I see that you keep CSIQ high, but bump JASO over the other two CN4. For Q3 ER I think the allocation match how they'll improve over Q2, except I think CSIQ is dominating all and JKS will be worst.

I'm curious about your CSIQ view now. I know you spent time working on Nano's modelling of CSIQ's future quarterly profits. Do you still do this and do you like what you see? When I look at 2013 and 2014 they show very consistent annual pattern with profit top in Q3 and bottom in Q1. It looks like the market played this pattern and that would mean that a sell after Q3 ER buy after Q1 ER scheme might be applied to CSIQ.

The other thing I think about when I compare CSIQ and Trina from a "who will lead perspective" is how Trina seems to be the 800-pound gorilla waking up now. They'll expand to 5 GW module capacity in 2015 and have built an 8 GW pipeline. Most of the pipeline is focused in China, which has a sustainable FiT and still double digit IRR for those with in-house panel supply. Although Ontario and Japan have very attractive FiT causing boom inducing IRR levels those markets feel unsustainable, i.e. government will have to do like Germany and slash FiT annually to not overheat and then IRR and developer margins will plummet and utility market segment will diminish.

I would not care too much about these worries that Trina might reclaim leader position in 2015+ if it weren't for the fact that they are around half the market cap of CSIQ and about the same as Jinko, but double that of JASO. Can these market cap relations remain for what's going to unfold coming 3 years? For now I've taken the less assuming bet of JASO only claiming a market value of 375m, but for diversification with long-term sustainable flavor I'm looking mainly at TSL although CSIQ looks very juicy (if Q3 performance is a sustainable repeat model a 3 times higher market cap would be warranted). TSL doesn't look very attractive on current performance and CSIQ looks dirt cheap now on Q3 delivery and is current leader as well, but I'm trying to understand the dynamics of progression for the different names coming years. It's not easy and I don't have a good view of CSIQ, which is the main reason I avoid it not that I don't like the view.

I also wonder a bit what will happen with supply-chain. Right now wafers and cells are cheap and CSIQ are light on such capacity and heavy on the module side and thus benefit greatly from this situation now, but as mentioned on their conference call all are adding module capacity only (Jinko goes for 4 GW in 2015, Trina for 5 GW and JASO added 1 GW this year) and wait with the expensive wafer and cell capacity (especially cell capacity as they don't want to add expensive capacity while trade picture is unclear, an expensive overseas factory overseas than will not be needed a year after built would be a huge loss of capital). Right now CSIQ can buy a CN or TW cell for $0.34, flip it to module for $0.16 and sell it with their Q3 ASP of $0.69 (although they said that ASP will be $0.64 in Q4). That's incredible margin on that module processing, beat even their Ontario plants. So for CSIQ right now everything is so perfect it is scary. To CSIQ's defence I've been saying this for 2 years and it just kept getting better for them and if we enter the next glut again they will be in best position to handle/exploit it.

The utility, which supplies power to the southern island of Kyushu, said the restriction begins tomorrow, according to a statement posted on its website. The decision doesn’t affect agreements already in place.

Kyushu Electric’s decision may also affect a 430-megawatt solar station under development on Ukujima, an island off Kyushu in southern Japan. The project, touted by developer Kyocera Corp. and its partners as the largest in the world to be built on agricultural land, has yet to secure grid access from Kyushu Electric.

• Jinko Power is not a pure yieldco, it encompasses project development, financing, EPC, and O&M. It's more appropriate to think of it as a growth IPP. In the future, her relationship with the parent company is in module sales and support in other areas, rather than project development and transfer by the parent company.
• Model shipment to Jinko Power will not be included in sales and continue to be capitalized into plant assets. It won't affect revenue and profit of Jinko.
• The preparation for the Yieldco IPO is ongoing and a top focus point of the management for the second half. Both US and Hong Kong are possible locations for the IPO. There is no specific time table yet.
• There is no plan for share offering before the IPO. The current funding is sufficient to support the development of over 800 MW projects using leverage, that is, to reach 1 GB prior to the IPO. Capital raises through other means such CBs are not excluded.
• The 225 M capital raised will be used only for projects. Jinko will gradually develop projects overseas after the IPO. A lot of preliminary work has been carried out.

• The new DG policy will benefit Jinko greatly, especially on the downstream business. Module sales is also going to grow with the takeoff of the market.
• DG is the trend of future. Jinko currently has a 600 MW pipeline. The scale of signed contracts and MOUs is even greater. The development cost is right now about YMB 7 to 7.5 per watt including modules. Jinko has been rapidly growing her DG teams, including development, financing, and engineering. Some are being outsourced.

• Jinko will use the capital used for bidding Topoint for capacity expansion. By the end of Q3, wafer, cell, and module capacities will reach 2.3 GW, 1.8 GW, and 2.8 GW, respectively.
• Regarding to retainage, it's a convention of the past. There are not many new contracts nowadays with retainage. Jinko has a total of about $30 M retainage revenue, which is expected to be fully collected within one year and a half. H2 is expected to see $5-$8 M being paid back.
• Regarding to the US tariffs, Jinko plans to pay the 2012 tariffs for the time being, which come to a combined rate of 29.18%. Jinko does not rule out the possibility of opening cell factories in the future overseas, including in North America. Jinko confirmed that the cost including the tariff is $0.62 (or higher). But with modules selling at $0.74-75+, it's still at quite profitable levels. In fact, it's similar to what it used to be using Taiwanese cells. (So GM is about 17% or even higher.)

• PT is based on stated PE on 12-month trailing EPS. Different PEs may be used to suit your own view.
• CSIQ and maybe surprisingly along with TSL (with the assumption of some plant sales) have the momentum going to the end of 2014. JKS and JASO, on the other hand, may be stories for 2015 when China demand picks up and accelerates, JKS's yieldco goes for IPO, and JASO's transformation gets close to the finish line.
• Lastly, the above is just my personal view based on assumptions which may not be correct or come to fruition.

I would add the point of low financial risk as another attractive point: Once a PPA is signed there is a beautiful cashflow visibility for the lifetime of the project because both the incoming cash and outgoing cash can be forecasted accurately (no volatile fuel cost component). This provides tremendous potential for the securitization of the cashflow and hence for low-cost financing.

However, this is not the point. My thesis was that under free market conditions there will no more profits in the PV power generating business than in the conventional power generating business. After all it is a commodity business with low barriers to entry. That is the sad truth.