Canadian Solar (CSIQ)

[SCSolar 83]

36 minutes ago, Klothilde said:

Well I'm starting to go over the numbers but first of all kudos to CSIQ for convincing the market that business is getting better and better despite margins heading south 😉

I need your help to make sense of the ASP/CPW numbers you guys.  They say ASP was 30 cts in Q1 and they state MSS GM at 21.6%.  Makes a CPW of 23.5 cts.  Which is kinda high to me in light of their low 20s CPW claim in the Q3 con call and the Q1 rev/shipment hack that points to 26-27 cts (pointed out by Philip Shen as well).

Also CSIQ's ASPs have historically been below the mono-PERC spot prices, and mono-PERC has been around 27-28 cts for some time now.

Help me out you guys please.

I have a hard time confirming the $0.30 ASP. I do have theories after looking at revenue segments from the PR. The theory is that revenues of module sales is included in some of the other areas of revenues such as Kits OAM and ESS Projects.

 

They shipped for Revenue 1423MW

$371M for MSS  solar modules and other products = $0.26 ASP

If you add in $25M for kits , the ASP is $0.278

if you add in OAM you get $0.306

if you take MSS $431M  you get $0.318

If you take MSS Kits OAM and Energy Projects  you get $0.323

 

Clearly in MSS numbers some of the products is inverters rails etc as part of the kits. Lets presume Kits are  1/3 other costs $0.10 for inverters maybe $0.05 mounts connectors etc. This would reduce 8.3M from the revenue side and $0.006 in lowe ASP

 

If OAM includes upgrades to modules, then there could be some module sales embedded inthose numbers. If say 20% of that is modules,  then you lose about $0.022/W

 

I presume that some of the Energy section could have embedded in it the module ASP for the NTP  projects as sales. This could be  as much as 70-80% of the services. That would lose about $0.01 on the above numbers.

 

Very odd but my best guess is  they took MSS numbers,  at $453.1. This is $0.318.

They subtracted out kits other costs and OAM non module costs(total of around $40M) or $0.028

And added in some NTP revenue say 50% or $0.008.

That would get you around $0.298 ASP.

 

 

This is just looking at the numbers and trying to figure where some of the module revenue would be recognized as a whole.

 

 

[SCSolar 83]

1 hour ago, Klothilde said:

Well I'm starting to go over the numbers but first of all kudos to CSIQ for convincing the market that business is getting better and better despite margins heading south 😉

I need your help to make sense of the ASP/CPW numbers you guys.  They say ASP was 30 cts in Q1 and they state MSS GM at 21.6%.  Makes a CPW of 23.5 cts.  Which is kinda high to me in light of their low 20s CPW claim in the Q4 con call and the Q1 rev/shipment hack that points to 26-27 cts (pointed out by Philip Shen as well).

Also CSIQ's ASPs have historically been below the mono-PERC spot prices, and mono-PERC has been around 27-28 cts for some time now.

Help me out you guys please.

I think you need to take with a grain of salt their stated costs. Typically costs are quarters end stated for internal production with current quarterly claimed margins.

 

They  have a large deferred revenue recognition with respect to their project builds and module acceptance periods of customers. This means that a large portion 30% or more could be from earlier builds as much as 6 months  earlier when their costs were much higher than Q4 stated.

 

For example it takes 1 to 2 months to freight ship. It could take another 1 to 2 months to get them installed. The contracts may also have acceptance clauses that require 6 months of running for baseline output. This suggests  that a good portion of the quarterly recognized modules for revenues could be deffered from quarters back.

 

Just look at Q1 where there revenue recognized modules were 1423MW and shipped is about 150MW more.

 

As they mentioned they have over 900MW of EPC projects going on. That is the potential of 250MW a quarter in deferred sales  until completed and acceptance. As they roll into more and more  NTP this will be worse as will some of their major customer contracts like the recent 1.8GW contract.

 

You also need to understand that not only is the deffered sales hitting their cost structure but it also props up the ASP. They are also being selective and selling to high margin markets and have built a direct sales team. These all lead to the higher ASP as well. The ASP was covered in a reply to Colin Rush.

 

 

[MVA 24]

During this weekend CanadianSolar EPS was UPgraded by analysts:

2019 = $2.23 > $2.36/share

2020 = $2.74 > $2.81/share

http://investors.canadiansolar.com/index.php/earnings-estimates

Probably on Monday stock price upgrades will follow....

[SCSolar 83]

11 hours ago, MVA said:

During this weekend CanadianSolar EPS was UPgraded by analysts:

2019 = $2.23 > $2.36/share

2020 = $2.74 > $2.81/share

http://investors.canadiansolar.com/index.php/earnings-estimates

Probably on Monday stock price upgrades will follow....

They see the same upsides I mentioned. There is another dime or 2 that could be had on top of the current guidance. I am not certain about the 2020 numbers.

[Klothilde 85]

On 5/31/2019 at 3:05 PM, SCSolar said:

...As they mentioned they have over 900MW of EPC projects going on...

Thanks for your thoughts on ASP and cost reconciliation.  I play with the same theories you've outlined.  In particular I think the $40M of EPC revenue under MSS includes some module shipments as well.

I also went over your EPS calculations and agree / disagree with some of the numbers and underlying assumptions.

At this point there's a lot of crystal ball involved since I don't have a solid feeling of how the market will evolve in H2, however I'm sceptic of CSIQ's ability to maintain module margins above 20% in the long run.  They managed to expand margins because multi wafers and cells collapsed in price but I think with improving market conditions those components will rebound in price.  I kinda get a feeling that we're already starting to see it with multi cell prices.

Also I think the EPC business ramping up will take overall margins down a bit, as EPC is traditionally competitive low-margin business (~10%).  EPC revenue is usually recognized at construction mile points (POC), so they will have significant EPC revenue already in 2019 (~700M?).

In my projections I deducted EPC module shipments from the overall module shipments figure and thus have lower module revenue than you.  Since I assume a lower module GM than you that takes the module GP down quite a bit.  However I think it is still very early to tell where things are headed, I just have a bad vibe.  Did I mention those 370MW of CN operating projects waiting to be monetized?

 

 

[SCSolar 83]

4 hours ago, Klothilde said:

Thanks for your thoughts on ASP and cost reconciliation.  I play with the same theories you've outlined.  In particular I think the $40M of EPC revenue under MSS includes some module shipments as well.

I also went over your EPS calculations and agree / disagree with some of the numbers and underlying assumptions.

At this point there's a lot of crystal ball involved since I don't have a solid feeling of how the market will evolve in H2, however I'm sceptic of CSIQ's ability to maintain module margins above 20% in the long run.  They managed to expand margins because multi wafers and cells collapsed in price but I think with improving market conditions those components will rebound in price.  I kinda get a feeling that we're already starting to see it with multi cell prices.

Also I think the EPC business ramping up will take overall margins down a bit, as EPC is traditionally competitive low-margin business (~10%).  EPC revenue is usually recognized at construction mile points (POC), so they will have significant EPC revenue already in 2019 (~700M?).

In my projections I deducted EPC module shipments from the overall module shipments figure and thus have lower module revenue than you.  Since I assume a lower module GM than you that takes the module GP down quite a bit.  However I think it is still very early to tell where things are headed, I just have a bad vibe.  Did I mention those 370MW of CN operating projects waiting to be monetized?

 

 

EPS is tough to estimate. The one thing that can be counted on is the module sales into the project. Since theyse projects are bid a year ago for many, then I can see them being higher ASP. The fact is they do not do the construction and I am not certain what the construction cost is. Many of these projects are being built for $0.80 these days. If you presume $0.30-$0.35 for the module ASP from a year ago, then there is $0.45-$0.50 in other costs.

I do not know how much would be cost in rack mounts buildings, inverters wiring road developent and land cost. The rest would be third party construction and the Engineering. I can't see the Engineering and Procurement being  anymore than say 3-5% of the cost to build as this is just paper pushing.

 

If there are 900MW of EPC @ $0.80, then I see $720M potential revenue.

Of this $270-$300M is from module sales(slightly more if all in the U.S.)

The EPC costs billable would be $36M @5%

The other materials and labor permits etc would be around $400M.

If they skim 2.5% commissions from the sub work, then they get another $10-$11M.

Total revenue to them would be  around $315M from 900MW of EPC work.

The rest of the monies goes to third parties.

If you presume 20% margins on modules they generate $55-$60M gross

The cash generation from EPC would be around $45M.

The EPC services I presume is built into the Opex.

This would suggest a cash generation of $100M from the 900MW.

 

These are my thoughts on where I am leaning about EPC

 

As for the module margins, I think you are giving to much to low cost dumping of upstream product. I believe that they have moved to more full vertical manufacturing in which with Si under $10/kg they have moved their costs towards or below $0.20 for 60% of their builds. This will aid them in cost management incase of rising upstream costs on wafers and cell rising.

 

My take is that JKS should be pushing 20% plus module margins as well if it was just upstream supply. I understand that JKS had been buying 25-30% modules at estimated 5% margins This would put JKS internal blend at around 18% margins to attain their 14-15% margins.

 

 

[Klothilde 85]

2 hours ago, SCSolar said:

As for the module margins, I think you are giving to much to low cost dumping of upstream product. I believe that they have moved to more full vertical manufacturing in which with Si under $10/kg they have moved their costs towards or below $0.20 for 60% of their builds. This will aid them in cost management incase of rising upstream costs on wafers and cell rising.

Well their MSS GM in Q1 18 was 16.7% and it dropped to 14.3% in Q2 (taking out the CVD reversal gain).  That puts the blended CPW in H1 18 around 30-31 cts/w.  A subsequent drop to 21 cts or so (~30%) in two quarters imo cannot be explained solely with cheaper poly and materials.   E.g. Poly halving from $15/kg in Q2 18 to $7.5/kg may only save you 3 cts/w at best.  In comparison multi wafers and cells have come down by 6 and 8 cts respectively since Q2  18, so I suspect this as their main cost reduction lever.  I suspect they even idled some of their own lines temporarily and loaded up on 3rd party wafers and cells.

[SolarRoof 87]

Third big project announcement in last 24 hours:

https://finance.yahoo.com/news/canadian-solar-signs-500-mw-110000891.html

[MVA 24]

CanadianSolar EPS was UPgraded again by analysts today:

2019 = $2.23 > $2.36 > $2.4/share

2020 = $2.74 > $2.81/share

http://investors.canadiansolar.com/index.php/earnings-estimates

[MVA 24]

USA grants tariff exemption for imports of bi-facial solar modules!

No link yet....

[Klothilde 85]

That would be horrible for FSLR.  What's the source?

[Mark 45]

https://pv-magazine-usa.com/2019/06/12/bifacial-beats-trumps-tariffs/

[Mark 45]

Well, I"m no longer a FSLR shareholder... made more than i lost overall, but i don't have the stomach for this.  

[SolarRoof 87]

I know their bifacial panels are a small part of CSIQ's sales, but wouldn't they suddenly become the panel of choice in the U.S. if they avoid the 30% tariff and are more efficient?  Not sure where they are on costs, but that 30% off seems like a great place to start.   Looks like it's up a bit after hours since news posted.

[SolarRoof 87]

I guess it's 25% now, but still significant.

[MVA 24]

"Canadian Solar's 2019 Earnings Could Fuel An Extended Cyclical Rally"

https://seekingalpha.com/article/4271475-canadian-solars-2019-earnings-fuel-extended-cyclical-rally?app=1#alt2