Editor's note: First Solar stock was trading at about $134 per share when John Hempton of Bronte Capital wrote this piece in April of 2010. It's currently trading at $25.80 per share, close to its 52-week low. I communicated with Hempton via email this week and he said that his "thesis turned out to be correct. Chinese manufacturers got to a price that killed First Solar's margins."

His strategy of shorting First Solar was not popular on our comment boards at the time. Here's his article in its entirety.

***

Investing in technology stocks has lots of traps for neophytes -- and by and large, we are neophytes, so we do not do very much of it. We do, however, spend a lot of time thinking about it, primarily because we are scared of what technology can do to other businesses. The demise of many low-tech newspapers provides a good demonstration of why -- as investors -- we should think about technology.

Technology offers a level of value creation that few other industries can match. In Australia, Cochlear has created enormous value and improved the world. It can literally plug a bionic ear into someone's brain stem and get them to hear. And the stock has paid about 20X, which is better than anything in our portfolio. Many of the biggest fortunes were made in technology. But technology -- and specifically, technological obsolescence -- has thrown many a fine company to the wolves. Palm, for instance, is likely to go bankrupt even though many of the concepts it pioneered are still in everyone's pockets.

We do, however, have a framework to hang around our (limited) technology investments. A technology, to be a really great investment, must do two things. It must change part of the world in a useful way (a big part of the world is better, of course, but you can be surprisingly profitable in small niches), and it must keep the competition out.

In technology, the competition is remorseless. In most businesses, the competition might be able to do something as well as you -- and it will remove your excess profit. People will build hotels, for instance, until everyone's returns are inadequate but not until everyone's returns are sharply negative. Even in a glutted market, a hotel tends to have a reason to exist: it still provides useful service. And someday, the glut will go away so the hotel will retain some value.[1]  In most businesses, the game is incremental improvement. If you get slightly better, you can make some money for a while. If the competition gets slightly better, you will make sub-normal returns until you catch up.

In technology, the threat is always that someone will do something massively better than you and it will remove your very reason for existence. Intel's Andy Grove, one of the most successful technologists of all time, titled his book Only The Paranoid Survive.  He meant it.

If your technology is obsolete, the endgame is failure, often bankruptcy. Palm will fail because Palm no longer has a reason to exist. If we wait 20 years, Palm will be even more obsolete, but the hotel glut will probably have abated. Nothing left in Palm is likely to have any substantial value. Businesses that produce plenty now will likely produce nothing then.

Surprisingly, changing the world looks like the easy bit. Plenty of companies do it. The problems are in keeping the competition out. Only a few do that (Microsoft and Google are some of the few that seem to). Hard drive makers changed the world (they allowed all that data storage, which made things like digital photography and internet multimedia possible). But they never made large profits -- and they trade at small fractions of sales.

The limited technology investments we have made are not driven by any real understanding of the technology. Sure, we try -- but if you ask us how to improve the laser etching on a solar panel, then we will not be able to help. The driver of our investment theses in almost all cases is watching the competition.

A simple example is Garmin. We have a small short position in what is a very fine company. Garmin, once a small avionics company, led the mass marketing of satellite navigation and allowed John -- without stress -- to find his son's Saturday sports matches. Sat-Nav, it seems, has saved many marriages and meant that school sports teams do not run short on players because Dad got lost.

Garmin has over a billion dollars of cash on the balance sheet -- and that cash represents past profits. It has changed the world, and thus far it has been well remunerated.

The only problem is that they can't keep the competition out. Nokia has purchased a mapping company. The iPhone now has a Tom-Tom app, downloadable for $80 in Australia. Soon Sat-Nav will be an expected application in every decent mobile phone. Google has mapping technology, too, and will soon embed it into their Android phone. Eventually, the maps will be given away because people might book hotels using their Sat-Nav device while they are traveling. (It is darn useful to know where a decent hotel with a vacant room is located when you are on the road.)

Garmin has a great product. It has improved my world. The only problem is that they can't sell their product at any price that competes with free. Garmin's business is going the same direction as Palm. Bankruptcy, however, is only a remote possibility -- they have a billion dollars on the balance sheet and unless they do something really stupid on the way down, they will remain a profitable avionics business.

Is it fair that Palm is facing bankruptcy? Or that Garmin is being displaced?  We don't think so; but then again, capitalism is not necessarily moral or fair. As a system, however, it does produce goods and services quite well. We don't invest on the basis of fair; we invest to make good returns.

The solar industry -- and the possible failure of the good

First Solar is a company that improved the world. It has driven the cost of solar cell production to quite low levels and has made utility-scale solar farms viable with only modest subsidies. There are some places where solar is now viable without subsidies.[2]

Our biggest short position, though, is First Solar -- a company for which we have little but admiration. There is a distinct possibility that First Solar's business will fail in the same way that Palm's or Garmin's has. It won't be fair -- but fairness has nothing to do with it. Like Garmin, First Solar probably won't go bust because it has a billion dollars in liquid assets on the balance sheet, assets which represent past profits.

Moreover, we suspect that First Solar's profits are about the same as the rest of the industry put together. The stock still trades with a high-teens trailing price-to-earnings ratio -- i.e., it's a fading growth stock. It hardly looks like a failure. It is a strange conclusion to come to, so we should explain how we got there. To do that we need to explain how a solar cell works.

How a solar cell works

To make a solar cell you need three things.

1) A substance which is excited (i.e., spits off electrons) when a photon hits it.

2) A layer which separates the electrons. This layer is usually a semiconductor, which means that electrons go through one way and cannot go back.

3) Something at the back which conducts the electrons away.

Thin film versus wafer

Traditional solar cells were made with a semiconductor ingot cut to a thin sheet. On one side was 'doped' with a substance that kicks out electrons. The other side was laced with wires to conduct the electrons away. This was expensive.

There were generally two types of ingot: monocrystalline, in which the wafer structure is perfect or near perfect, and polycrystalline, which has visible crystals in the wafer. Monocrystalline wafers are primarily used for computer chips (where atomic level imperfections are problematic) and are expensive. Polycrystalline silicon is cheaper.

For most large-scale uses, polycrystalline wafers are sufficient. These have about a 17 percent conversion rate, which means that 17 percent of the photon energy that strikes them is turned into electricity.

The ingot itself was a substantial part of the cost of a photovoltaic cell. Polycrystalline ingot used to sell for $450 per kg. [That was a spot, not contract, price. --- ed.]

First Solar (and others) developed a process for making solar cells with considerably less semiconductor material. The company has a cadmium telluride process which vapor-deposits a semiconductor at atomic-level thickness and comes up with a cell that is now exceeding an 11 percent conversion ratio.

This company is a technological wonder. Glass goes in on one end of the manufacturing process and comes out as solar cells at the other with next to no human intervention. Labor is used only when it comes to putting frames around the glass and for similar tasks.

This was revolutionary. It made cheap solar panels and hence made possible large, megawatt-size, commercial-scale plants. This is enough to supply a few hundred households -- not earth-shattering, but it represents a complete revolution in the solar industry.

We can think of few other companies that have pushed a technology so far and with such high environmental benefits. Companies like this will allow us to maintain a modern lifestyle while addressing greenhouse emissions issues.

Still, for all the benefits of First Solar’s cells, they are inferior in many important ways to a polycrystalline cell. Their efficiency is lower, which means you do not get as much solar energy off the constrained roof space. Secondly, while they save a lot on the semiconductor part of the manufacturing process, they have to use more glass, more wires, etc., to generate the same amount of solar electricity. Each cell generates less electricity, as well, so inverters, connectors, and installations all cost more with thin film. Thin film also degrades over time. First Solar warrants its cells' performance over their lifetime -- but with the warranty accounting for lower levels of performance in the second decade of operation. Thin film does, however, have some advantages in low light, such as the ability to retain a slightly greater proportion of their peak capacity.

Indeed, the main advantage of thin film is cost -- and that cost advantage has been driven by the cost of the semiconductor component. After all, ingot did cost $450 per kg at one point.

That cost advantage made First Solar absurdly profitable, and they have used that profit to grow into a behemoth. Revenue has grown from $48 million to over $2 billion. Gross profit (before selling and administrative costs) has grown to over $1 billion. We do not want to tell you how far the stock ran for fear of invoking insane jealousy. This stock would have made Berkshire Hathaway shareholders jealous.

But remember: all of that was predicated on a cost advantage (almost all other things being inferior). And that cost advantage is predicated on expensive semiconductor material.

Competition cometh

To recap, in order to make money in technology, you need to do two things. First, you need to change the world (which First Solar clearly did), and secondly, you need to keep the competition out. Alas, very few businesses manage the second trick.

In First Solar's case, the competition has come in a couple of forms. First, it came from Applied Materials. Applied Materials (or AMAT, as the company is known) is the most important company in the world you have never heard of. It is the dominant maker of capital equipment that goes into semiconductor factories and it is thus the company that, more than any other, provides the kit to keep Moore’s law active.

AMAT has tried competing head-on with First Solar in the thin film space. AMAT developed the vapor deposition equipment that made large-screen LCD televisions possible. This entails deposition on large sheets (5.6 square meters), which are then cut down into several large-screen TVs. An imperfection in the vapor deposition shows on the TV as a bad pixel.

AMAT appropriated this technology for solar. The silicon semiconductor is not as efficient as First Solar’s cadmium telluride technology and it is equally subject to the Staebler-Wronski effect. However, AMAT can make much larger panels than First Solar (with comparably lower wiring, inverter, and balance of system costs). AMAT’s thin film business could do some damage to First Solar, but it is unlikely to kill it. (Indeed, AMAT appears to be de-emphasizing that business for the reason discussed below.)

Far more important have been developments in the wafer business. AMAT (often the protagonist) has developed wire saws for cutting wafers into thinner and thinner slices. They are now 80 microns thick. These wafers are so thin that they flutter down in the air and break if held on their side. AMAT will, of course, sell the whole kit for handling these wafers, including laser-etching material and the equipment needed for other steps in the manufacturing process. Much less semiconductor material is needed in the wafer business.

However, the price of ingot has fallen, and spot prices are now $55 per kg -- a lot less than $450. First Solar’s advantage is entirely dependent on the fact that the company uses much less semiconductor than wafers -- an advantage that disappears entirely as wafer prices fall. At that point, all of First Solar’s many disadvantages will begin to shine through.

We are trying to work out the cost structures of the polycrystalline manufacturers, but it looks to us that the extra glass and other balance-of-system costs that First Solar panels have are now getting close to completely removing the advantage of low semiconductor material usage.

If that happens, First Solar is toast. It probably won’t file bankruptcy, because it has so much in past profits to fall back on, but it will be every bit as obsolete as a Palm organizer is now or as a Garmin car-based navigation system might be in five years.

We do not wish failure on First Solar, and if we are right in this assessment, it could not have happened to a nicer company (no irony intended). Capitalism is not fair -- and technology investment is particularly unfair.

We don’t make money from fairness. We make money from getting the business analysis right and betting on (or betting against) the right business -- and in this case, we are betting against the most successful company in a massively important growth industry.

If we are right (and we think we are), then we will make money from the demise of a company that has much improved the world. We like to think our business is noble, and sometimes it is -- but in this case, we can see why people dislike short-sellers. Their opinion, however, is not our business.

***

[1] Unfortunately, the hotel is usually mortgaged -- and the value often reverts to the debt holder.

[2] One way amuses us greatly. Walmart started putting solar cells on the rooftops of many of their super-centers in the southeastern United States. They did this originally because of implicit subsidies. However, the test centers showed something quite interesting. Good solar panels turn quite a lot of the energy hitting the rooftop into electricity that is conducted away. That energy does not get turned into heat in the building, and as such, the cooling load of the building went down. The rooftop solar installation may not have been justified by the electricity output alone, but combined with lower cooling bills, it worked out great.

***

John Hempton is the Chief Investment Officer at Bronte Capital. He has extensive experience in global markets with Platinum Asset Management, an Australia-based global fund manager. John was the chief analyst for financial stocks globally and a partner in the business. A more extensive curriculum vitae can be found here.