Mary Ma - Manager, Investor Relations Shawn Qu - President and Chief Executive Officer Huifeng Chang - Senior Vice President and Chief Financial Officer.

Philip Shen - ROTH Capital Partners Colin Rusch - Oppenheimer Paul Coster - J.P. Morgan Jeffrey Osborne - Cowen and Company Gordon Johnson - Axiom Capital Carter Driscoll - FBR & Co. Brad Meikle - Coker Palmer Institutional Sophie Karp - Guggenheim Securities.

Ladies and gentlemen, thank you for standing by. Welcome to Canadian Solar's third quarter 2017 earnings conference call. My name is Edison and I'll be your operator for today. At this time, all participants are in a listen-only mode. Later, we will conduct a question-and-answer session.

As a reminder, this conference is being recorded for replay purposes. I would now like to turn the call over to Mary Ma with Canadian Solar's IR department. Please go ahead..

Thank you, operator. And welcome everyone to Canadian Solar third quarter 2017 earnings conference call. Joining us today on the call are Dr. Shawn Qu, our Chairman and Chief Executive Officer, and Dr. Huifeng Chang, our Senior Vice President and Chief Financial Officer.

Before we begin, I'll remind our listeners that, in today's call, management's prepared remarks will contain forward-looking statements which are subject to risks and uncertainties and management may make additional forward-looking statements in response to your questions.

Therefore, the company claims the protection of Safe Harbor for forward-looking statements that is contained in the Private Securities Litigation Reform Act of 1995.

Actual results may differ from management's current expectations and, therefore, we refer you to a more detailed discussion of the risks and uncertainties in the company's annual report on Form 20-F filed with the Securities and Exchange Commission.

In additions, any projections as to company's future performance represents management's estimates as of today's call. Canadian Solar assumes no obligation to update these projections in the future, unless otherwise required by applicable laws. At this time, I would like to turn the call over to our Chairman and CEO, Dr. Shawn Qu.

Shawn, please go ahead..

Thank you, Mary. We appreciate everyone taking the time to join us today. The third quarter results reflect our continued business focus and strong execution of [indiscernible]. We have exceeded our gross margin target. Several key efforts in our asset monetization plan crossed the finish line in Q3, with others now ready to cross the finish line.

I will give you details on our results and plans in a minute. But if you take away one thing from the call today, it should be that Canadian Solar has executed its plan and the company is in an excellent position. We are delivering value to shareholders today and laying the groundwork for continued success as we recycle and reinvest our capital.

We exited Q3 with 1.87 gigawatts of total solar module shipments. We energized a total of 559 megawatts of solar power projects in Q3, which brought the scale of our solar power plants in commercial operation to 1.42 gigawatts. Our late-stage utility scale project pipeline stands at approximately 1.6 gigawatts as of the end of October 2017.

We're pleased with the performance of our solar module business in the quarter. Our solar module business continues to give us stability and helps us to engage customers worldwide. Solar module shipments revenue and gross margin both exceeded our expectations in the third quarter.

The strong performance was driven by the better-than-expected demand in China and the pull-in effect in the US market ahead of the Section 201 ruling. As a result, the average selling price of solar modules was substandard in the quarter.

While our shipments to third-party customers in the US will moderate in the third quarter as we supplied modules to our own 281 megawatt Tranquility 8 utility-scale project in California, the pull-in effect appears to have driven the spot market module price higher globally and pushed low price solar module demand into 2018.

We did, however, encounter headwinds, including the cost increase of certain raw materials, such as high-purity polysilicon and aluminum extrusion products. In addition, the appreciation of Chinese yuan and Canadian dollar against US dollar resulted in foreign exchange losses to the company and drove up our production cost.

We responded by continuously expanding on the manufacturing steps in which we have technology advantages such as diamond wire-saw wafering and black silicon solar cell. These efforts helped partially offset the increase of raw material costs.

On the energy business side, we successfully closed the sale of some of our operating solar power plants and we're close to the finish line with transactions of some other projects. In the US, we signed definitive agreements with two Asian buyers to sell a portfolio of six solar power projects in California, totaling 703 megawatts.

These transactions are subject to various government approvals. We hope to close the transaction in the fourth quarter of 2017 or the first quarter of 2018, depending on the timing of the required governmental approvals.

In Japan, Canadian Solar Infrastructure Fund, or CSIF, successfully priced its initial public offering with 13 operating solar power plants with a total installed capacity of 72.7 megawatts as its initial portfolio. The listing was completed on October 30, 2017. This is the fourth and the biggest solar asset infrastructure fund in Japan.

Net sale proceeds to Canadian Solar from the initial portfolio amounted to ¥30.3 billion or approximately $270 million. Japan continues to be a good market for us. We have developed strong market reputation and goodwill in this market.

We are positive on the forward potential in Japan and committed to growing our quality project pipeline and selling operating solar power plants to CSIF to fuel its growth. India is another important market. We have been developing opportunities by working through the government bidding process. Our efforts have paid off.

As an example, during Q3, we completed the sale of a 108-megawatt project in India. We look forward to updating you on our progress and success in India as the market continues to develop.

Moving on to our manufacturing capacity, we have completed the ramp up of the new multi-crystalline silicon ingot casting workshop in Baotou where we can benefit from a lower cost of electricity. The total annual capacity is expected to reach 1.2 gigawatts by the end of 2017 through production debottlenecking.

The company plans to further increases in ingot capacity for approximately 1.7 gigawatts by June 30 and may expand to 2.5 gigawatts by the end of next year if market conditions justify.

Our internal wafer manufacturing capacity is now 3.0 gigawatts and is expected to reach 5 gigawatts by December 31, 2017, reflecting the benefit of production debottlenecking. All of our wafer capacity uses diamond wire-saw technology, which significantly increases the wafer yield and reduces the silicon usage, therefore, reducing our wafer cost.

Our solar cell manufacturing capacity at the end of the third quarter 2017 was 4.7 gigawatts. We plan to add production capacity to our Funing and Southeastern Asia plants later this year to bring the total cell manufacturing capacity to 5.45 gigawatts by December 31, 2017.

The company plans to add another 1.5 gigawatts to reach approximately 7 gigawatts by the end of 2018. Our total global module manufacturing capacity will reach 8.1 gigawatts by December 31, 2017. We will further increase it to over 10 gigawatts by the end of 2018 if the market conditions justify. Now, let me comment on our business outlook.

Looking into Q4 2017 and next year, the solar industry continues to face both opportunities and challenges. We believe that solar energy will be adopted by more and more people and the long-term aspect of this industry is bright, given the compelling fundamentals.

However, the environmental and trade policy of certain countries will likely continue to cause uncertainties.

Separately, while Canadian Solar remains an industry cost leader, the unexpected raw material cost increase and the appreciation of Chinese currency over the past few months will make it challenging for us to reach our original target of solar module manufacturing cost by the end of 2017.

Canadian Solar will continue to prioritize profitability rather than market share, focus on research and technology and selectively invest into certain manufacturing process to optimize our supply chain and cost structure. This business strategy has proven successful for us and positions Canadian Solar for continued growth and success.

With regard to the fourth quarter of 2017, we currently expect our total module shipment to be in the range of approximately 1.65 gigawatts to 1.75 gigawatts, including approximately 16 megawatts of shipments to our own utility-scale solar projects. We expect the module ASP to remain stable.

However, higher raw material cost will affect the profit margin. Revenue for Q4 2017 is expected to be in the range of $1.77 billion to $1.81 billion, including revenues from solar module sales, the proceeds of the sales of initial project portfolio to CSIF in Japan and certain project asset sales expected to close in Q4.

The management estimated the timing for transaction close of the project asset sales with our best efforts. However, uncertainty remains. As noted in our press release, revenue will be affected if some of the project sales slipped into 2018. Q4 gross margin is expected to be between 10.5% to 12.5%.

For the full-year 2017, we expect our total modular shipment to be in the range of approximately 6.7 to 6.8 gigawatts come back to 6 to 6.5 gigawatts as previously guided. We expect our revenue for the full year 2017 to be in the range of $4.05 billion to $4.09 billion.

Module shipments recognized in revenue and total annual revenue will depend on market conditions, including ASP in this quarter and the timing of close for the sales of certain solar projects. Let me now turn the call over to our CFO, Dr. Huifeng Chang, for a more detailed review of our results for the third quarter. Huifeng, please go ahead..

Thank you, Shawn. As Shawn just outlined, we have taken actions to improve Canadian Solar's financial strength as we move to the final phase of asset monetization. While there are challenges, we are confident that Canadian Solar will continue to expand upon our global leadership position by prioritizing profitability over market share.

For the third quarter, net revenue was $912.2 million, up 31.8% sequentially and up 38.8% compared to a year ago. Gross profit for Q3 was $159.8 million compared to $167.8 million in Q2 and $117.3 million in Q3 of last year.

Gross margin in the third quarter of 2017 was 17.5% compared to 24.2% in the second quarter of 2017 and compared to 17.8% in the third quarter of 2016 and the third quarter of 2017 guidance of 15% to 17%. Total operating expenses were $102 million in Q3 compared to $84.1 million in Q2 and $90.3 million in Q3 of the prior-year.

Income from operations was $57.8 million in Q3 compared to income from operations of $83.7 million in Q2 and $27 billion in Q3 of last year. Foreign exchange loss in Q3 was $16.5 million compared to foreign exchange loss of $11.6 million in Q2 and a foreign exchange gain of $4.4 million in Q3 last year.

As Shawn noted, this was due to the appreciation of the Chinese RMB and Canadian dollar against the US dollar. We recorded a gain on change of fair value of derivatives of $1 million in Q3 compared to a loss of $1 million in Q2 and a gain of $2 million in Q3 of last year.

Income tax expense in this quarter was $6.2 million compared to an income tax expense of $9 million in Q2 and income tax expense of $16,000 in Q3 of last year.

Net income attributable to Canadian Solar for this quarter was $13.3 million or $0.22 per diluted share compared to net income of $38.2 million or $0.63 per diluted share in Q2 2017 and a net income of $15.6 million or $0.27 per diluted share in the third quarter of 2016.

Moving on to the balance sheet, at the end of Q3, our balance sheet of cash and cash equivalents was $614.6 million compared to $496.6 million at the end of Q2. Restricted cash balance was $539.5 million at the end of Q3 compared to $464.9 million at the end of Q2.

Trade accounts receivable balance was $457.4 million at the end of Q3, up from $367.6 million at the end of Q2. Accounts receivable turnover was 47 days in Q3 2017 compared to 57 days in Q2 2017. Inventories at the end of Q3 2017 was $301.5 million compared to $283.2 million at the end of Q2 2017.

Inventory turnover was 37 days in Q3 2017 compared to 52 days in the second quarter of 2017. Excluding liabilities held for sale, short-term borrowings at the end of Q3 totaled $2.14 billion compared to $2.04 billion at the end of Q2. Long-term borrowings at the end of Q3 was $318.2 million compared to $273 million at the end of Q2.

Non-recourse bank borrowings at the end of Q3 were $1.29 billion. Senior convertible notes outstanding totaled $126.2 million at the end of Q3 compared to $126 million at the end of Q2.

Short-term borrowings and long-term borrowings directly related to utility-scale projects, which include $1.22 billion of non-recourse borrowing, totaled $1.43 billion at the end of Q3 compared to $1.3 billion at the end of Q2.

The value of our build-to-sell project assets at the end of Q3 was $2 billion compared to $1.8 billion at the end of Q2 2017. We are pleased with our considerable achievements in the monetization of our solar power projects. We plan to finalize the sale of solar power project assets in the US, Australia, China and the UK in the coming months.

Our focus remains on delivering a healthy return to our shareholders, maximizing operating cash flow and deleveraging our balance sheet in order to fund the next phase of the company's growth. With that, I'd like to open the call to your questions.

Operator?.

[Operator Instructions]. Your first question comes from the line of Philip Shen from ROTH Capital. Please ask..

Hi, everyone. Congrats on the successful J-REIT IPO and progress you're making on the monetization efforts. The first question, you've had a successful IPO.

Can you talk about the revenue and earnings recognition used for the project sales to the J-REIT? And also, looking ahead, I see an estimated 80 megawatts of Japanese projects projected for COD in 2018.

How much of that 80 megawatts could be dropped into the J-REIT in 2018?.

The revenue recognition of the J-REIT's IPO will show up in Q4 results. So, we have to wait for a couple of months to see the final numbers. Basically, we recognized 85% of the proceeds because we own 15% of the CSIF, the Canadian Solar Infrastructure Fund, so we recognized 85% of the sales.

The regulation of the J-REIT is that after 9 months we can drop in new assets, which means from Q2, Q3 next year, you should see us dropping certain additional assets into this fund if the market conditions justify..

Great. And then, for Q3, can you share what the geographic revenue mix was? If I missed it, I apologize.

And then, what projects were recognized in revenues in Q3? And also, what projects are in Q4 revenue expectations? I know you mentioned the J-REIT there, but was wondering what else you have in there as well?.

For Q3, we didn't provide the geographic distribution, but China will probably be the number one market in terms of the total module sales. If it's just module sales, then China will probably be number one. And I think India and Japan will be the number two or three, I would guess. We shipped a reasonable amount of modules into US.

And as we mentioned in the press release, a lot of modules goes mainly into our own projects. So, the third-party module shipments in US is not that high..

And can you comment on the projects that were recognized in Q3?.

How many projects? All right. We recognized one project in Q3, which is this GPOE project in California..

Great.

And then, beyond the Japanese REIT, what do you expect in Q4 revenue expectations for projects?.

At this moment, we expect some of the US project sales to close in Q4. Some closing Q1 because those project sales are still pending due to some government approvals. So, we made a best guesstimate and we think probably half of them will happen in Q3. So, we estimate that three projects will get closed in Q4 and three projects closing in Q1 next year.

We also forecast to have our UK project asset closing – all right, okay. I was told that the UK project, current forecast, will put it into Q1 next year. And then in Q4, Q4 will include the Japanese IPO. I think Q4 should also include another US project. It's a 90 megawatt project in US. And Q4 also includes some project sales in China..

Okay, great. And then, just another quick follow-up on that.

Can we do the same exercise for the amount of debt that could be sold or taken off your balance sheet as a result of the project sales that you expect in Q4?.

We can provide you this number later. But we can – if the US project sells, all the debt associated with the three US projects – actually four US projects and also the Japanese projects..

Your next question comes from the line of Colin Rusch from Oppenheimer. Please ask..

Thanks so much, guys.

Could you talk about where your processing costs were ex the supply-chain impact and the currency impact, just to give us a sense of progress on the cost reduction from the manufacturing side?.

From Q2 and for Q3, I don't think we have seen much processing cost reductions. We have cut down our own processing cost, but the raw material price go up to more or less offset. Meanwhile, the ASP – the module average selling cost also remained stable..

Okay. And then, payables have been creeping up here for the last two quarters and you're seeing just a little bit above $1 billion here.

Can you talk a little bit about what's going on with the supply chain there? Are you getting better terms from someone or are there some longer items that you're able to extend terms on?.

We always enjoy good payment terms from our suppliers. And the Q3 module shipment increased significantly over Q2. That also resulted in some accounts payable increase. But I think it's also a little bit to do with the cutoff. I think it just happened that September 30, that particular cutoff, our accounts payable was high.

And then, it went down in October, November time frame. So, I think when we report Q4, you will see our accounts payable lower, probably lower than September 30. So, more or less, it's just timing..

And then, just a final one from me, just the cadence of the portfolio buildout in Japan, how should we think about how quickly you're building and completing those systems through the balance of this year, into next year? Is this an area where you can accelerate things, now that you've got the exit vehicle? Or should we continue to think about this kind of grinding out the way it has in the past?.

I think you should consider it still at the same pace that in the past because Japanese projects just take long from the grid permit to the EPC itself. It doesn't really benefit from the IPO of this infrastructure fund. But on the other hand, the Japanese project, we receive good leverage, which receive a very good financing support.

Very low interest rate. So, it doesn't really add too much burden to us..

Great. Thanks so much, guys..

Thanks..

Your next question comes from the line of Paul Coster from J.P. Morgan. Please ask..

Yeah, thanks. First off, the US asset sales. So, I assume that this is – all of the cash is going towards deleveraging. There is no sort of sense that the cash will be returned to investors directly.

Is that a correct statement?.

I think all the debt associated with the US project will be retired for sure. And for the capital, some capital may be spent on new projects, new utility solar power plant projects. Some may be spent on the manufacturing expansion project..

You've successfully sort of made huge progress actually on this vertical integration and scaling, Shawn.

As you look into 2018 and in the context of the higher input costs and the unfavorable foreign exchange shift, do you think scale can still deliver reduced unit cost or are we sort of getting to the point where that no longer delivers?.

We will continue to reduce our own processing cost. For example, our diamond wire-saw capacity will go from today's 3 gigawatt to 5 gigawatt at the end of next year. And by cutting the wafer ourselves internally, we estimate that we can save around $0.02 per watt just by cutting the wafer ourselves and using diamond wire-saw.

So, all of these will save the processing cost. But, meanwhile, we don't have direct control to the polysilicon price, for example, and we didn't expect the polysilicon price to go so high inside China and nobody expected. By June 30 of this year, we all expect the polysilicon price to go down after June 30. But it didn't go down; it actually go up.

In Q1 and Q2, the polysilicon price inside and outside China, the difference is $2 per kilogram. So, if polysilicon sales were, let's say, $12 per kilogram outside China, typically, it sells $14 per kilogram inside China. But, right now, outside China, it's still $12, but inside China it become $19.

So, this is part of the issue created by all this ADCVD, those trade war between China and US, for example. And, I guess, US thought by introducing the duty on the Chinese total modules, it will hurt Chinese volume. But, actually, we were not hurt by the ADCVD in US or rather by the Chinese ADCVD duty on the solar grade polysilicon imports on US.

So, it's just ironic, but that's what happened. And I also mentioned aluminum extrusion product, but that's more because China has been – reinforced its environment protection laws in the past six – three, four months. So, a lot of the factory were closed down in China just because they exceeded their wastewater or whatever discharge limit.

So, all the raw material, lots of the larger commodity or raw material price went up and silicon – and aluminum is just one of the item. But that's related to the big picture, macroeconomy, not just solar. So, we don't have that much control of it.

I think – but, meanwhile, there has been lots of announcement – quite many announcement on polysilicon expansion in China. So, by end of 2018, we should see much more polysilicon output inside China. I hope that helps.

But who knows? Maybe the demand in China also go up because, this year, it looks like the US – looks like the solar market in China will be 50 gigawatt and nobody expect that. I remember early this year, everybody thought China would install maybe 20 gigawatt this year.

We'll be happy if it's 20 gigawatt, but the new number seems to point to 50 gigawatt. So, who knows? Maybe the market continue to grow in China and we still don't have enough polysilicon by end of 2018. So, there are some market uncertainties.

I think long-term, things will balance, but we don't have direct control on those raw materials unfortunately..

Your next question comes from the line of Jeff Osborne from Cowen and Company. Please ask..

Hey, good morning, Shawn. Just a couple of questions from my end. I was wondering if you could just bridge the gap between the reported gross margins in the third quarter and then the anticipated roughly 500 to 700 basis point drop in the fourth quarter.

You've got a couple of moving pieces with some of the projects that you mentioned in the US as well as the Japanese REIT.

Can you just give us broad-brush strokes on what's driving that drop in gross margins?.

The module ASP pretty much remained stable from Q3 to Q4. So, we are okay there. The market sustained. But the material price, as I said, material cost increased a lot. I mentioned that the polysilicon price in China went from around $14 to $19 in the past three or four months. In Q3, we still benefited from some of the low-price polysilicon.

Some of the supply contract we signed by the end of Q2 benefited us into Q3. But Q4, we don't have that benefit anymore. The spot market will be fully reflected in the Q4 cost. And I also mentioned, for example, aluminum. We also now – all this cost increase will show up in Q4. So, our cost base, harmonic manufacturing cost did go up from Q3 to Q4.

On the project side, we forecast to close three out of the six projects in US. No, three plus one. So, probably, four projects in the US. I want to caution that this is only our forecast because we can't really control the government approval process. However, those projects in US, those are large projects.

You'll see lots of revenues, but you don't see high gross margins on those projects. However, I think the good part is that we're not going to see too much operating cost increase in Q4 either because we are closing a lot of large solar power plant sales, but the operating expenses, the people cost, all those costs are still more or less the same.

So, I think, although Q4, the gross margin is low, on the net profit side, the bottom line, we should still see a reasonable bottom line.

Does that answer your question?.

That does. Very helpful. Maybe just two follow-ups about that then.

So, first of all, I assume the same trend for the other three projects in Q1, assuming ASPs are kind of flattish and the polysilicon environment is elevated, we should think about a similar 10%, 11%, 12% gross margin for Q1 of 2018, just given the dynamics at play? So, that's question number one.

And then, just the other follow-up is, can you just touch on – again broad strokes – how we should think about the Japanese drop-downs and what the margins are of that business there?.

Yeah, the six projects, six recurrent projects to be sold in Q4 and Q1 next year, the margin is different project by project. So, when we report Q4, you will see the profit on the three projects we recognize in Q4 and, later on, the profit we recognize in Q1. But the profit are different from project to project. That's one comment.

And the second, about the cost and the pricing of the module business, that's a very good question. First of all, it looks like we should have a decision on the Section 201 in US by the middle of January. So, maybe this pull-in effect of the Section 201 should be over. That's what people believe once the decision is delivered.

Once White House delivers this decision, then it doesn't make sense for people to rush anymore. That may mean the module shipment into – the US – maybe that means the US module demand will go slower. But if the US module demand goes slower, maybe the polysilicon price also will go lower. So, this is the dynamics.

I think Q1 dynamics will be different from Q4, but whether the module pricing and the material cost go in step or have some kind of misalignment, that's something we can only find out when we get into next quarter. .

Your next question comes from the line of Gordon Johnson from Axiom Capital. Please ask..

Hey, guys. Thanks for taking the questions. I guess, first, on the interest expense and the FX impact this quarter, those are a bit higher than, I think, some people expected. Can you talk about how that is expected to trend maybe in the fourth quarter and maybe through next year, i.e.

as you sell assets, will the interest expense fall as you retire debt or maybe you can break down how much of that capital you expect to reinvest in your projects or back into the company? And then, with respect to your CapEx, you guys are ramping up capacity ingots, so module, can you talk about what your CapEx expenditure is going to be maybe next quarter and throughout next year? And then I have a couple of follow-ups..

I would like to have Huifeng to answer this question..

The first part of the question, regarding the FX loss, to put this in perspective, that RMB fall quite a while and we had the correct view that it has been in depreciation trend since early 2012. And then, we have enjoyed that benefit because a lot of our expense are RMB and our revenues US dollar. So, for many quarters in the past, we had an FX gain.

But somehow, since early this year, I think it started April, the RMB had reversed, appreciated against US dollar. Now, 80% of that appreciation of RMB actually really not RMB appreciation per se, it's the decline of the US dollar. So, we had quite a big loss number of this position.

Of course, right from the beginning, we adjusted our exposure, but we didn't go, like, 100% neutral. So, we had two quarters of loss. But coming into the third quarter, we actually can see that in the markets. RMB again started this depreciation trend from like 6.5 RMB per US dollar now to like 6.6.

And I believe, going forward, due to all the political, economical, macro factors, at least we don't see RMB again started appreciation. So, right now, our position is much neutral. We never speculate. It's always hedged on the currency we have in account. It's just the degree, the percentage, we do hedging.

So, right now, we're taking a more neutral view. So, I don't see a big number in the Q4. So that is to answer your question regarding our currency loss. Thank you..

Your next question comes from the line of Carter Driscoll. Please ask..

Good morning. Thanks for taking my question, gentlemen. So, just want to ask a question about – obviously, you're facing higher raw material costs for some period of time.

Does that play – your expectation of that, lasting beyond a temporary period, play into your decision to kind of ramp up internal wafer and cell manufacturing to try to better control cost from just a better mismatch – I mean, a better match of your internal manufacturing capabilities and/or be able to bring an elevated level of product to market to support your development activities? Just trying to get a sense of how much of that buildup is to drive down costs versus continuing supply of very robust pipeline? And then I have a follow-up.

Thank you..

All of those efforts are targeted to drive down costs. But if the raw material price continues to go up, then it can only partially offset the cost. Personally, I guess, the polysilicon price can't go up anymore. It is probably – I hope that it's how much it can – this is more or less high as they can go. From this point on, it should only go down.

That's my personal view. However, as I commented in one of the previous question, Q1, maybe the market dynamic is also different because people expect that, whilst the Section 201 decision is handed down, the rush in the US market should be over. So that market should slow down too. So that will be my thought..

Your next question comes from the line of Brad Meikle from Coker Palmer. Please ask..

Hi. Good evening, gents.

First question, could you talk about the Japan profitability with the drop-down and the listing and how that revenue is recognized and what the margins are roughly in Q4?.

Huifeng, do you want to handle this question?.

Actually, you answered earlier. Well, first of all, the transaction is in Q4, right? We usually don't talk about profitability forward. But I can give you some color, how you can – so that you can see our performance better. That is from the megawatt, close to 73, total megawatt. And we also published the transaction value, which is about $270 million.

And then, the other side, you can find out [indiscernible] in Japan. So, you have the revenue, right? And then, you have a roughly estimate of cost and then you can figure out how much profit we can get. Of course, we don't book everything. Right now, we're still part of – the 14% shareholder of the J-REIT. So, simple math. It cannot go far away.

So, you actually – other information you can find out from the Tokyo stock market..

Right, thank you.

And would a cost estimate of $170 million be roughly in the right neighborhood with a 15% discount on that $100 million in profit?.

I can't comment on that. But you see we have published all the numbers – the equity proceeds we raised and also the leverage of the J-REITs, which is about 40%. And so, you have the total number, the cash raised on the platform and how much will come to Canadian Solar to buy that 73 megawatts. It's actually very transparent.

But I cannot comment on your calculation..

Your next question comes from the line of Sophie Karp. Please ask..

Hi. Good morning, guys. Thank you for taking my question. I had a question on J-REIT also. What kind of yield do you expect it to trade at in the – I believe, at IPO, it was above 5% indicative yield.

Is that, I guess, normal for Japan? Or would you expect it to be a little bit lower at some point?.

Actually, what our IPO price, it's priced about 6.8% cash distribution, which is a great deal for the investors. However, for us, it's actually not bad at all because for that equity – in addition to the equity raised, somebody else's money, right, investors' money, we also can raise a lot of debt.

And the debt, the rate, somewhere around only 40, 50 basis points and then we swap the debt which is a floater into fixer. So, the total cost so far only 80 basis points. That's all market information. So, combine the two, all together, right, it's about close to $300 million cash and $270 million will be used to buy our assets.

So, then there are expenses to operate all these assets from the cash flow. You can also model from the FIP. So, overall, it's a win-win game between – as a developer and also the investor.

And then, actually, one thing that I'd like to point, actually we also disclosed that, is that in addition to the 73 megawatts we dropped into the J-REIT, we have about 100 megawatts COD project at this point. We have about 35 megawatts COD several months ago, but not a drop in the J-REIT.

But also after the deadline of J-REIT listing, we have added about 60 megawatts-ish new projects. So, going into 2018, you can see that we have – already have something already to move in. .

Okay, thank you..

Your next question comes from the line of Gordon Johnson from Axiom Capital. Please ask..

Hey, guys. Thanks for the follow-up. Just had a follow-up on the CapEx expenditures you guys expect maybe in Q4 and next year. If you could just give us an update on what the cost per watt in the quarter was as well as the ASP on a blended basis. Thank you..

Huifeng, do you have that number on your finger tip?.

Sorry, can you repeat the question.

Sure. If you could just address the expected CapEx expenditure in the fourth quarter and maybe next year, given the capacity ramps? And then also, what the cost per watt on a blended basis was in the quarter as well as the ASP. Thank you..

Okay. Let me comment on the ASP first. We believe the ASP going forward will be stable, probably at the level close to what we have witnessed in Q3, for both Q4 and early part of 2018. Of course, as Shawn said, that everybody's speculation.

But we see that the manufacturers, by their pricing behavior this year, in the last two or three quarters, are very different from what was really expected 12 months ago, okay? We didn't go down below $0.30 per watt. So, it's somewhere around $0.35 or high 30s. So, ASP side, I think we're okay.

The cost side, as Shawn pointed earlier, it's mainly driven by something we cannot control, the commodity market, the overall macroeconomic climate. Regarding the – you ask about the CapEx. That is something, obviously, has many uncertainties because Section 201, we are waiting for the outcome.

And then, we are also evaluating opportunities in domestic China. One thing for sure is that to maintain our global leadership position, we need to be the cost leader. And to be the cost leader, the lesson we learned from past is that we have to change our strategy in the early years, the reverse pyramid capacity strategy to a vertical integration.

So, for that, we will continue to add our capacity on cell, wafer and ingots. So, going into the new year, 2018, probably that will be our main project, that is to complete the vertical capacity integration project and then to be the cost leader..

Thank you. Please note, we have reached the end of the conference. And I would like to turn the call back to Canadian Solar's CEO, Dr. Qu, for closing comments. Thank you. Please go ahead..

Thank you, everyone, again for taking the time to participate in today's call. We hope you have a great day. Thank you..

Ladies and gentlemen, that does conclude the conference for today. Thank you for participating. You may all disconnect..