SolarPVInvestor Forums

I noticed that and I was ready to send a private email as site refreshes were taking forever. Now it's back to normal.

Hi, I'm new to posting here. Being enjoying reading and lurking over posts these last few months. Great to finally have an active solar investor board.

My thoughts on the best bets for the coming up cycle (hopefully)

By far my largest holding is SOL. Has been for a long time. Management, low cost poly, wafer tech, tariff exemption, emerging brand with VIRTUS, relatively low CAPEX spend (and DEBT) for what they have built leads to the most cost effective fully integrated solar company by far.

My second largest holding is CSIQ. Good project business and close supplier agreement with GCL.

Both of the above will benefit hugely form any tariff on imported polysilicon.

Finally my third holding is JKS, which I think is another well managed company with a relatively low debt and good brand image.

I firmly believe in solar and find it hard to believe that the all incost of a PV module will be approx 50 cent a watt be Q1-Q2 of this year. The cost reduction in the past 18 months has been nothing short of phenomenal.

I also think Trina will do well, despite its large debt and await with interes to see how LDK is going to compete with its bloated balance sheet.

Hi griff,

Same view as I have on which 3 are best positioned for profits.

I'm also of the view that 50 cent all-in cost for a module is phenomenal. Two years ago I had this target for this decade and saw it as the first inflexion point for a new level of demand stimulation, since it can enable module prices as low as 70 cents. This enables good economic sense for major PV installation in large portion of geographies and energy markets. Can't believe we're here now. Next level is when module price is 50 cents, then PV will be the default alternative almost everywhere (i.e. 10-20% of global generation might come from PV after the great install). There's not capacity for this yet, so I think we'll see yet at least another cycle, with module prices risking shooting above $1 within 2 years, triggering a new expansion wave.

Has anybody already hit 50 cents all-in cost? Maybe I missed out on something.

From my understandting JKS currently has the lowest cost at 59 cents as of Q3 (12 poly, 47 non-poly). For Q4 they may get to 54 (9 poly, 45 non-poly). Is anybody else close to 50? With stable or increasing poly prices I don't see the 50 cents becoming very likely before the next poly investment cycle, however. Also bear in mind that the 9 cents poly from JKS are at cash cost level, and the big 4 poly guys need at least 14-17 cents to feel cozy.

As to who will perform best coming out of this cycle I have a crazy theory going: I think that in a demand recovery scenario it is the big poly manufacturers and the competitive thin film guys who will profit the most. At 40-50GW demand runrate (End of 2013) silicon prices will have already gone up significantly and poly players will be able to make good profits. Higher poly prices also translate directly to higher margins for the thin film guys (FS and SF). And prices will stay high until the next poly investment cycle runs its course.

As for the c-Si panel makers I think that the winning teams will have a balance of good market presence, healthy balance sheet, and downstream integration. I'm not too keen on technology and lowest production cost, since I fear that the spot price of components will continue to be saturated at or near cash cost level. I see CSIQ very strong in downstream as well, but also have my hopes in Trina due to the brand and the healthier balance sheet. Of course they need to speed up their work on projects, which is something that management seems to have understood.

Not really 50 cents, yet. As you say some are close to hit 55 cents right now. SOL said they think they will hit all-in cost around 52 cents this year. YGE said they'll hit 45 cents processing now and can chop 5-8 cents off that by end of 2013. So 50 cents is within reach this year.

Wasn't Hanwha talking about being below .50 but above due to lower utilization levels worth $0.06
http://seekingalpha.com/article/1058411-…ipt?part=single

Yes they too have signalled low cost. In order I have YGE, SOL and HSOL that have signalled lowest poly to module conversion costs. The supply cost can vary though. SOL will have 95% low cost (10 + 11 cents) in-house supply of poly and wafers (but only 25% cell). YGE no low cost poly so 50% and HSOL no poly and 50% wafer so only 25% control over poly and wafer cost.

Like Klothilde mentioned it is not reasonable to assume c-Si can buy poly at 10 cents and wafers at 20 cents other than during this extreme phase. So to truely claim 50 cents all-in cost vs ThinFilm you would have to make all poly, ingot, wafer, cell and modules yourself and at industry leading cost - such a company would allow 50 cents all-in cost. Right now top players will be around 55-60 cents. JKS is already there at 59 cents in Q3 and can chop another few cents off. SOL and YGE have guided that they can make the 50-55 cents range. HSOL if they add to full wafer capacity like YGE and JKS could also come within proximity of that range.

I think that you can buy wafers from GCL cheaper than one can make them. I think that scale does reduce costs if fully utilized and scale is the wafer company money maker. It is simply based on the motion of how product is made, as there is not recipe (not yet) for efficiency rating to produce confirmed or designed yield.
Good wafer is really a purity of the poly and perhaps chemistry of the junction. Good poly wafer is one sorted from the core of ingot, where mono seed was closest. Sorting is means to find it not a produced it. All the claims are still a lot of hogwash. The cell is the heart, where technological know how, can take the wafer of choice and improve it. Cell "dressing" is what is going to revolutionize things. This is why I see companies buy wafers more than ever again making them. Wafering has a byproduct (low efficiency) which gets melted again, or thrown away for due to poor purity and added cost when reprocessed. Those are costs. When it comes to making cells, you buy what you need only, same goes for module making. This is why the full integration is not a desire is carry-on. Of course in the days of few companies, integration will be a necessity versus not having wafers at all. However intuitively we know there is a lot less wafer makers than module makers, because of the by-product effect. Cell technology exclusivity is also curtailed by "me too" equipment, the same cost of labor and the same pricing factor. Module sales are brand, efficiency and channels. Finally EPC is money, ability and offered savings through material procurement. I propose that companies developing cells, and ones building EPC backlog will reach further. CSIQ, YGE, HSOL first row. TSL and SOL second

odyd, it is not just the cost per watt to make a module that determine its product line's profitability. It also the power of that resulting module that matters. A standard module like HSOL make is now 240W. YGE will make 250W average in mass production and SOL average 255W in mass production already. If a plant cost 1.60 $/watt to build with module cost at 60 cents and the BOS at $1.00, then 10W extra in the YGE module save 4 cents on the BOS and the 15W on the SOL module save 6 cents per watt in plant cost. The module sell can basically add these BOS savings to the price of their module. This gives the different ASP ranges for the different power rating ranges we see in the ASP reports from Solarzoom. To simplify the analysis with an assumption that all have same ASP per watt this BOS cost saving power premium can be taken off the panel cost instead, i.e. actual per watt cost for YGE panels are 4 cent lower than their stated $/W module cost and 6 cents lower for SOL when looking at the profitability.

This in combination with SOL controlling wafer cost to 20-21 cents regardless of component market price development and having low opex and interest is why I reach the conclusion that they will have the biggest profit margins. For their Virtus A++ wafers they save another 2 cents for cell and module maker, i.e. an 8 premium on Virtus A++ wafer compared to standar multi is warranted from a plant cost savings perspective, again leading to big profit margin opportunity for this product line.

Note that $1 BOS is counting low, i.e. for large utility scale projects. On small residental scale installs, those 10W and 15W extra power from YGE and SOL modules will save even more costs.

These effects should start show in 2013, maybe a glimpse for SOL already in Q4.

Those partnering with GCL gets quality wafers (the yield of quality is still lower than SOL is my impression, SOL gets 0% standard quality part in their ingots) at discounted price and has a competive advantage among all those not fully integrated upstream and those integrated, but small scale or high cost for other reasons.

My standpoint on GCL remains unchanged. They have a multi billion market cap and multi billion shareholder capital investment to defend. Their price will, as it has down from 80 cents two years ago to 25 cents in Q3, remain a function of market price and a partner discount (say 5 cents or percent), not a function of their cost and a fixed low gross margin target (then price would have been much lower 2 years ago). They are not FoxConn. Right now GCL is losing money and it is because of the market price for wafers, not because they want to enrich their partner's shareholders instead of their own.

Still GCL is the best partner to have if you don't want to expose yourself with too much own poly and wafer capacity.

The inverted pyramid capacity model of HSOL, that TSL also has adopted to now, is well structured to balance outsourcing penalty and underutilization risks. But if you are end-to-end integrated and able to lead in cost and quality all the way, there's higher potential for the retained profits over time. IMO.

explo I did not mention cost with exception of added cost due to byproduct which you reprocess or discard when making wafers. How you obtain efficiency in wafering? By sorting or poly purity. How you obtain efficiency in cell, by technological built on sorted or pure wafers. Called raw material in each process, in cell processing is 100% efficient by design and not 100% in wafer. The mass processing costs of course on the depth of the technological process will vary for cells. In case of wafers, what you can recover due to quality of your selling wafers, you lose in the inefficient production format, which can be somewhat hidden by the scale (read GCL) and cheap poly manufacturing (read GCL again)
Therefore your statements completes my view. By design cell process is not only producing more efficient product it produces more efficient product because of the designed rating of raw material you put on line. Module almost the same if we talking automation, beyond inclusion of cells of the same rating range, but less to the CTM effects and substandard quality of sum of parts of the module making. Logistically with poly pricing and wafer being in low levels wafering is a cost not added value. Poly goes up, dynamic changes. This is why I do not see anyone else but wafer company (SOL) designing ways to have efficient production of wafers. Anyone can pick the segment and get into it, getting into wafers is probably last thing anyone would do. Ironically this maybe the very reason why SOL can make headlines. Yet I do not see wafer company lead the solar revolution, and revolution makers will command high appreciation on the market, two cents of mine.

odyd I think we wrote at the same time, so only my last message about GCL was in reference to your prior message. Might have been confusing.

Anyway I think you're right, many can buy new multi furnaces and maybe reach the 10 cents wafer processing cost that HSOL did. But that extra 8 cents of value that SOL create is much harder to replicate. So say that SOL process its HP wafers at 10 cents (11 now including non-cash depreciation of quite high capex for purchase and retrofitting older furnaces) and the other guy processes standard performance wafers at 10 cents. The the real advantage difference is actually 8 cents, since the extra power you get from the SOL wafer (in standard multi cell printing process) has that value (assuming $1 BOS cost). So I agree only those who achieve this will have this benefit, e.g. SOL already and as an example GCL (if they we're to go mid-stream) and maybe LDK, so integration with wafer capacity like JKS, TSL and HSOL has done might not be a good idea, since you will have inferior modules from those wafers compared to CSIQ and JASO based on high quality GCL wafers. But this still puts SOL in the best position, both for the wafer division and module division, right?

By the way, I would like to know more about the sorting strategy for wafers. Do you have some links?

Wafer technology I see much as purity and structure focus in the ingot growth step to minimize recombination loss and similar on efficiency. Then a lot off cost focus on the slicing step, like lower consumption from thinner wafers and wire-saws that reduce kerf loss, lower breakage, higher yields, continuous process etc (last part applies to the ingot growth too).

Cell technology I'm not so familiar with, but so far it seems recent avances are quite basic, like increasing active area (less from busbars, back contacts, square mono) and reducing reflection and other losses (surface texturing).

There's quite a lot left to do in terms of more advanced technology for the real leaps in efficiency, but for now focus is on doing many small things that gain a little more than they cost. That's my impression.

Note that the earlier in the value-chain you can add the conversion efficiency increasing measure the more value is created because of savings goes deeper down in the value-chain. That's why an edge upstream can have big profit potential (value added is a high percentage of the revenue level at that stage).

Quoted from "spiritcraft"

Quoted from "odyd12"

Yet I do not see wafer company lead the solar revolution, and revolution makers will command high appreciation on the market, two cents of mine.

This begs the question... who might be that leading revolutionary in your opinion?

If cell equipment makers will stand for the revolution in conversion efficiency I guess you want to wait for that revolution equipment generation by Applied Materials (or who it will be) before full expansion in that vertical.

Look at YGE. Their power bump of 10w was made in the ingot step. For panda it is mono ingots providing the high efficiency.

So I pointed out to the limitation of the wafer processing. All those improvements are about yield, purity leading to efficiency. Sorting is a mechanical process of finding efficient material, by measuring tools in automation or by hand using tools to recognize characteristics required to process it further. The percentage of quality wafers is less than total production. Cell is based on 100% quality material based on your design needs. This is where the technology takes a hold of wafer and can make huge difference. There is process damage and its elimination is also a factor measured. In module you lose some or more of that efficiency tightness to processes or chemistry or physics, all affecting costs and ultimately leading to greater conversion or lack thereof.
There are so many options, that none of them is clearly the lead, MWT and the PERC, PERL etc are some of them, in case of cell processing.
As I mentioned for revolution requirement you need in-house, technology design. How you get it is not by way of Applied Materials, which by nature of it is for sale to wide spread of operators, eliminating differentiation. I also said that wafering internally is a cost as opposed to add-on value, today. As the necessary evil Yingli cannot stop developing its fully integrated value chain, to not stay behind. This explains its ingot improvements. You will not see wafering capacities grow, but you will see cell and module eventually, while mass produced wafers will be bought by their specs and its limited availability will be essential in creating new relationships among companies.
In case of SOL their path includes building relationships with companies to offer SOL cell technologies to keep up with the market trends. However in my scenario, of in-house created advantage, the top players will leave the solution providers behind, another words leaving the purchaser with secondary or as we see today, vanilla type equipment accessible to all. I would be looking for cell developments and module developments from companies which have a selling channels, brand. There could be exclusive agreements, and cooperation deals precluding for years technological expansion to others in specific or multiple of faucets. Naturally because of it, equipment costs will raise again. Equipment manufacturers will want to overcome leaders and attract purchasers. However due to limited buying pool, pricing with have to come up to make those efforts viable. Again top players may buy those and make them exclusive to close off other operators. Of course there will be as many other scenarios as participants and human ingenuity allows.

Anybody know what LG & Samsung bring to the solar table, other than loads and loads of cash and top notch engineering??? Should the Chinese be afraid??? Samsung spells trouble in any market they decide to get into, as their engineering is second to none IMO.

https://solarpvinvestor.com/spvi-news/65-…ng-solar-energy

I think we are getting away from original question, which was who can make panels for 50 cents all-in cost and my point there was that not only the 50 cents cost per watt decides which panel provides lowest $/kWh for the plant. The power of the modules play a role too. So if SOL can make 52 cents all-in cost as they say and HSOL will make 58 cents with poly outsourced at 14 cents. The difference is actually bigger, since SOL makes 255w for their 52 cents compared to HSOL 240w for their 58 cents. Klothilde brought this up for the same reason I think, power saves BOS, that's why a TF panel at 12% module efficiency must cost much less than a c-Si module at 16% conversion efficiency in order for two plants built with the two solutions to have the same cost per kWh they produce. What causes the cost and power properties of the panels is not relavant in that plant economics analysis. The SOL solution makes most money for plant owners, that's why they grow fastest now. That's my conclusion at least, could be wrong. This is just the current state. In the future someone else might have the best solution to maximize kWh production per $ cost. In my opinion TF are the ones in most trouble here. SOL is first to sell, there's still 30gw+ left for others, but TF produces less kWh per $ than c-Si and there's a lot of c-Si capacity to fill the demand. For same reason SPWR with over 1 $/w in panel cost is not so hopelessly far behind the Chinese 55-60 cents as it seems, since their panels are already above 300w, way ahead of the standard 240w.

The name of this thread is the question I am replying to. The matter of the owning the lowest cost is one of the answers. It is semi-logical to assume that cheapest things are in most demand, but that is not true unless they are equal in quality. Nothing you wrote about is in any contradiction to what I have written. I talk about a corporate roadmap or perhaps industry roadmap, you are looking into getting measurement of one company against another on current factors.