Canadian Solar (CSIQ)

[explo 706]

It looks like the Micro Replus is not as integrated physically into the panel to be classed as an AC module. CSIQ is probably one step ahead here. SOL is doing storage this year too. I think there's a lot happening here. Panel makers are not just going into EPC and project development, they focus on taking more of the BOS kit even when not doing the project. I think this is an easier way to capture more of the value-chain. Trina has their mount product as an early move into BOS example.

[Guest spiritcraft]

I would imagine it cuts BOS and install costs heading towards a more plug and play type of solution. It is all exciting news and gets us closer to that tipping point.

[Guest nanofrogfish_spf]

Things have been fairly quiet in the CSIQ press room lately, so I thought I’d go back and highlight what the major achievements for CSIQ were in 2012, and how it has positioned them for a return to profitability in 2013 and beyond. There are many other smaller announcements and achievements which are not listed, but which all collectively contribute to CSIQ’s sunny outlook.

All the negative doom & gloom articles and rhetoric on the global solar industry dominating the printing presses in 2012 overshadowed much of CSIQ’s transition from being considered just another high-cost, undifferentiated Tier-2 module maker, to one of the top Tier players with a great combination of low-cost, high efficiency and low-BOS product offerings. Combining this with their differentiated project business model has resulted in the company having by far the most advanced project pipeline in the best markets and most visible and bankable revenue streams and GM’s of all the Chinese-based solars for at least the next 2 years, at least at this point in time.

I have some problems with formatting when I’m cutting and pasting to here, so it was just easier to pdf and attach my commentary...the following part may be of most interest;

“Combining the utility scale and non-utility scale data, and assuming modest growth in total MW shipments, for the full year 2013 I come up with;

- Full year revenue of $2.174 Billion

- Full year EPS of $3.65 (excluding forex, taxes and any surprise 1-timers)

These numbers are 2H back-end loaded.”

Heck, even if they just got to $2 EPS (incl everything), I’d be pretty happy...

[Guest nanofrogfish_spf]

It just occurred to me that you can't download the pdf unless you're a member. So for the lurkers, the part that matters the most (2013 profitability/2014 outlook) is as follows:

Business Outlook – 2013

- I have CSIQ generating about $1.2 Billion in revenue in 2013 for its NA utility scale project business, with well over 20% average GM.

- Expect news on CSSKY regarding project development and construction in SA. Africa, Mid-East. It will probably start off slow, then accelerate quickly similar to what happened in Ontario.

- CSIQ will continue to expand into the high growth global markets, in addition to benefiting from the huge China build-out.

Profitability and EPS -2013

I have generated a quarterly non-utility scale revenue and profit estimate for 2013 that is based on my belief that “stabilization” is here, and will remain throughout 2013. Primarily, that means that inventory costs catch up with real-time costs. I don’t really expect a real increase in prices (maybe a little short-term), but the ASP’s drops from here will be much more gradual. The big target in 2013 for cost reductions will be on BOS and soft costs. Stabilization will be supported by a 40GW+ demand profile for 2013 (my guestimate).

Combining the utility scale and non-utility scale data, and assuming modest growth in total MW shipments, for the full year 2013 I come up with;

- Full year revenue of $2.174 Billion

- Full year EPS of $3.65 (excluding forex, taxes and any surprise 1-timers)

These numbers are 2H back-end loaded.

Since almost 60% of the projected GP is coming from the visible project pipeline, and to meet these EPS numbers they just have to break even by Q4 on the non-utility scale business sales, I believe this is a very realistic number.

Some of the risks to this outlook would be if some Q4 revenue recognition getting pushed out to Q1-2014 (primarily the lump-sum TransCanada projects, as the rest would be recognized on % of completion), ASP’s suddenly resuming their tumble, or a harsh ruling from the EU that would seriously impact sales to the EU (a highly unlikely scenario IMO).

CSIQ FUTURE – 2014 AND BEYOND

Right now I estimate that they have about $1 Billion in the utility scale project pipeline for 2014, and $140 Mil. in 2015. These numbers will increase throughout the year. Right now, if they just break-even on modules sales, the EPS for 2014 (excl. forex, etc.) would be about $3.77.

Moving forward, the “next wave” demand profile should keep all Tier-1 solars busy. FIT-free markets should see GM for project development eventually level-out in the 12%-15% range. With many now predicting that the GW demand could be pushing 300 GW by 2020...well there’s certainly a lot of money to be made out there.

[Guest Klothilde]

...with a great combination of low-cost, high efficiency and low-BOS product offerings.

Could you please explain what you mean exactly with low-cost, high efficieny, low BOS? I just went over their product portfolio and it appears absolute standard to me. Standard poly modules (60 cells) ranging from 230 to 250 W. They source a lot of wafers and cells from the spot market, probably much more than any other Tier 1. This allows them to have a good cost position like JKS, on the other hand this questions their technology edge and their internal cost position. Also, the high degree of dell outsourcing has resulted in more quality issues than at other companies, e.g. in snail trails: http://www.photovoltaikforum.com/pv-module-f2/hilfe-meine-canadian-solar-csi-module-veraendern-s-t65910.html

[explo 706]

Thanks for sharing Nano. Good stuff.

With many now predicting that the GW demand could be pushing 300 GW by 2020

Do you mean 300 GW installed annually by 2020 or cumulative installs reaching 300 GW 2020? The former sounds high (but nice) and the latter sound low (not nice), so I wonder where you saw these predictions?

[explo 706]

Also, the high degree of dell outsourcing has resulted in more quality issues than at other companies

There's a recent quality report from measurements by Photon made in December available at http://info.renesola.com/. Just click on the download banner, but I think you have to enter an email address and the report will be sent to that address instead of downloaded directly. Anyway, on the poly crystalline modules test CSI did not perform well. Note that the results presentation seems to have been filtered by SOL, so it doesn't show for example how CSI did on mono modules.

[Guest nanofrogfish_spf]

Do you mean 300 GW installed annually by 2020 or cumulative installs reaching 300 GW 2020?

Sorry explo, that was supposed to be 2025. It's from the 300 GW initiative. Many think it's an ambitious goal but doable. The point really is that solar will continue to grow rapidly for at least the next decade, and there is a lot of money to be made being a leader in this industry during this period.

[Guest nanofrogfish_spf]

For the lurkers, this is a quick overview of the year in review for CSIQ in 2012, plus my expectations in 2013 concerning manufacturing. The foundation for their success in 2013 and beyond was being laid in 2012...so I think it's good to see how strong the foundation is:

CSIQ PAST – 2012

Manufacturing andR&D - 2012

- Manufacturing hit “fantasy” target of 55c to 60c all-in costs....Q4 2011 was at 84c/W, Q4 2012 at 57c/W (or less)...a 32% reduction.

- Hit lab cell eff. 2013 target of 21% in 2012, increased 2013 target to 23%. Mass producing of ELPS cells at over 19.5%.

- Multiple new certifications and performance ratings as further proof of their high quality modules.

Canada Projects and EPC –2012

- Bought, built, commissioned and sold their first 11.2 MW DC utility scale project in August, completing their first project as a developer (appropriately named CS1...”Canadian Solar-1”). This milestone marks the beginning of a whole new chapter in CSIQ’s future business model.

- Two (2) TransCanada plants totaling 20 MW AC bought, built and commissioned. Revenue recognition expected Q1. Approx. $110. Mil. rev. with GM over 25%.

- Purchased 16-plants totaling 157.5 MW AC from SkyPower, and 4 more plants from SunEdison Ontario totaling 46.5 MW DC. Total Revenue approx. $1 billion dollars at over 20% GM.

- Providing EPC services on 3-plants totaling 28.7 MW AC for Penn Energy.

- Increased output of Canadian module factory 50% from 220 MW to 330 MW...operating at 100% output, and sold out for at least the next year.

- Received loans from Bank of China ($120 Mil.), China Development Bank ($93 Mil.) and Deutsche Bank ($139 Mil.), totaling $352 Mil. for the Ontario projects.

United States Projects and EPC - 2012

- Closed deals to be the developer and/or EPC contractor on 246MW DC worth of projects, mostly in the 10 MW size, which is CSIQ’s “sweet spot”.

- Partnered with the largest privately held roofing company in the United States—PetersenDean—on a new roofing program the company is introducing.

- Entered partnership with De Lage Landen Financial Services, Inc., a global provider of financing solutions new program designed to make it even easier for commercial installers and solar developers to bring more solar projects to completion in the U.S. market.

Global Systems and Solutions business - 2012

- 50:50 Joint venture with SkyPower, called CSSKY, to develop projects in South America, Africa and the Middle East. They presently have 12 offices in these areas. SkyPower’s majority shareholder is US CIM Group, with over $9 Billion in assets under management. SkyPower boasts 8.4 GW of projects in applications and various stages of development.

- Good penetration in high margin Japanese market, supported by shipment data, primarily system kits.

- Penetration into Abu Dhabi (with Al Fahad Group), Turkey, Puerto Rco and other high growth potential areas.

Negative Business Developments - 2012

- Reduced 4Q/YE guidance 20% during Q3 (appears this was due mostly to factors other than just demand).

- TransCanada project revenue recognition slower than anticipated. In the future this should be mitigated by recognizing project revenue on % of completion basis, for both own/sell and EPC.

- CSIQ loses arbitration to LDK. This was about $30 Mil. plus the loss of a $10 Mil. deposit. I would guess this will show up on the Q4 numbers, so it could be a big one-time hit to close out last year.

CSIQ PRESENT - 2013

Manufacturing - 2013

- Presently at about 55c/W all-in costs...expect that to hit new fantasy target of 45c-50c/W YE, achieved with a combination of improved manufacturing efficiencies and higher module efficiencies. Lab cell eff. target of 23% by mid-year.

- Ramp up of ResidentialAC and later Commercial AC module product lines. Expect options that reduce BOS (New Edge/Zep Solar), and systems kits, to become increasing more popular (very popular in Japan right now). The utility scale market gets all the headlines, but it’s this point-of-use rooftop or small ground-mount that have the greatest upside potential IMO, with “grid parity” for residential spreading rapidly (and even more-so as storage becomes a more integrated part of these systems).

[Guest Klothilde]

- Presently at about 55c/W all-in costs...expect that to hit new fantasy target of 45c-50c/W YE, achieved with a combination of improved manufacturing efficiencies and higher module efficiencies.

I can see the strenghts of CSIQ in the project business but still don't understand fully the product side. Maybe you can help me out, since you seem to have in-depth understanding. Just a few questions: - Where have you heard about 55 all-in costs? The latest figure I have is 57 cents from the updated company presentation. - Why do you expect 45 -50 cents? Imho this is not realistic. In Q4 (company presentation) the cost breaks down to 22 cent wafers (mostly purchased at spot), 15 cents cell conversion, and 20 cents module conversion = grand total of 57 all-in cost. If polysilicon goes up to $25-$30/kg by year end then wafers will hit spot prices of 24-27 cents/W. If CSI can lower cell and module processing costs by 10% by year end to roughly 13 cents (cell) and 18 cents (module) then their all-in cost will be around 55-58 cents at year end. Could you please state your assumptions (cost breakdown) for your 45-50 cents? Do you expect the polysilicon and wafer spot prices to decline substantially in 2013?

Lab cell eff. target of 23% by mid-year.

I like what I'm hearing about the ELPS cell but am a little bit concerned about the relatively low production volumes so far. During last quarters CC they stated that until Nov. 15 (day of CC) they had shipped only 13.6 MW of ELPS modules cumulative. Thus my question to you: How fast will the conversion of the mono lines to ELPS and the ramp-up of product shipment occur and what will be the impact on production costs (specifically cell conversion cost)? Thanks in advance for your help.

[Guest nanofrogfish_spf]

Why do you expect 45 -50 cents? Imho this is not realistic.

Maybe the 45c/W is reaching a little...I would put that more in the “wouldn’t be surprised” category, based on Qu’s notoriously ambitious goals. But 50c or less would be my “expectation”. I saw all the same comments and calculations last March on how 55c-60c was impossible by the end of 2012...turns out they not only hit the target range, but actually almost hit the bottom of that range. I don’t expect further reductions to come from upstream, but rather it would more likely come as a result of more automation, better manufacturing techniques, lower use of materials, and from better panel efficiencies...all areas you can expect small, but steady improvements. And this should be helped by higher utilizations due to higher global demand. And who knows what else they may have up their R&D sleeves... This is just my opinion...we’ll see in March what they actually have in mind...if it’s 50c or less, you’d have to ask Qu if you want specifics. But I also think 50c/W all-in cost by YE is much less of a fantasy than thinking that any part of the value chain will be doubling their selling costs in the next 11 months... This thread was not intended to be a discussion on what CSIQ’s YE cost targets may or may not be, or an in depth discussion on any of the other micro-parts of the equation, but rather to highlight CSIQ’s potential profitability and EPS in 2013 and beyond due to the accomplishments they made on both sides of the business. Bottom line, this is what most investors ultimately care about anyway. If you want to dig down and discuss any of these other more focused topics, please start a new thread to do so...and if it all revolves around SOL, then on the SOL board. I also don’t see how repeatedly attaching links to SOL’s website on the CSIQ board offers anything of substance to the topic being discussed...which on this board is CSIQ. Thanks for your understanding nano

[Guest Klothilde]

ok, I understand that 50 cents is your opinion and that the company so far hasn't given any guidance. My personal opinion is 55-58 cents all-in cost by YE 2013. Since we shouldn't discuss costs any further is it ok if we at least discuss ELPS ? What is your take on the low production volume so far and how fast do you think production will ramp and at what cost?

[Guest larryvand]

Any tier 3 has the modules that CSIQ has. Nothing special. I also noticed that if you search Google for "Canadian Solar", their website says "Canadian Solar Inc. is a vertically-integrated manufacturer of silicon, ingots, wafers, cells, solar modules (panels) and custom-designed solar power". When did CSIQ become a "vertically-integrated manufacturer of silicon"???. Is that dishonest or what? https://www.google.com/search?q=canadian+solar&aq=f&oq=canadian+solar

[dydo 1,558]

Larry, could you give me the reason for your post? I haven't seen a tier three have ELPS modules. Unless there is a value to teach the reader something new, I see no point. this is what it says on the front page about Canadain Solar http://www.canadiansolar.com/en/our-company/about-us/ Once again this forum does not need this. I will ban members for this type of postings

[dydo 1,558]

Link to the datasheet for ELPS http://www.canadiansolar.com/en/products/new-product-announcements/ Nano, interesting part about the industry is that you have retail sellers and EPC sellers. The retail sellers are the ones which would want to boost the efficiency for their product. EPC sellers is about overall quality, but it is also ability to gather contracts, sales channels etc. No doubt CSIQ, the best in the category from global Chinese. YGE at home. CSIQ high efficiency products sell a lot in JP. Pretty good recogntion for the company.

[explo 706]

I also don’t see how repeatedly attaching links to SOL’s website on the CSIQ board offers anything of substance to the topic being discussed

You must mean this:

There's a recent quality report from measurements by Photon made in December available at http://info.renesola.com/.

The Photon report provided some independent test data on CSIQ's modules conducted as recent as December. I thought it would add to the CSIQ discussion, since a question about the quality of their panels was raised. I'm attaching the pdf directly instead, so you don't need to use the link. Didn't work (40 kB too big). Attached is a snapshot of the poly modules result table, which included CSIQ (aka CSI). I have the mono and quasi results too, but not featuring CSIQ.

[dydo 1,558]

Question, how is that result translates to sales of their modules? Do you have any measurement that gave the company called Jinko in this case, the advantage? It happens that Jinko, Renesola and Hanwha sell their modules at the lowest ASPs. It would mean those fractional differences in this test as in many others have done thus far very little to overall take on ASP?

[Guest larryvand]

Odyd, ELPS is contributing very little to their bottom line so be it because of high costs or low yields, ELPS is having issues for mass production. BTW, we've heard about ELPS from CSIQ for a very long while, and yet the numbers show that they have difficulty in ramping it up. So I was referring to the bulk of the modules that CSIQ does and there is nothing special about them versus everyone else. On the question of CSIQ being a "silicon manufacturer" as per their web site states, are they or are they not? It was a question. This would be news to me so the follow up question if they are, what's their cost per kg?

[dydo 1,558]

Larry, thank you for the reply, The bulk of the modules presented by companies on their datasheets is the same for all of them, all companies listed here. Exception here is the SOL presentation, which shows Virtus II modules and chooses not to show other ones, but still sells them. China Sunergy has also limited datasheets and QSAR ones are one of most powerful ones. I think you very astute to know the difference between keywords for one’s website and company description page. I linked here what “about us” page states. Keywords were built by the webmaster. They need to update this. Now how this came to be part of company’s dishonesty, I am not sure but this type of stinginess has no purpose here. I am familiar with all types of issues. I wrote about it here https://solarpvinvestor.com/spvi-news/300-canadian-solars-q2-projects-trouble-ahead-for-chinese-companies and the classic Pierce Lee here. https://solarpvinvestor.com/spvi-news/220-canadian-solars-elps-not-ready-for-mass-production-yet Constructive criticism is in, pettiness and one-stock cult be discriminating against other ones is out.

[Guest nanofrogfish_spf]

explo...I think you’re missing the point. This thread was not started to debate the difference between performance ratings between SOL and CSIQ. I didn’t even mention SOL once, either directly or indirectly, in my overview. Not to mention that info posted on another company’s site will obviously be filtered and presented in a biased format, not only by what’s shown, but also by what’s not shown (your own comment; “Note that the results presentation seems to have been filtered by SOL”). I’m sure there’s more to the story than meets the eye. Btw, according to this test, both the CS5 mono and CS6 poly CSIQ panels did extremely well. An even better example...can you please explain to me how the thread started by spiritcraft last week on CSIQ’s ResidentialAC launch, quickly turned into a discussion on SOL’s microinverters (which btw aren't even integral and must be externally mounted), with links again to SOL’s website, and then talking about SOL’s storage solutions, etc.? It’s just a pretty obvious pattern of attacking CSIQ and pumping SOL at every opportunity. Maybe it's because CSIQ has become one of the top leading solar companies, and therefore it naturally becomes a target because it's the example everything else is compared to (bash the leader to make one feel better about their own investment). I didn’t come here for endless SOL vs. CSIQ debates. Quite frankly I don’t care one bit about SOL. And I hold nothing against you for that being your #1 choice. But please keep all the stuff about SOL on the SOL board. I hope you can see this from my point of view... nano

[dydo 1,558]

Nano, It seems I also went into the the same corner rating ELPS. Sorry about it.

[Guest nanofrogfish_spf]

no problem...it's easy to get drawn totally off topic, and some response was warranted. It would be an interesting discussion, involving all the solars, to discuss and compare their product offerings, R&D breakthroughs, and future expectations (on a separate thread of course... ).

[explo 706]

Question, how is that result translates to sales of their modules? Do you have any measurement that gave the company called Jinko in this case, the advantage? It happens that Jinko, Renesola and Hanwha sell their modules at the lowest ASPs. It would mean those fractional differences in this test as in many others have done thus far very little to overall take on ASP?

It's just a test. Beside's publishing the report on the blog, SOL actually issued a press release about it. Here's a quote from it to put the numbers into context: "Performance ratio represents the actual amount of solar power produced in comparison to the maximum possible power output. With more than 170 different modules from over 100 manufacturers, the PHOTON study is widely regarded as one of the largest and most accurate of its kind. The complete results are available in the December issue of the PHOTON International magazine." (Just to provide a source: http://phx.corporate-ir.net/preview/phoenix.zhtml?c=210622&p=irol-newsArticle&ID=1771500&highlight=) I think CSIQ previously have scored high in similar test so I wouldn't make too much of it, but the 88% does not look very good an might have created some chatter, regardless of validity.

[Guest ILOVEPV]

it is nice to read so bold predicition. I'm puzzled that a share price is still so low. With EPS $3.6 a forward PE is less than 2. Any idea why?

[explo 706]

It’s just a pretty obvious pattern of attacking CSIQ and pumping SOL at every opportunity.

I think you are missreading me here, I've focused on CSIQ, not SOL, in this thread. The panel quality question came up and there's been a big recently conducted and published test by Photon. I just shared the result for CSIQ in that test. How is that not a contribution to a discussion about CSIQ's outlook?

[dydo 1,558]

It's just a test. Beside's publishing the report on the blog, SOL actually issued a press release about it. Here's a quote from it to put the numbers into context:

The question is valid and results have minimal contribution today, so I would say none. Time will tell but it is one of the tests, and many tests show high efficiencies (or output) from companies I have never heard of.

I just shared the result for CSIQ in that test. How is that not a contribution to a discussion about CSIQ's outlook?

guess your question could be taken as a point. It could simply have no impact on anything at all today. We ELPS has excellent scores but it is not a module which has a demand, with exception of Japan perhaps. I think we are all ahead, when we are discussing efficiency, as a selling medium. We are not there yet, companies are not there yet. Not to the scale. I must admit that Virtus II is a part of revolution, which has not taken off quite yet in my opinion. I hope technological revolution will take place, as to the degree this will be the differentiation we are all looking for.

[Guest nanofrogfish_spf]

it is nice to read so bold predicition. I'm puzzled that a share price is still so low. With EPS $3.6 a forward PE is less than 2. Any idea why?

Excellent question PVLover...I wish I knew the answer. Just a month or 2 ago, I had a much lower 2013 EPS estimate. This was because I had a much lower GM modeled for non-utility sales. which was a drag on the project profits and overall OM (just wasn't sure when the ASP free-fall was going to stop). Now that it looks like things are stabilizing, and demand predictions are rising quickly, I can see a big improvement on that part of the business...I've now modeled this with increasingly better margins, but still only with a breakeven 4Q (non-utility scale GM covers all operating costs). Believe it or not, there's also upside to this number if they can actually turn some type of profit on this part of the business at any point in the year. So this may be part of it (estimate revisions by many haven't happened yet...still based on the falling ASP model) My estimate is based on this pricing stabilization remaining throughout 2013 (prices can even gradually drop, but the free-fall is over, and if they stay steady or even rise a little all the better), and coupled with most of the earnings being primarily in the 2H, many of these funds may just still be in a wait and see, or nibble only mode. I'm sure there's many other factors at work also, including the cloud of negativism and denial that still covers much of the industry, and uncertainty about the upcoming EU decision. I do think some funds have been slowly accumulating though based on the charts, and I just noticed that 1 of the 2 analysts covering CSIQ (lol), recently changed their red to black on 2013 EPS...one is now at 15c and one is at $1.77. btw, with only 2 analysts (and 4 listed on their website), they could just be flying under the radar screen because no-one is really covering them! As far as the project revenues and GM's go...it's all very visable and well documented...you can easily verify all of this yourself...they just need to break even on those module sales by the end of the year! I'm betting on my own DD and calculations, and that at some point the fundamentals will win out...and that's why I believe this is just the beginning of a multi-year run for CSIQ...

[explo 706]

odyd, I think the Photon ranking is not about how much power a module produces. I think the ranking is about the difference between how much power it produces compared to how much it was supposed to produce according to its power rating. It's more of a brand trust test than a technology test.

[Guest spiritcraft]

I had to go check after reading the above Nano... you are absolutely correct. CSIQ seems to be the only Chi-Solar to have even one analyst predicting a profit. It was an -$0.80 loss a week ago. I of course hope the $1.77 analyst is correct as they are one of my four Chi-Solar holdings. So who are the analysts?

[dydo 1,558]

Explo, I got to be honest. I did not care if this was a test how they tasted. There are many tests out there, measuring many things and I am glad companies meet expectations more or less. The biggest test is selling modules. Does CSIQ sell modules? Yes. If any of those tests point to loss of sales, I will pay more attention. Thus far being the best in any of those has not increased sales for anyone (judging that some of the winners of those are nearly bankrupt or went bankrupt). That is the real meaning of my reply.