    Interest bearing debt (short+long) Q42011:1855,3 mil $ Q12012: 1936,5
    Cash and Restricted Cash Q42011: 891.9 Q12012: 674.7
    Increased debt, less cash.

    Debt of Trina from Q42011 to Q12012:
    Interest bearing debt (short+long) Q42011: 909.6 mil $ Q12012: $1,020.4 mil $
    Cash and cash equivalents and restricted cash Q42011:896.4 mil $ Q12012: $748.3 mil $$
    Increased debt, reduced cash.

    Debt of CSUN from q42011 to 2012:
    Cash and cash equivalents Q42011: 209.5 mil $ Q12012:233.2
    Interest bearing debt (short+long) Q42011: 475,1 mil $ Q12012: 460,3 mil $

    Increased cash, reduced debt slightly.

    Debt of Hanwha Solarone from Q42011 to 2012:
    Cash and cash equivalents Q42011: 314 mil $ Q12012: 303.1 mil $
    Interest bearing debt (short+long) Q42011: 613 mil $ Q12012:534,7 mil $

    Slightly less cash, less debt.

    Time to take a look at another cost of modules, the operational expense (opex):

    The theory behind this value metric is to spread the operational costs thare are not one time across the modules shipped for that quarter. That gives an idea on how effective sale channels, adminstration etc is and how much added cost these type of operational expenses add to the modules per watt.

    First a summary of the findings:

    CompanyProduction costopex costs in mil $Average sale prices (ASPshipments q1
    Rec1,022,920,98Shipments 191MW
    Yingli0,8760,30,96*Shipments 520MW*
    Trina0,8658,10,92Shipments 380MW
    China Sunergy0,85*18,80,86Shipments 79,9MW
    Hanwha Solarone0,92*23,170,83Shipments 160,7MW

    *shipments calculated shipments based on revenue/ASP
    *ASP Yingli calculated on gross margin adjusted for one time effects. (11,5 % adjusted gross margin)
    *China Sunergy cost: ASP*1-gross profit margin)
    *Hanwha solarone production cost if only using internal wafers 0.78

    We can then compare producers at production cost and then also add opex cost to compare them:

    CompanyProduction + opexProduction cost
    Rec1,121
    Yingli0,990,87
    Trina1,010,86
    China sunergy1,080,85
    Hanwa solarone1,060,92

    Note:
    Opex is adjusted for one time effect where possible.

    Sources:
    Yingli Solar q1:
    media.corporate-ir.net/media_files/IROL/...
    ir.yinglisolar.com/phoenix.zhtml?c=21301...=

    Trina Solar q1 phx.corporate-ir.net/External.File?item=...

    ir.trinasolar.com/phoenix.zhtml?c=206405...=
    http://ir.trinasolar.com/phoenix.zhtml?c=206405&p=irol-newsArticle&ID=1664310&highlight

    China Sunergy q1 chinasunergy.investorroom.com/file.php/1...

    REC q1 www.recgroup.com/PageFiles/156/Investor_...

    Hanwa Solarone files.shareholder.com/downloads/SOLF/189...

    http://investors.hanwha-solarone.com/releasedetail.cfm?ReleaseID=678001

    http://investors.hanwha-solarone.com/releasedetail.cfm?ReleaseID=657438

    Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

    Additional disclosure: I am long in REC in the Norwegian stock exchange.

    Jun 05 8:14 PM | Link | 1 Comment
  • An Historical Prognosis Of Rec Cash Flow Q2 2012 Until Q1 2013

    First the results of the historical prognosis:

    - REC SOLAR -

    World average prices per quarter:

    q12012: 0,90 $ per watt. - OK - (adjustet for current euro/dollar values)

    q22012: 0,81 $ per watt. / -10% /

    q32012: 0,77 $ per watt. / - 5% /

    q42012: 0,73 $ per watt. / - 5% /

    q12013: 0,69 $ per watt / -5 % /

    Rec ASP per quarter:

    q12012: 1 $ per watt. - OK - (adjustet for current euro/dollar values)

    q22012: 0,89 $ per watt. / -10% /

    q32012: 0,85 $ per watt. / - 5% /

    q42012: 0,80 $ per watt. / - 5% /

    q12013: 0,76 $ per watt / - 5% /

    Rec production cost per quarter:

    (not included deprecation)

    q12012: 0,97 $ per watt. - OK - (adjustet for current euro/dollar values)

    q22012: 0,87 $ per watt. / -10% /

    q32012: 0,78 $ per watt. / -10% /

    q42012: 0,72 $ per watt. / -5 % /

    q12013: 0,68 $ per watt / - 5 % /

    Rec cashflow margin per quarter:

    q12012: 3% / 3 cents per watt /

    q22012: 2,2% / 2 cents per watt /

    q32012: 8,2% / 7 cents per watt /

    q42012: 10% / 8 cents per watt /

    q12013: 10,5%/ 8 cents per watt /

    Rec cashflow at 160 MW sold:

    q22012: 3,2 mil $ / 2 cents per watt /

    q32012: 11,2 mil $ / 7 cents per watt /

    q42012: 12,8 mil $ / 8 cents per watt /

    q12013: 12,8 mil $ / 8 cents per watt /

    - REC SILICON FBR A-

    Rec ASP per quarter:

    Q12012: 30/24? $ per KG

    q22012: 24$ per KG / 0% /

    q32012: 23$ per KG / - 4% /

    q42012: 22$ per KG / - 4% /

    q12012: 23$ per kg / + 4% /

    Rec production cost per quarter:

    (not included deprecation)

    q12012: 15,5 $ PER KG. - OK -

    q22012: 15,0 $ PER KG / -3,2% /

    q32012: 14,5 $ PER KG / -3,3% /

    q42012: 14,0 $ PER KG / -3,4% /

    Q12013: 13,5$ PER KG / -3,5% /

    Rec cashflow margin per quarter:

    q12012: 35,4 / 8,5$per kg /

    q22012: 37,5% / 9 $ per kg /

    q32012: 36,9% / 8,5$per kg /

    q42012: 36,3% / 8 $ per kg /

    q12013: 41,3% / 9,5 $ per kg /

    Rec cashflow t 3650000 KG sold

    q22012: 32,9 mil $ / 9 $ per kg /

    q32012: 31,0 mil $ / 8,5$per kg/

    q42012: 29,2 mil $ / 8 $ per kg /

    q12013: 34,7 mil $ / 9,5$perkg /

    Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

    Additional disclosure: I am long in REC on the Norwegian stock exchange.

    Jun 03 7:01 AM | Link | Comment!
  • A closer look at Renewable Energy Corporation wafer division

    In my first "a closer look at" I looked at the silicon division of REC and found that the current value of REC was less than what I believed the silicon segment to be worth. Link: seekingalpha.com/instablog/901705-eystei...

    This time we will look at the wafer segment at REC. This segment is the story of falling EBITDA margins due to investing during high cost years of 2007/2008, a market confrontation over margins and a very rapid fall in wafer prices more than REC anticipated in the 2010 roadmap they had starting 2006. But it is also the story of hidden underwritten value in book values, a hope of relief in the price war, technology investment that will pay dividends later and achievement beyond the 2010 roadmap.

    This article assumes you know the basics of solar modules making. If you don't feel you have a good grasp on this I made a basic introduction to it for amateur investor from an amateur investor: seekingalpha.com/instablog/901705-eystei...


    Current book value of wafer segment: "Too much write downs with hidden value?"

    The current value of wafer in Norway is 1.05 billion NOK.  Q2 saw write downs of 4.7 billion NOK in wafer Norway. The reason was the 45% drop in wafer prices from may to June 2011.  Here is a lot of hidden value should the current wafer prices improve or the costs in production be reduced. And in 2009 a 991 million write down happened with the mono wafer plant in Glomfjord due to failure to ramp up in time. The investment on the mono plant was 1.3 billion NOK in 2007 and 500 million NOK in 2008. In Herøya 3&4 the investments in 2008 was 2.5 billion NOK. The capacity of Glomfjord mono is 300MW but currently estimated 2011 total will be 146 mw under half the capacity. Again if the ramp up should gain speed there is a potential for hidden value to become clear. There is a current trend happening on this now as I will show later in the article. In Singapore 1.4 billion NOK was written down this quarter and this was from the most modern equipment, probably in the wafer segment as this saw heavy underwriting this quarter. Again here is a hidden value should the operation be improved and wafer segment coming online again. In 2007 REC invested in a R&D center worth 210 million in Herøya focused exclusively on the wafer segment. In 2007 Herøya 3 and 4 was invested at 2.5 billion NOK (Herøya 3 built in 2008 and Herøya 4 built in 2009and together with Singapore form the most modern wafer segment at REC.

    Since the current price of REC is less than the silicon segment alone as I argued earlier then the wafer segment is basically "free" and anything positive that happens should be a good catalyst for market price increases.

    "Story of falling EBITDA margins"


    Graph 1: "Ebitda trend downward" -% EBITDA by year.

     

    Year Ebitda wafer wafer production multi mono % mono
    2006 34 317 286 41 14,34 %
    2007 42 507 469 38 8,10 %
    2008 34 582 541 41 7,58 %
    2009 18 799 758 41 5,41 %
    2010 12 1210 1121 89 7,94 %
    2011 15,7 1123 977 146 14,94 %

    Table 1: "Ebitda 2006-2011 (estimate) "
    Currently the trend has been since 2008 of a rapid decline in wafer prices as EBITDA margins show clearly. From May to June 2011 wafer prices dropped 45% while polysilicon spot prices dropped 32%. Since REC have stable silicon prices of 50$ per kg this hit the wafer segment hard as most of the decline in price was taken in this segment. I believe that currently the trend is for a price war happening where some few will remain victorious and the rest will have so low margins as to quit the business.

    REC took a drastic step in q2 shutting production in its oldest plants Herøya 1&2 and Glomfjord multi starting q3:

    "REC will temporarily shut down production at the oldest wafer plants at Herøya in the third quarter due to the weak market conditions. In addition, a decision has been made to temporarily shut-down production at the multi wafer plant in Glomfjord" (q2 report)

     

    This means Herøya I & II and Glomfjord multi wafers are closed with around 185mw less production. For 2011 I have estimated a growing EBITDA at 34,6q3 because of closing the oldest plants and inventory sales. Or 10,76% EBITDA excluding inventory sales. Ironically this estimate is actually conservative because there could be increased sale prices due to higher efficiency wafers while I have calculated with 9% further price drop and increasing demand could make prices slightly higher. I have also noticed that they have phased out the older 225W modules at a local retailer list www.ecobusinesslinks.com/solar_panels.htm and now only sell 240W panels at a higher cost than earlier. (1.75$ per watt compared with earlier 1.55$ per watt.) so this could mean prices are on average higher now for REC.


    One question that pops out is have we witness a bottom in the REC EBITDA and will it improve from now? The trend have been for faster and faster reductions but this can only go to a certain point before the trend will slow down. The reasonThere simply is no current technology or easy approach to make costs reduce at the current pace. A recent trend in the Chinese companies is to focus on efficiency gains as the price drop based on cost reduction have slowed down. (source: www.electroiq.com/articles/pvw/2011/08/b...) I believe the current trend will benefit REC wafer segment as they have plenty of room to improve the wafer costs due to failure of achieving the original 2010 cost road map for wafers. Much more compared with its Chinese competitors. And REC efficiency is on a steady rise, currently at 16.7% and gaining to 18% in 2012. REC simply choose the strategy of taking the hard path first with increasing efficiency instead of the simpler measures to decrease prices like an in-house wire production. (More on this later in the production cost discussion.)

    "Ebitda 2011 estimates"
     

    quarter 2011 revenue cost ebitda ebitda margins
    q1 sale price @ 5,49 cost @ 4,345 1 846 1560 286 16
    q2 sale price @ 4,47 cost @ 3,56 
    1 190 1226 -36 -3
    q3 sale price @ 4,0677 cost @ 3,63 ( I have assumed 9% fall in price levels) 1189,8 778,64 411,16 34,6
    Q4 sale price @ 4,0677 cost @ 3,45 (I have assumed stable q4 prices) 876,59 743,48 133,11 15,2
    average 1275,5975 1077,03 198,5675 15,7
    Q3 excluded inventory sales @ 4,0677 cost @ 3.63 872,52 778,64 93,88 10,76

    Table 2: "Ebitda estimates for 2011" Please note q3 includes inventory sales 
    Sale price calculated as (income without special items - EBITDA) / sales in mw and costs is incomeproduction mwSo both are average prices for all REC wafer products. Note that sale prices are in NOK. (So for q1 around 1 $ on average.) I have assumed Singapore produced 170MW of modules and sells 15MW external for the wafer segment to increase its revenue. If singapore consumes all wafers (185MW of modules) revenue and costs both would be reduced 7% for Rec Wafer.


    Graph 2 " Ebitda 2011 estimates"
    Total production q3 will be at 214,5MW and inventory at 78 meaning 292,5 MW of wafers have to be sold for these estimates to be valid. In q1 rec sold 336 MW so this is quite achivable.

    During q3 sale prices have fallen because it was only in the middle of q2 sale prices started dropping. Rec stated they dropped around 21% without heding effects and 17% with. Since the fall startet slightly after the middle of the quarter  I have estimated another 9% drop in prices for q3 but even with this fall when REC decided to sacrifice revenueproduction to just use the more modern plants we see a total new EBITDA picture at 411,16 million NOK with inventory sales included but  93,88 million NOK without inventory sales. (Inventory costs have been written down q2 already.) For quarter 4 I have estimated stable sale prices but this scenario I am not sure will happen, due to a higher mono wafer percentage the average price should be higherFurther more without the 220watt modules I suspect was produced at the old Herøya plants the sale price is more likely to increase due to higher watt modules. With my conservative estimate Q4 REC will gain 133,11 million NOK q4I think the potential is positive for q4 but I would rather err on the negative side than on over enthusiastic estimates. Ironically the write downs of wafer segment in q2 2011 was far more than the old factories leaving only 1 billion NOK left out of the value of the new plants invested for 2.5 billion NOK in 2007 and finished built in 2008 and 2009.



    "Falling wafer production for first time"


    Graph 3: "Wafer production stagnating and falling slightly 2011 for the first time due to market pressure(wafer production for REC WAFER)

    I have estimated that q3 and q4 the oldest factories will not produce wafers and total for 2011 will be 1123 mw (977multi 146 mono) for mono i have estimated a growth of 1 mw q3 and q4 (35+36+37+38mw q1-q4.) This will be the first time REC actually slows down wafer production as this had gained speed since 2006. This indicates quite clear that drastic measures was put into action when they temporary shut down the oldest wafer plants. Since these oldest plants was upgraded in 2010 with retrofits to the ingot furnaces the efficiency of these wafers where on pair with the rest of the wafer production (more on this later in the technology segment of the article) but these factories have less automation and have a higher cost per watt than the newer plants, taken together with the fact that the Norwegian government pays worker salaries after 2 weeks for up to 54 weeks  (www.morningstar.com/earnings/earnings-ca...)  this alone means wafer cost should drop q3. Also the fact that most of the wafer segment is written down means deprecation costs for q3 should be much lower also giving a lower production cost


    Graph 4: "Singapore and Herøya 3&4 the major capacity achievers"

    The production capacity at Herøya is around 1.1gw. 650 mw is from the newest Herøya 3/4. (162.5 mw per quarter) leaving 450mw to the older two plants (112.5mw per quarter) taken together with the fact that in q2 the reduction means 185 mw less capacity this leaves Glomfjord multi at 72.5 mw per quarter (290 mw annually) Singapore have a capacity around 740 mw (185 mw per quarter.)  If production halting at the older plants is not temporary then the current yearly production of wafers at REC total (Rec wafer and REC solar)  is 1680 MW or 347.5mw per quarter.


    Graph 5: "More modern capacity than old"

    The fact that the current value of Herøya 3&4 is 1.05 billion NOK compared with the fact this plant contributes 650mw says to me that there is a lot of hidden value in this plant. The plant was built in 2008 and 2009. And saw some investment of 2.5 billion NOK and currently it and the mono plant is worth less than half the investment made 2 and 3 years ago.  Q3 will be very interesting because it will reveal what is the cost of the more modern wafer plants alone not dragged down in result by the older plants. I suspect the reason the 220watt panels are no longer listed for sales in online retailers is because this was produced in the older plants while the new 240watt and above is from the more modern Herøya plants & Singapore. (The new modules have a higher average sale price of 1.75$ or 9,625 NOK per watt at a total cost of 1,51 $ / 8,3 NOK per watt q2.)

    "What if: Wafers are used for internal modules and average selling is 240watt 1.75$"
    As previously mentioned REC now sales 240watt modules for minimum price of 1.75 (9,625 NOK) per watt according to www.ecobusinesslinks.com/solar_panels.htm august 14. From hugin.info/136555/R/1531802/466302.pdf the REC q2 report we know cost at Singapore for the module totaled 8,3 NOK or 1.51$ and out of that cost 4,74 NOK came from Singapore as wafer cost was 3,56. Selling prices all wafers where used for modules they would sell at 4,885 NOK. As Singapore have it own supply of wafer this scenario would mean that REC can increase its wafer values just by adding capacity in modules to Singapore.

     Increasing mono wafers production

    Graph 6: "Mono wafer share increasing of total production"

    From the lows of 2009 at 5% estimated 2011 mono wafer is around 15% of total production. In line with the current trend of mono wafers demand due to Europe demanding high efficiency modules.  "This week, the cell and module market prices remained high at $0.9/Watt due to Europes strong demand for high conversion efficiency solar cell."  (source Energytrend: www.energytrend.com/node/1817)  Spot prices of mono wafers are some 20 cents higher than its multi wafer counterparts currently according to Energytrend and 60 cents higher on average according to Pvinsights. (Source: pvinsights.com) This should help increase margins for REC going into q3. If Glomfjord manages to ramp up quicker to its capacity of 300 mw annually then this could also be a positive surprise.In my estimates I have not considered the effect of more mono wafers sold in increasing average sale prices (i reduced sale price by 9% in q3.) so that could be a positive suprise.

    In the next section I will try to look at production costs for wafers, so first a very small introduction of the wafer making process is in order:
    "A short introduction to the wafer process reveals the other areas of improvement. First silicon is melted in furnaces into big ingots. These ingots are then cut into big clumps that fits in a sawing machine that cut them into thin wafers. During the cutting a slurry is used together with a wire to allow cutting into thin wafers. Finally the wafers are separated from each other, washed and dried."

    "Rec wafer production costs"


    Graph 7 "Raw material costs of wafer production - is there potential?"
    RecQ1 production costs. (Data from investor may pdf.)
    Rec2: estimates modern production and 50$ kg poly price (what we will see q3)
    Rec3: Estimates modern production plants only  & cash cost poly @ 26$ per kg. (What we would see if REC decide to go for the same strategy as GCL with cash cost to focus on wafer sales.)
    Rec 4: potential further cuts wire / slurry / labor discussed as possibilities.


    Company Admin and deprecation  in cents ingot wafer in cents poly cost in cents Total cost in cents NOK cost@ 5.5$ 4watt wafer in $ Ingot size
    Rec 18 33 28 79 4,345 3,16 1200
    Rec2 10 31 25 66 3,63 2,64 1200
    Rec3 10 31 13 54 2,97 2,16 2000
    Rec4 10 24,4 13 48,4 2,662 1,936 2000
    Ldk Included -> 25 28 53 2,915 2,12 800
    Renesola Included -> 26 31 57 3,135 2,28 650
    Gcl poly Included -> 24 14 38 2,09 1,52 1000
    Typical cost Included -> 28 28 56 3,08 2,24 650
    based on solarpvinvestor.com/asp-cost-data.html            

    Tabel 3: "Summary of production costs for REC and major wafer competitors"
    * Values in $ cents but 4watt wafer value in $ dollar. For NOK currency used is 5.5$ = 1 NOK. Ingot size in kg.

    1. Polysilicon discussion.

    From March 2011 to July wafer spot prices dropped 45% according to q2 report of REC. At the same time polysilicon spot prices dropped 32% but is still above REC contract levels around 50$ per kg. At the same time REC polysilicon cost is around 27$ per kg expected to drop to around 25$ per kg in q3 or q4 for REC. I think its only a matter of time before polysilicon drops to new levels of 45$ per kg. With 5.6grams per watt this will reduce costs from 0.28$ to 0.252$ per watt. With only newest Herøya plants & Singapore and estimate of 5 grams per watt this will be 22.5 cents per watt. If REC decided to go for cash cost production they could go as low as 13 cents per wafer in costs beating even Gcl-poly and having worlds best cash polysilicon cost for wafers. (This is scenario REC 3 in the graphs.) If i am not correct in my estimate of 5grams per watt then prices will be off by 3 cents per watt. I do belive this is not a over optimistic estimate because the more modern plants have a higher average effiency and since the wafers produce more watts they also on average need less silicon per watt.


    Graph 8 "size of ingots due to furnace size"
    Ingot size from Renesola public announcement, Ldk investor pdf.

    Gcl-poly ingot size source:
    pacificepoch.com/china-investment-resear...
    jgtec.en.alibaba.com/product/239490362-2...
    gcl-poly.com.hk/eng/press/media_photos_v...


    2. Ingot discussion & electricity

    In the furnace process Rec have an advantage in Norway with stable electricity (There have been several power outages in Chinese solar energy as of late http://www.forbes.com/sites/williampentland/2011/08/14/china-braces-for-escalating-power-shortages/?partner=yahootix) and that in Norway industrial energy is actually cheaper than consumer electricity while its the exact opposite in china. Also REC is looking to push this advantage further with the development of the worlds largest 2000kg furnace that is stated in the q2 release as entering anIntense developmentphase. LDK solar reveals a lot about Chinese costs for wafer in they investor presentation they have had recently (A side note is that a former REC leader Goran Bye is now leader of Ldk polysilicon division.) http://media.corporate-ir.net/media_files/irol/19/196973/2011InvestorDaypresentation.pdf here one can see that for a typical Chinese wafer maker 7% of costs is from electricity. Bigger furnaces allows the heating energy to be used more efficiently and should give modest gains in costs. The conclusion is that when it comes the ingots REC is ahead ofchina incand have an advantage.

    3. Automation advantage at REC

    In the next phase where the ingots are cut then REC also have advantage in patented highly automated process for doing this and should be at very low costs. Especially for the newer plants Herøya 3&4 and Singapore and this should be revealed in q3 when only this production is sold and not averaged down with the older production with higher costs. Many Chinese manufactures use manual labor for these tasks. In short term labor can be beneficial but in long term the constant same quality from automated process have an upper hand in producing quality goods. I believe the automation approach from REC is the correct strategy. One should also note as the standard of living in China is growing fast and wages are increasing.

    4. Cutting process ( Wire and slurry )

    When it comes to the cutting process REC have a small advantage in that the new plants have very good capacity to cut thin wafers down to 120 μm in size (I think now the standard is around 160 um) when this becomes viable in the future. Also how thin the wire cutting influences grams of silicon lost in the cutting process and currently REC have a very good 5.6g per watt. I have estimated with using only new production plants for q3 they use around 5 grams per watt. (REC2 on the graph.) On the other hand I dont think they have a wire manufacturing facility like Renesola that reduced cost (probably around 2-3 cents per watt) so when it comes to the wire consumable they are at a disadvantage to some Chinese competitors. The same goes for the slurry recycle process where they have opted for a partner with sic processing. (http://www.sic-processing.com/) making them average and at a disadvantage to companies with an in-house slurry recycling process facility.  

    5. Labor costs

    When it comes to labor costs REC is at a major disadvantage but they try to make up for it with automation processes. Ldk a typical Chinese manufacture have 3% of total wafer processing cost due to labor. It is probably not an overestimate to estimate at least three times as high labor costs in Norway/ Singapore combined on average. Of the 40 cents (q2 wafer processing estimate cost) in costs 3.6 cents comes from higher labor costs (9% of total costs perhaps?) and means REC needs a technological edge or electricity edge to make up for this small cents disadvantage. But luckily this is a small advantage for the Chinese companies and with electricity at around 7% of costs for Chinese competitors REC could nullify this advantage with the new 2000kg furnaces. In the graph 6 with the label REC4 I have tried to estimate cost reduction of an in-house slurry recycle & wire production and the labor moved to Singapore.

    Summary

    The processing cost for wafers around q2 was around 51 cents per watt. This includes depreciation and administration costs. Renesola and Ldk have around 26 cents costs q2 and Renesola is going for 19 cents per watt in q3 due to in-house wire production and new furnaces. It will be very interesting to reveal wafer processing costs for REC without the older factories making the price higher. Since the newer plants have higher automation labor costs are probably lower, with bigger furnaces (currently 1200kg in the modern plants) electricity costs are lower too. I also have a suspicion the sawing process is more efficient in the new plants. Taken together with the fact depreciation that for Ldk consists of 9% and Rec having written down most of its value already this should lower the costs even more. Rec have written down 80% of the value of wafers so depreciation should see a 80% reduction from around 14 to 3 cent per watt.

    Summary of potential reductions in cost for REC:

    Sales admin: 8 cents per watt (can be reduced perhaps?)

    write downs: from 10 to 2 cents per watt (this will happen q3)

    Rec poly cash cost: 26$ per kg (=13cents@5gram vs 28 cent @ 5.6 gram q2) Happens if REC goes for the Gcl-poly strategy

    Rec modern p: 0.05 grams @ 50$ = 25 cents (3 cent save q3 if really the more modern saws have less kerf loss and less silicon is spent per wafer.) If I am not correct my estimates are slightly off in total cost for q3.

    Rec wire: 3 cents less costs if Rec produce in-house wire production

    Rec labor move: 3.6 cents, what if Rec moved equipment to Singapore?

    Rec slurry: 2 cents with in-house slurry

    Rec electricity: 2 cents (4 % estimate drop in costs with 2000 kg furnaces)

    Rec increased efficiency: 17% average = higher sale prices due to more watt per wafer


    Graph 9 "Wafer cost for a 4 watt wafer in a 240W module"

    Q2 Rec sale price on average was 4,47$ per wafer and production cost was 3,1744$. Due to selling less than produced and writing down inventory margins where negative 3%. Q3 estimates 2.64$ cost of production and selling off inventory causing big margin jump. I am positive for REC wafer q3 and have reflected this in 34.6% margins in what I believe is a conservative estimate if I am correct that inventory will be sold out.

    Partner discussion:
    Two major Taiwan partners Gintech and Motech have had increased sales in July. Motech  increased revenue by 18% from the previous month and Gintech increased revenue 52%. This also hints that q3 should be good for REC wafer. solarpvinvestor.com/news.php?id=39
    BP Solar as partner of REC is a potential buyer of REC at current market prices in a move mirroring the TOTAL acquisition of Sunpower.

    Discussion about thin film adventures:

    REC have tried thin film and other technologies in the past but concluded the c-si route was more cost efficient:
    "REC established CSG Solar back in 2004 together with Q-Cells and the former management of Pacific Solar and REC currently holds 23% of the company. REC has, through CSG Solar, invested in micro crystalline thin-film technology based on silane gas being deposited directly on a glass substrate (from gas to finished module)." www.recgroup.com/view?feed=R/136555/PR/2... In 2007 the share was reduced to 8.7% and they actively decided not to pursue this technology. Rec also had a share in Everq string ribbon technology before selling off Sovello (owner of everQ.) REC later sold its shares in Sovello.

    I think personally REC made the right choice in not pursuing thin film as the future of this technology is more risky than the c-si that has a stable cost reduction road map ahead of it. Currently REC is experimenting with a cleaver process technology instead of sawing that have the potential to produce very thin wafers without kerf loss. (loss of silicon due to sawing.) This could of course be a positive surprise trigger for the future stock price of REC if it proves successful.

    Discussion about outlook possible trends:
    - china inc

    Currently my thinking is that in china we will see an increased internal market as the price point is within reach of what the Chinese government wants to start supporting the industry with FiTs. I think the Chinese manufactures will reduce another 10% in 2011-2012 to really stimulate this market and that the hope is for polysilicon prices to drop to allow for this reduction in prices without squeezing market margins for the more downstream manufacturers. So at least for China we will maybe see 45$ per kg poly prices in the end of 2011 or starting 2012. I also think that it will be hard for REC to gain market shares in this market when the Chinese government is heavily supporting theyre own solar industry. I also think a bigger chinese market will stabilize prices more and we wont see a 45% drop in 2012 like we did 2011.

    - combined mono/multi wafers

    Another trend with the release of a old patent on combined mono/multi wafers into cells that have a higher efficiency at reduced prices combining the best of both types of wafers have been since in new products like Renesolas Virtus wafer, Rec hints that its new furnaces have capability of combined production of both mono and multi wafer ingots. Possible this could mean REC will launch a new type of module in the future based on this combined wafer approach.

    - Europe demand higher efficiency, America in the middle and rest of the world cheapest watt

    Europe is growing residential projects and this demands high efficiency. So mono cells or combined mono/multi based cells will gain market shares in Europe. America seems to have both types of markets with thin film of first solar dominating utility projects in big land areas and high efficiency cells from Sunpower dominating in the home residential market. In Asia and the rest of the world I believe currently cheapest cost per watt types of wafers will dominate sales. REC has not the highest efficiency but is around the top high efficiency producers and that with modules that work even in small irradiance levels. One could say the modules of REC are suited for the European and American market and not so much towards the low cost markets with its current strategy of quality over price.

    Conclusion: I believe currently the price of REC is justified alone based on the silicon segment and the wafer segment is a free bonus that can be a huge market trigger if more successful. As I have tried to show q3 will be revealing in what the true production cost is of the more modern wafer plants compared with the older plants. And also a trigger in the future is the potential of the new 2000kg furnaces coming online that could even spark a new product line from REC.  Finally the potential of cost reduction is easier to reach for REC now than its Chinese counterparts that are now struggling to further reduce in areas REC have superior advantages: Polysilicon prices and efficiency of the wafers. Also one should keep in mind the heavy write downs in value when on looks at the book value of REC. This equipment is state of the art technology and if paired with better wire & slurry recycling and perhaps lower labor costs should produce much lower wafer processing costs. I believe the write downs have been to high and that if someone was to buy out REC then these values would be more appreciated.

    Appendix:

    I also have look at the annual reports and found some technology developments in the later years and hints of what is to come:

    2006: "Thinner wafers and
    wire are the single largest cost savings
    areas in REC Wafers 2010 Cost Roadmap."
    Herøya 3&4 " The new production lines have been designed
    to handle 120 μm wafers and cutting wire"

    2007: new crystallization technology for higher quality ingots. Advances in using thinner wafers and wires.
    210 million NOK r&d center at Herøya for wafer

    2008: new wafer sawing process

    2009: new crystallization technologies in the ongoing ramp-up of new production at Herøya REC expects to make further progress in sawing technologies and is also progressing with tests of sawless wafer cutting processeswww.sigen.net/articles/SiGen_REC_final.pdf www.renewableenergyworld.com/rea/news/ar...

    2010: ingot furnace upgrade "Furthermore, in the second half of the year the oldest generation ingot furnaces have been upgraded to enhance process control and improve product quality."

    "REC Wafer has developed a low cost retrofit solution to its oldest generation of crystallization furnaces, which will be installed in the Norwegian plants built prior to Herøya III and IV."

    "Finally, REC has commissioned and started up its new laboratory in California, dedicated to the development of thin silicon PV technology. The first devices were completed in the quarter."

    2011: Increase in slurry recycle. New furnaces: "introduction of the worlds first 2000 kg crystallization furnacesup from todays 1200 kg"

    Speculation: New laboratory in California more research together with sigen who is located in California? The cleaver technology allows very thin silicon wafers and this could be classified as "thin film."

    Appendix 2: In the presentation q2 on page 29 they claim to produce 240 MW for q3 as a target. hugin.info/136555/R/1531802/466302.pdf
    . The most obvious target for how to gain this increased capacity is in glomfjord mono who if going at full capacity produce 75MW of mono wafers quarterly. While I think more realistically glomfjord is having a major ramp up to 55MW from 35 currently and Herøya 3&4 increase wafer production from 162.5MW to 185MW due to increased effiency of the wafers next quarter. I will try to show some implications of this if it proves true.  It is less conservative than my original scenario but since REC itself states this as a target maybe we can look forward to quite the nice q3 for REC wafer. In this scenario I have calculated no surpluss wafers from REC singapore.

     

    quarter 2011 revenue cost ebitda ebitda margins
    q1 sale price @ 5,49 cost @ 4,345 1 846 1560 286 16
    q2 sale price @ 4,47 cost @ 3,56  1 190 1226 -36 -3
    q3 sale price @ 4,0677 cost @ 3,63 1293,53 871,2 422,33 32,65
    Q4 sale price @ 4,0677 cost @ 3,45 976,248 828 148,248 15,19
    average 1326,4445 1121,3 205,1445 15,47
    Q3 excluded inventory sales @ 4,0677 cost @ 3.63 976,25 871,2 105,05 10,76
             
    Note: q2 sale price lower due to currency exchange 5.4 and q3 5.5 exchange rate estimate      
    Note: Conservative estimates on sale price as increased mono wafers and demand should increase sales    

    Table 4: New estimates if REC produce 240MW of wafers and no singapore wafer contribution.

    Graph 10 "New estimates if REC produce 240 MW of wafers q3 / q4"

    Average ebita would actually be down 0.4 % due to the inventory effecting ebitda less. And q3 would be unchanged in % but 94 million nok without inventory previously now is 105 million NOK EBITDA and q4 from 133 million nok to 148 million NOK.
    Tags: S, REC
    Aug 13 1:06 PM | Link | 3 Comments
Full index of posts »
Latest Followers

StockTalks

More »

Latest Comments


Posts by Themes
Instablogs are Seeking Alpha's free blogging platform customized for finance, with instant set up and exposure to millions of readers interested in the financial markets. Publish your own instablog in minutes.