Solarworld has on the face of it, just made a very innovative, shrewd move.

They have bought up a failing japanese silicon chip factory in Oregon, USA, for $40 million. The factory - or should we say Fab - had already had investment dollars haemorraged into it by Japan's Komatsu-Group; $794 million until the crash in chip prices made it financially unviable.

The deflation though in semiconductor prices has not been reflected in the solar pv market, where supply clearly cannot match demand. So Solarworld are going to convert the factory to a silicon wafer and cell production unit at a cost of $397 million. The future output of this converted solar plant has been put at 500 megawatts per year.

This reminds me of a conversation I had many years ago with a very successful entrepreneur - of all things in high-end bespoke wooden furniture - who shared with me his plans to buy a restaurant to entertain his clients and friends in. The key to owning a successful restaurant he said, was not to set it up yourself. Ideally, you buy a 2 year old restaurant that had just about failed. That way, someone else had haemorraged all the investment money on location, fittings, marketing, staff etc.

Just when they are at the point of collapse - because restaurants have such a high failure rate - this was the time to move in, buy it up for a song and get into the business. Apparently, all you really needed to then do was to invest in a good chef !

So there you go - like restaurants, semiconductor fabs are very, very expensive to build and the world is awash with them. Buying one off the shelf and converting it for your own purposes is potentially extremely clever. I suspect the rest of the solar world will be watching this experiment very closely and may already have started scouring redundant semiconductor fabs in Europe, Asia and the USA.