Turbine Availability

I don’t get it. There are hundreds of turbines on the ground in North America, awaiting permits to be erected. Every wind publication you open talks about a new factory being announced or opened in Pennsylvania, Iowa, Fort Erie, the Gaspe, Oklahoma, or Colorado and expansions of factories are routine. The production capacity of the wind industry must surely be expanding at a rapid rate. The largest markets, such as Germany or Spain, are maturing, and will be lucky to install as much wind this year as they did last year.
But wind turbine lead times are at an all time high. Most vendors are sold out through 2008, with only dribs and drabs available. The contract terms in turbine supply agreements are increasingly favouring the vendors, with less warranty, earlier and larger pre-payment requirements, and higher prices.

Last year, the wind industry worldwide installed 15,017 MW, a 30% gain on the year before. To give perspective, Ontario has a total installed capacity of 30,000 MW of all types of generation, so 15,000 MW is a pretty big number. Based on the number of factory expansions announced, I believe the capacity of the wind industry will expand by at least 25% this year, or about 4000 MW, to over 19,000 MW.

Where are all these turbines going? In Canada alone, there are several hundred MW of wind turbines in lay down yards waiting to be installed. It would be hard to believe the firms that own them would be ordering more turbines. The US has several firms with a similar situation. And there are several large framework agreements for turbines placed by some of the large developers, giving them a pipeline of supply, and in some cases not enough places to put them. Why is there a shortage of supply when capacity is expanding, and there are all sorts of signs that the capacity of the manufacturers exceeds the capacity of developers to find homes for turbines?
I used to be in the computer business, and I remember well the cycles in the memory market. Memory chips follow Moore’s law, which says that the number of transistors on an integrated circuit doubles every 24 months, at the same cost. Or the cost falls in half for the same number of transistors. So the long term price trend for computer memory is down. But despite this underlying cost trend, the spot memory market used to rise and fall dramatically. In a rising market, inventory was valuable, and you could sometimes raise the price on inventory on hand. But in a falling market, inventory was bad, and sometimes you sold at a loss, just so you could buy the new lower cost memory chips. It was a risky business. We tried to turn inventory 3 times a month, which is almost as fast as inventory turns for perishable food in the produce section of the grocery store.

The long term cost trend for wind turbines is also down. Some studies suggest that the long term cost per Kilowatt hour should drop 3-4%/year, or 15-20% every time the industry doubles in size. With a 30% annual growth rate, the wind industry doubles in size every 2 1/2 years. Growth drives innovation and economies of scale. In short, the more we build, the better we get at it. Against this underlying cost trend the industry has faced higher commodity prices, especially steel and copper. And of course, prices and costs don’t always have a direct relationship.

The memory market, like all markets, was driven by supply and demand. When Samsung built new fabrication plants, and tried to buy some market share from Toshiba or NEC, prices would soften. In the early years, IBM drove the market. If they had a new model they were about to introduce, they would accumulate inventory, making the spot price of memory very high. But if sales softened, they would cancel orders, and the market would be flooded with product. The equivalent in the wind business is the framework agreements signed by some of the large developers, and even some financial institutions like Goldman Sach’s. In 1987, IBM introduced a new architecture called Microchannel, to replace EISA. It was unclear which of these two architectures would dominate. And so for 9 months, corporate procurement froze. The memory market collapsed, and consumers had some tremendous buys. In 1995, Microsoft finally got its Windows software right, and introduced Windows 95, a big hit. Windows 95 needed lots of memory. The market soared.

In the wind business, in 2002, the US Production Tax Credit expired, and the US market for wind turbines collapsed. Installation of wind in the US dropped from 1700 MW to 400 MW. It became a buyers market overnight. Wind manufacturers faced tough times, and several had to merge, of go out of business.

Many people in the wind business today have not seen the down cycles that can occur in markets. But it will come. It may be the expiry of the PTC in the US, which is scheduled to run out at the end of 2008. It may be that the rush to install new manufacturing capacity may finally catch up with demand. It may be a temporary collapse in natural gas prices in North America or Europe due to a warm winter. Lower natural gas prices often make electricity prices drop as well. Markets never go straight up. And the wind market is no exception.

But these are difficult times for developers trying to buy turbines. My advice to them is be patient. The market will turn – it just may take awhile. Meanwhile, a window of opportunity has opened for 2nd tier vendors, which will encourage innovation and competition in the market in the future. And my advice to turbine vendors is never lose sight of your customers. Difficult decisions are being made now on allocation between customers. And while you can sell all you can produce today, your customer relationship will be needed more than ever when the market turns down. And it inevitably will.

Leave a Reply