Amtech Systems, Inc. (ASYS) makes horizontal diffusion furnaces and related equipment used to manufacture silicon-based photovoltaic (PV) cells; they have a 40% market share globally. ASYS is also a net-net. Actually, ASYS is cheaper than a typical net-net when one accounts for the $21.6 million of deferred profit on its balance sheet as of 12/31/11. Assuming they collect their deferred profit, here is what one gets for $8.23 per share:

• $5.95 per share cash and securities (no debt).
• $2.08 per share deferred profit (contingent on customer acceptance of recently installed furnaces).
• A business with a leading market share of equipment critical for making silicon-based solar cells that generated $2.28 earnings per share (GAAP) for the 12 ME 9/30/2011.

ASYS is especially interesting when one considers that a) solar cells are just now becoming economically competitive with fossil fuels, b) to remain competitive, ASYS customers must continue to buy new furnaces which produce more efficient solar cells at lower costs. We plan to buy ASYS shares in the near future.

Valuation & Capital Structure

The following table summarizes ASYS’s current valuation and capital structure.

Note that ASYS has no debt and very modest lease obligations. While the company does not pay a dividend, management announced a stock repurchase plan on 8/17/2011 in which they plan to purchase up to $5 million dollars of common shares (6.2% of diluted shares at the current price) between 8/18/2011 and 8/18/2012. ASYS currently sells for less than the difference between its current assets and all liabilities, but how does its valuation relate to earnings?  The following table summarizes ASYS’s earnings since 2003.

When reviewing these numbers, it is important to make allowance for the company’s over-capitalized balance sheet, i.e., adjust for excess cash. On this basis, ASYS sells for less than 10x 2007 – 2011 earnings.  In fact, its valuation is absurd based on 2010 and 2011 results (selling for a fraction of one year’s EBITDA). How can this be explained?  We think the answer is that ASYS is in a cyclical business and they are not likely to sell many diffusion furnaces during fiscal 2012 due to a glut of PV modules coupled with softening demand for PV modules.  As explained below, the demand softening is temporary because the economics of generating electricity with PV cells is just now, after 57 years of falling prices, becoming compelling. Given ASYS’s balance sheet, it is likely a matter of time before they start selling more diffusion furnaces. Risks worth worrying about do exist and include the potential of a competing technology that could displace either ASYS’s furnaces or silicon-based PV cells. For instance, there are a number of thin-film PV cells (a different technology that do not require ASYS furnaces) which are available at lower costs than silicon-based PV cells. However, to date, thin-film PV cells are significantly less efficient than silicon-based PV cells (lower efficiency reduces end user return on investment). Customer concentration is also an issue, e.g., one customer accounted for 1/3^rd of accounts receivable. Never the less, at this point, ASYS looks well positioned within a growing industry.

The Business

ASYS makes horizontal diffusion furnaces used to make silicon semiconductors, primarily silicon-based PV cells. ASYS has a 40% market share of horizontal diffusion furnaces used to make silicon-based PV cells. Their chief competitor is Centrotherm Photovoltaics (CTN, Frankfurt Stock Exchange) which also has a 40% share. ASYS’s furnaces sell for about $1 million each and are used to oxidize, dope and anneal silicon wafers. The company also produces automation equipment for loading and unloading wafers (into diffusion furnaces) as well as semiconductor lapping equipment and materials.  

Recently, ASYS developed a furnace which produces higher efficiency N-Type solar cells (superior to the traditional P-type cells). ASYS developed this furnace with two partners including Yingli Green Energy Holding Co. Ltd. (YGE, NYSE), one of the largest PV manufacturers. Yingli had exclusive rights to this furnace technology through September 2011 and made extensive use of it to increase the efficiency of their premium Panda PV modules. ASYS is now free to sell N-type furnaces to other PV manufactures. ASYS also purchased 55% of Kingstone Technology in February 2011 for $4.1 million; Kingstone specializes in ion implant technology for semiconductor manufacturing – a technology that is likely to further boost PV cell efficiency. From a business standpoint, ASYS’s N-Type furnaces and ion implant technology are likely to drive furnace sales as PV manufacturers are forced to upgrade to remain competitive. From a valuation perspective, note that Applied Materials acquired Varian Semiconductor (an equipment manufacturer with similar technical expertise to ASYS) on 11/10/2011 for 7x tangible book value and 22x trailing EBITDA.

PV background

The PV effect is the creation of an electric current in a material upon exposure to light. Some discoveries take time to commercialize – the PV effect is one such discovery. French physicist Alexandre-Edmond Becquerel first observed the PV effect in 1839. The effect was explained by Einstein, who won the Nobel Prize in Physics in 1921 “for his services in theoretical physics, and especially for his discovery of the law of the photoelectric effect.” In 1954, three scientists working at Bell Laboratories invented the silicon-based PV cell, which was further developed by NASA in order to provide electricity for satellites and spacecraft. The following table summarizes the price development of silicon-based PV sells.

The table shows that since the invention of silicon-based PV cells in 1954, the cost has dropped from $250 / watt to $0.51 cents / watt (a factor of 490x).  Prices for wafers (the building block of PV cells), PV cells and PV modules (a grouping of PV cells mounted in a frame and covered with glass) have fallen 40% - 60% since the beginning of 2011. This is interesting because solar modules at $1 / watt are able to produce electricity at prices which are competitive (sans subsidy) with the retail price of electricity in many countries, including the U.S.  Let’s examine this assertion. 

The economics of PV electricity generation: approaching parity

Industry data indicates that as of this writing, a PV system (a system includes solar modules, a rack to hold the modules, and inverters to convert the DC output of the modules to AC) may be installed in California for $4.40 / watt  resulting in a levelized cost of electricity (LCOE ) produced $0.26 / kWH. Even at an installed cost of 4x the cost of the solar modules (modules are now selling for $1 / watt wholesale) the LCOE is approaching parity with retail electricity prices. Indeed, 22% of Californians currently pay more than $0.26 / kWH for electricity. The above data was provided by Renewable Energy Corp., a vertically integrated manufacturer of PV systems, so it is wise to get some data from an independent source: the writer’s household.  Over the last two years, the price we pay for electricity averaged 17cents per kWH (Fairfield County, Connecticut). We recently solicited a quote for installing a PV system on our roof.  Here is what we received:

Quoted PV system for our house, 11/7/2011

There are a number of issues that we have with this quote, not the least of which is that the installed cost is $5.57 / watt: 30% higher than the theoretical installed cost in California and more than twice the installed cost in Germany. But let’s assume for a moment that $33,977 is the best we can do – how much electricity is the quoted system likely to generate and what is the value of that electricity?  Using a publicly available model developed by the U.S. Department of Energy, the following table summarizes the likely output of this system. 

Expected Electricity Production, 6 KW system, Fairfield County, CT

So, paying $5.57 / watt installed for standard efficiency silicon-based solar modules produces a 3.6% return ($1,219/ $33,997) without subsidy (68 basis points north of 30 year treasuries). Based on further price checks, we can probably get the installed cost down to $25,000 (about the California price in the Renewable Energy Corp. data) which would boost the unsubsidized return to 4.9%.  We don’t get too excited about 5% returns, but the point is, solar power is now approaching cost parity with retail electricity prices and the economics of installing PV cells is likley to improve for the following reasons:

1. Higher retail electricity prices in the future. Electiricity prices tend to increase with time.  For instance, from 1980 through 2005, electricity prices in the U.S. doubled. To the extent this trend continues, PV system returns will improve.

2. Falling PV module and installation prices and more efficient solar cells. If installed system costs in the U.S. were to fall to the current costs in Germany of $2.86 / watt (Germany is not a low labor cost economy, so this is a reasonable assumption), then we would earn a 7% return on the above system. Considering PV prices have been falling for 57 years, it is probably reasonable to asumme they will continue to fall (though this pattern will not continue forever, there is no data which suggest we are approaching an endpoint). Our research suggests that the PV installation costs are billed at a premium to other construction trades – we suspect that PV installation costs will converge with the costs of other roof and gutter work.  Technology continues to boost the effeciency of PV cells (the amount of solar energy the PV cell converts to electricity). The following table shows the current efficency ratings of PV cells by technology.

PV cell efficiency by technology: 2009

Since PV module costs now represent a fraction of PV system installed cost, higher efficency cells boosts returns on capital by leveraging installation costs and roof space constraints.

3. Sunnier locations. All things equal, the economics pf PV sells are better in Florida than Connecticut. Of course, all things are never equal: electricity is cheaper in Florida than it is in Connecticut.  

And one more thought.  There is currently some debate regarding the appropriate hurdle for generating electricity with PV cells: should electricty cost using PV cells be compared to the retail price of electricity or should it be compared to the wholesale cost of electricity generated using traditional fuels? This debate is easily settled: since there are little or no economies of scale involved in deploying PV cells, the appropriate hurdle is the retail price of electricity.

How to make money from this?

The PV industry looks like an industry that is likely to benefit consumers: as electricity costs generated from PV cells fall below retail prices, consumers have the option to reduce their electric bill by installing a PV system. The PV industry also has the potential to disrupt investment returns from traditional generating methods, at least at the margin. Finally, the PV industry has characteristics we typically shun as investors: commoditized products subject to technological obsolescence characterized by falling prices – this sounds like the DRAM business.

In the case of ASYS, however, their business is likely to benefit from technological improvements, assuming ASYS is able to keep pace with innovations. This is because PV cell manufacturers are forced to upgrade their production equipment to remain competitive.  For example, ASYS’s N-type furnace, developed in conjunction with Yingli Green Energy Holding Co. Ltd., has boosted the efficiency of Yingli’s Panda modules by 1-2%.  Yingli is producing, on a commercial basis, multi crystalline PV cells that achieve 17% efficiency as of this writing, and can produce multi crystalline cells in the laboratory that operate at 19.7% efficiency. Yingli’s mono crystalline product (Panda) currently operates at 19% efficiency. As of 9/30/2011, ASYS is free to sell N-Type furnaces to Yingli’s competitors. Given the dynamics of competition among PV manufacturers, Yingli’s competitors have little choice but to upgrade their production technology to N-Type furnaces (or some equivalent offering by an ASYS competitor) in order to stay in business.  (Assuming they are priced the same, why would customers buy panels with 14% efficiency when 17% efficiency panels are available?). So these harsh dynamics enhance the businesses of equipment manufacturers like ASYS so long as they can keep developing furnaces which produce better PV cells at lower costs. In addition to the N-Type furnaces, which have been commercially deployed, ASYS is developing an ion implant furnace which they think will be commercially viable by 2014 (ion implant technology enhances PV efficiency).

A few words about polysilicon

Polysilicon is the raw material used to make silicon-based PV cells as well as computer chips. Polysilicon is a highly refined version of silicon metal, which is made from sand. About 28% of the earth’s crust is comprised of silicon, so it is unlikely the PV industry will run short of raw materials (peak sand? not likely). However, there was a shortage of polysilicon in recent years. The following table compares polysilicon consumption in 2010 to consumption in 2006. 

Polysilicon consumption: 2010 vs. 2006

Until 2004, polysilicon used by the PV industry represented about 1/3^rd of total polysilicon consumption. The data in the table indicate that the PV industry consumed 41% of the polysilicon produced in 2006.  Between 2006 and 2011, the PV industry increased its polysilicon consumption by 625% and now consumes 81% of the polysilicon produced: the tail became the dog. During this time, polysilicon prices rose from $30/ kg to $400 / kg.  The polysilicon industry responded and ramped capacity. As one would expect, polysilicon prices have fallen and are now back around $30 / kg.  Interestingly, note that even as polysilicon prices went through the roof, PV module prices only rose slightly before continuing their price decline.

ASYS History

Jong Whang, ASYS CEO, founded the company in 1981 to make quartz racks (“boats”) and accessories used to hold semi-conductor wafers while they are processed in diffusion furnaces. The company sold shares to the public in1983 and purchased intellectual property from Intel in 1984 which led to the development of their Atmoscan product line, a wafer processing system for use in horizontal diffusion furnaces. ASYS began buying components and assembling horizontal diffusion furnaces and in 1995 they purchased the assets of Tempress B.V., a maker of horizontal diffusion furnaces. In 2004, ASYS acquired Bruce Technologies, another maker of horizontal diffusion furnaces, from Kokusai Semiconductor Equipment Corp., for $3.6 million.  The timing of the Bruce Technologies acquisition was auspicious.  The combination of government incentives and falling PV cell costs caused the PV market to grow exponentially driving demand for ASYS’s furnaces.  At this point, many government incentives are still in place, but as described above, are becoming unnecessary due to the improving PV economics.

ASYS Revenue by Country