Saturday, August 28, 2010

Solar reaching GW scale

One of my favorite aspects of solar technology is that the more we make/use/buy the cheaper it gets.

There are several reasons for this from the trivial--buying any product in bulk is cheaper than buying one, two or three at a time--to the more profound like the experience curve: you get better at making any widget the more widgets you make. Also as the solar industry grows it continues to improve the efficiency of the cells that go into the panels as well as how efficiently it makes the panels (yield).

Semi-related to the experience curve is that scale brings certain advantages--I'm alternately annoyed/amused when people compare the "high" cost of solar for a home (3-5 kw system) to the "low" cost of electricity from a GW sized coal power plant and reach the conclusion that solar is simply a bad investment. This is like comparing the cost of transporting an item from NYC to LA by bicycle to the cost of doing so by freight train and then deciding that a bicycle is simply a bad investment. It is clearly comparing apples and would need to compare the cost of electricity from a GW solar plant (and preferably the cost from the nth GW solar plant not the first one) to the cost of electricity from a GW coal plant to reach a meaningful comparison. Or you would need to build a house sized (3-5kw) coal fired power system and compare that to the home solar system.

Several recent news items appear to show that we will soon have GW (or near GW) scale solar plants to compare with the coal variety...from a 1 GW solar thermal plant to multiple 0.25 GW solar plants (PV and CSP) under review out in CA. Although this won't yet be a perfect comparison it will at least become a meaningful comparison.

Just as you would organize the installation plan and procedure for installing 1.6 MW of solar PV differently from installing 2-3kw of solar, so I expect another similar leap going from 1-2MW to ~250MW of PV. It was only 4 years ago that Google's 1.6MW planned solar installation at their headquarters (completed by mid 2007) made headlines around the country/world, and now we are clearly well along on plants that are two to three orders of magnitude larger.

One thing that comes to my mind is...why not automate the assembly of the panels into blocks (at or near the site) just as solar panel manufacturing has become automated? Each solar panel is basically the same as the next one--just as each cell in the panel is equivalent to the next, and that is very automated. A 1 MW install consists of ~4,000 panels (250kW ea), while a 250MW install consists of 1 million panels (250kW ea)...maybe the assembly is not 100% fully automated, but there will be at least a dozen "activities" in the installation porcess that could be individually, or partly automated, saving a million person iterations of the same procedure (and likely improving consistency).

At the moment this belongs to the realm of science-fiction, but one could certainly imagine a room-sized machine that travels along one row of (ground mount) panels with a fully automated sequence of actions to locate, one panel next to the prior panel, mount it (i.e. screw it in place) mechanically and then make the electrical connection and perform a quick test to ensure a minimum mechanical stability and electrical performance is achieved, and then rinse and repeat...

Tuesday, August 17, 2010

Markets are clearly not rational

Solar stocks are having a terrible year even as installations of solar go through the roof!

Most market analysts are expecing at least 50% more solar to be installed in 2010 than in 2009. Prices for panels have so far held surprising steady after last year's 40-50% plunge. Now that panel manufacturers are running at high capacity, they are making money again.

Solar stocks are almost universally down this year, some are down substantially...

Take a long time favorite of mine: SPWR (A or B, they have equal claims on equity). In the past the B shares traded at a substantial discount, now the discount is only about 5%.

Today the B shares trade around 11.10/share and the A shares at 11.70/share

On the conference call a week ago, SPWR reported that it has $737 Million in cash and investments at quarter end. That represents ~ $7.40/sh in cash and investments.

The company reiterated FY 2010 earnings guidence of ~1.50 to 1.65/share.

Take the A shares (11.70) and subtract C&I (-7.4) and you get an adjusted value of $4.30/share.

Now take the $4.30 and divide by the lower $1.50/share in earnings and I get a P/E (price/earnings) of less than 3! Less than 3! LESS THAN 3!!! and that was using the expensive shares and the lower earnings number.

Even assuming SPWR can only earning $1/share in 2010 (i.e. misses FY estimates by 1/3) this is still selling for an adjusted P/E of less than 4.5!

Also note that 16 million shares of SPWR are sold short according to Yahoo (i.e. people borrowed shares they did not own and sold them). While that is down over 1 million shares from the prior quarter, I have to ask why are short sellers sitting on shares of an industry leader (in a period of booming demand/installation activity) that is selling for less than an adjusted P/E of 3?