IDC is out with its estimated numbers, based upon 3 quarters of data, and one quarter of estimated data. The summary is amazing. 20+% CAGR for this market. It is about 9B$ (yes, that is a B meaning billion or 10^9, which is 10**9 for old timers). It is growing about 1.8B$/year at the present. Clusters are growing 60-90% per year.
And so on.
This is tremendously exciting. HPC demand is huge and getting larger. It is strong, year over year, and getting stronger. IDC admits that this demand is simply too strong to be “bounce back” from weak demand in the 2000-2003 time frame.
This means that we are seeing real growth here. There is wealth being created. There is real growth going on for smart companies/vendors. Bad ideas will always get routed, poor implementions or engineering decisions will largely be ripped apart by market forces, see Orion Multisystems for an example of this.
One would also hope that in a market growing like this, that capital would be in greater supply. Good ideas, well thought out plans targetting this market should be able to attract capital. This has not been the case in recent years, we have been told a number of times that HPC isn’t exciting. Can’t speak to that. It is a great place to build a company of value and valuable products. It is obviously generating returns to investors. The question is when will the capital market realize this and start working in this area.
IDC itself points out that lower costs had been driving earlier portions of the market, but now serious differentiators are needed for these lower cost systems. One 1U server pretty much looks like another. One blade unit looks like another though they are largely electronically incompatible. You aren’t going to get get low volume high margin sales in this market, those days are long gone, and the vendors still practicing that are either barely breaking even in their category, or augering in for rough landing.
What we see now are high volumes of supercomputing being purchased. The market was elastic to a degree, but actually something more interesting occured. As the price dropped, more customers entered the market. Not in the view of a constant dollar value market, so that the appropriate number of customers would enter the market to offset the price reductions and hold the market value the same. More customers entered the market to drive the market to higher consumption rates of lower cost systems such that the overall market grew. 20+% CAGR, for the past 3 years.
So what happens if a serious differentiated solution were to exist in the 10-25k$ segment (largest/fastest growing portion), and the 25k$-50k$ segments (second largest and fastest growing) that offered significant value to end users? Our belief is that this would push volumes rather nicely. Take a look at small to medium sized cluster installations using Infiniband networks which would qualify in this area. It appears as if these have helped drive the market forward. Now what happens if you have something which can extend the computing power of the cluster by orders of magnitude in that price range … shouldn’t equity people be all over that? Wouldn’t that be a seriously disruptive technology? What would that computing power at a low price drive the market to do?
Hopefully we will find out.