Contradictions in the Solar Market

The solar industry is faced with a huge oversupply of solar panels planned for production in 2008, but no one seems to notice… or care. Shares in many solar companies such as Evergreen Solar (ESLR), First Solar (FSLR), SunPower (SPWR), and Suntech Power (STP) have surged with the booming solar market .

In the past few years, we have witnessed a stampede of startups entering the solar cell market using thin film technology because of a shortage of polysilicon material used to make crystalline cells. At the same time, existing thin film solar suppliers have announced large expansions as a means of reducing production costs and gain a competitive edge. This has resulted in thin film solar panels reaching 9.4% of the 3.8 gigawatts [GW] of power generated worldwide in 2007, up from 7.6% of 2.5 GW produced in 2006. In 2008, worldwide solar power generation will grow 50% to 5.6 GW, but thin films as a percentage of panels will grow to 14.4%

At the same time, polysilicon suppliers have also initiated competitive capacity expansion plans. 2008 will be the turning point when polysilicon capacity actually exceeds demand by a mere 4,700 metric tons using a calculation that thin film panels at 14.4% of the market. If thin film solar continues at its same growth rate, in 2009 thin film will make up 17.8% of all solar power generation. That would leave a capacity of polysilicon exceeding demand by 17,000 metric tons, based on capacity expansions announced by the polysilicon manufacturers.

There are a great deal of contradictions in the solar marketplace based on these scenarios. The first contradiction is a result of the huge expansion in the industry. Total capacity announced for nearly 100 panel manufacturers (thin film, monocrystalline, and polycrystalline), comes to 10.7 GW for 2008, while solar production is forecast to reach only 5.6 GW. That equates to a capacity utilization of only 52%. By way of comparison, capacity utilization in the semiconductor industry was 90% at the end of 2007.

The low capacity utilization in the solar industry will only slightly improve in 2009 and 2010, growing to 0.57 and 0.63, respectively.

Traditional monocrystalline and polycrystalline silicon solar panels with efficiencies between 15% and 22% compare to thin film amorphous silicon of 6% to 7%, which will possibility increase to 10% efficiencies in 2009 using bilayer micromorph structures. CdTe (cadmium telluride) technology, led by First Solar, is already achieving 10% efficiency. Thus, amorphous silicon is two years behind CdTe.

The second contradiction is that if thin film solar continues at its same growth rate, in 2009 thin film will make up 17.8% of all solar power generation. That would leave a capacity of polysilicon exceeding demand by 17,000 metric tons, based on capacity expansions announced by the polysilicon manufacturers. With all that polysilicon capacity on hand, and thin film solar panels having such low efficiencies, why should thin film continue to increase? If thin film solar stayed at 14.4% of the market in 2008 as it was in 2007, that would still give an excess of nearly 15,000 metric tons of polysilicon. Even if 100% of solar panels made in 2008 were non-thin film, there would still remain an excess of 5,000 metric tons of polysilicon. So that’s the third contradiction: why such an increase in capacity by polysilicon manufacturers when it’s not needed?

The fourth contradiction is why is there a thin film market in the first place? If efficiencies are so low, particularly in the amorphous thin film sector, why do we need them in the future? They served a purpose in the past few years because there wasn’t enough polysilicon around. But why in 2008 and beyond? Clearly an efficiency 1/3 that of crystalline solar means that 3 times more amorphous silicon panels need to be used to generate the same amount of solar power. That’s a lot of residential rooftop area. Solar farms have a lot of real estate to generate electricity, but you need to install 3 times the panels there too. And that’s a lot of copper cable connections and mounts, each connection costing money for materials and labor. Is there a price benefit of thin film panels one may ask? Monocrystalline and polycrystalline panels are priced within a few cents of each other, whereas thin film panels are about $1 cheaper.

However, the mono and poly panels supply 200 W of power each, while the thin film is 60 W. Again, one needs 3 times more thin film panels for the same wattage. But the lower panel price is quickly eroded by hookup costs. Thus, the cost per watt of power generated is about the same, whether it is a thin film or non-thin film panel. And so these technologies will continue to co-exist until the shortage of polysilicon works itself out. That would be 2009 at the latest.

The fifth contradiction is the high equipment costs to make an amorphous silicon thin film panel. Up the food chain, solar thin film equipment suppliers such as Applied Materials (AMAT) of the U.S. and Oerlikon of Switzerland are selling amorphous silicon technology. Equipment costs in the neighborhood of $200 million to make 60 MW of panels. Add to that the costs of consumables. Cost to manufacture panels of amorphous silicon is about $1.70 per watt, depending of the size of the factory (First Solar, which uses cadmium telluride, has reduced its cost to $1.20 per watt). The profit for a panel selling for $2.50 per watt would be $0.80 per watt or $50 million per year. But with the equipment costing $200 million, it would in reality take 4 years just to recoup the equipment costs. And as more capacity is added, competitive pressures will drop the selling price further, not to mention Chinese manufacturers selling their product at under $2 per watt.

The overcapacity should impact equipment and materials sales in the amorphous silicon thin film area. As the shortage of polysilicon dissipates, due to ramped production and a semiconductor slowdown, prices of mono and polycrystalline silicon solar panels will drop and become even more economically competitive with thin film technology, further exasperating thin film equipment sales and the thin film solar market. With the industry having twice the capacity as it needs, expect some rethinking on the part of investors.

Disclosure: none

Comments

[fxtrader07]

great article, though i have to believe your numbers as i have no way (and not time) to check them.
based on your analysis, there aren't too many bargains in the solar stocks universe, despite the sharp drop of the past weeks.
polysilicon suppliers with cost advantages, low p/e panel manufacturers. everything thin film, especially FSLR with its obscenely high p/E ought to witness their stock prices collapse over the next 2 years ?!

[oatleaksbeans]

You should take a look at Michael Rogol's article in the February edition of Photo International (the best source for in depth information in the solar industry). Rogol (google him - he has been right before and spends much of his time in China at solar factories) writes that there is no oversupply, spot silicon prices are heading toward $500, and the profit margins in the solar industry will remain above 30% through at least 2010.

So brighten up!

[InflectionResearch]

Dr Castellano, I read several articles that you wrote last year of production under-capacity of 71,000 MT in 2010. How much of a downturn in semiconductor manufacturing does this anticipate? How are so many additional wafers going to be available?

[x3mblack]

I'm from both the semiconductor and solar industry. I don't see an immediate drop in silicon spot price. Recently announced NorSun new plant in South East Asia scheduled for 2009 Q3 production has its capacity completely sold out!
Your estimation of the equipment cost and thin film module price is not exactly right. Its higher than reality and I can't disclose the actual pricing here.
The mono/poly silicon panel may be up to 200W each but it is constructed from many smaller wafers that is 125mm sq. A single thin film panel of 5.7sqm will give you about 350W and since it is a single piece, it cost much less to produce than smaller individual pieces where eventually you need to stick them together to make a big panel.
What ever you think the contradictions might be, this whole solar craze is good for the planet.

[GH]

Mr. Castellano,

Look at the end uses of amorphous silicon thin film technology. It's building-integrated. Over half of the cost of PV power generation is in installation. Building-integrated PV replaces an existing (roof installation) cost with a slightly higher one, rather than a large incremental cost to install PV panels above an existing roof.

Your numbers may be accurate but you're not applying them correctly; ie. you're comparing apples to applesauce ;-).

[elfie]

oversupply, first sentence in Castellano's piece

this is the most ridiculous comment and is just stupid

there is substantial demand and as prices come down the demand will accelerate (increasing capacity is not the basis for saying oversupply exists)

[zawy]

I agree that the seeking alpha article that discussed oversupply is probably wrong. The reported demand appears to be a gross underestimate of actual demand. The solar companies themselves know better than financial institutions what their customers are scrambling for. There's no incentive for buyers to tell outsiders what their needs are. They just need to tell the solar manufacturers, who are reporting a very bullish outlook.

For the 1st contradiction, who's forecasting the 54% utilization? Certainly not any individual solar manufacturer. So i would question your industry-wide numbers It's also important to compare last quarter's capacity with next quarter's production. Not this year's year-end capacity with this year's production.

In the 2nd contradiction, the thin film expansion in light of poly oversupply, you're going backwards towards the same data used in contradiction number 1. It's not a different contradiction if the problem is simply an underestimate current demand. The 30% lower cost/watt of installed thin film explains the increasing market share.

The 3rd contradiction is still not a different contradiction. Over supply of poly is inherent to the argument of contradiction number 2 and still explain by an underestimate of coming production and demand.

For the 4th item, thin film (FSLR) has 2 times the profit margin and their costs will be cut in half in the next 5 years. Lower mono and poly prices will still not enable poly and mono manufacturers to keep up. That is not to say I'd buy FSLR at this PE. In fact, I sold FSLR and bought CSIQ right after FSLR jumped, but also right before CSIQ fell. Thankfully it's coming back today.

[User 160607]

Amorphous silicon panels produce 87 watts of power not 60.
Efficiency does not correlate directly to production of power. Amorphous panels operate in low light and produce more power annually . Check the Japan and Netherlands studies.

[alpha24seven]

There will be a huge oversupply of everything from Si to finished panels. It is a fact. The "huge demand" someone above mentioned is predicated upon a government subsidy somewhere. Problem is the subsidies will only subsidize a portion of total output. The global subsidy spend is flat contrary to what the media and this industry would have you believe. There is not a new Germany out there to absorb the massive new flow of product. Solar is far too expensive to sell without a subsidy. So what of the excess product? Glut leads to margin compression leads to share collapse leads to company failures leads to no interest in external funding. Solar has always been boom and bust. This time it's the mother of all busts.

A few other considerations. Solar 2.0 (I termed that phrase) will include new technologies such as Nanosolar and other innovations that aren't that far off. They will orphan both Si and thin film and makes us say "remember those big ugly panels". Not today but not ten years from now either.

[jimmy46]

Great article, maybe that's why investors are selling the solars this year. Keep us updated please.

[sunnny4ever]

According to DOE, there are 104 nuclear power facilities in USA with average capacity of 1 GW producing 20% of total USA electrical needs. This past month it was disclosed that 24 of those generators may have insufficient cooling water for continuous operation. Only offsetting those with solar panels would require 24 GW of solar panels.

Total global energy needs exceed 84 TW ( 84,000 GW),,, it will be many years of production to achieve even 25% of these requirements.

Good luck picking the winners.

[frflyer]

I don't understand the argument, that solar is still expensive because is needs subsidies. Oil and gas company subsidies are in excess of $80 billion annually, coal is subsidized and nuclear is subsidized. Some estimates of the hidden costs of oil are over $800 billion annually just in the U.S. That includes subsidies, cost of military protection of shipments, health and environmental and other costs from pollution.
The idea that the market will fix everything is not realistic when the market is already rigged in favor of existing entrenched industries. If anything, we should be subsidizing renewable energy at a much higher rate than current proposals. It currently is a small fraction of other energy subsidies.

[Lady Trader]

What's your take on EMKR? Thanks!

[User 161624]

My panels are not ugly, they're much better looking than a coal mine or a nuclear power plant. Why are people so in love with these dirty technologies?

oversupply? We have too much coal, thats why it was so cheap for so long. If the price of silicon goes down, and the cost of silicon is the major factor in the panel, and the world needs electricity, what factory would go out of business selling their panels?

[Re-REMIC]

Beauty is in the eye of the beholder 161624. What's so dirty about nuclear power? It has no emissions. Sure a little waste to attend with, but the dirty little secret about solar is that those panels don't have an unlimited lifetime. There is plenty of waste from solar, albeit perhaps not as toxic (although Cadmium is pretty nasty stuff). The bigger issue is how much capacity can we really expect from a solar setup versus a nuclear power plant. We need all the energy solutions we can implement, so don't go pish-poshing perfectly good solutions like nuclear power. Unfortunately this applies to coal as well.

[cminoc]

Anyone care to comment on the reliability of thin film? Seriously, I have an opinion, but it may not be well informed.

If they improve efficiency, but if thin film still tend to peel up, then what good is that? Mono & Poly have proven reliability. Think homeowners and business owners care about such things?

[gaucho]

One comment that surprised me was that the one about thin solar. Companies like NANO SOLAR have their prodiuct sold out for 18 months. They will not even ship samples. This material will be incorporated in roofing materials.
The demand for electricity will skyrocket in the next three to five years. The new lithium battery that is being introduced into the market will have 10 time the storage capacity of current batteries used in autos. Plug in hybrids are going to introduced at an accelerating rate starting next year. This technology will use every bit of the capacity that the electric utilities can produce including all the capacity planned for solar.

Another are for solar growth is the second and third world countries where the infrastructure does not exist and abundant sunshine does. Imagine a carport with solar cells on the roof directly charging a plug in hybrid while the driver is at work.

[1000pluspercentgains]

My bet is with ESLR for 2008. They're builing a plant in Mass. with new thin film panels that will have the best technology.

[mkreisel]

Nuke plants don't spew emissions, fine. But when they go off near a population center, 100k - 1 million dead are easily achieved.

In the market terms, it's the equivalent of selling naked calls for a living.

[supershort]

Mkreisel,
Yeah sure, nuclear power is so dangerous and unpredictable that they use it to power most naval ships that have more bombs and other dangerous stuff on them. Think about it. Even TMI was completely contained and not one person was injured. Chernobyl was a mess because it was a Russian desin that was inadequate, which I am sure the engineers have corrected by now in other plants if it needed to be.

I like this article. Am short SPWR and have puts. Crash and burn baby.

[supershort]

Also, look at the heavy insider selling. Why would an insider exercise options and immediately sell if he/she thought the stock would go up astronomical as the longs say its going. Heavy short interest in these companies is also ominous. Institutional selling is also another telling sign. The bubble has burst and looks like it will slowly fizzle.