Addison Snell at Tabor Research published an interesting read last week at the HPCwire blog.
With the economy beginning to crash down around us, Tabor Research issued a press release in October stating that the majority of users expected to increase their spending on HPC, with most of those expecting a significant increase. Three months later, we stand by that insight.
They aren’t the only ones saying that HPC is a lone bright spot in the otherwise dreary IT market. IT analysis firm Waters said in early January that HPC in finance would be strong in the near future. Another major, soon-to-be-released study predicts that while HPC sales may fall in the short run (a result of deferrals), they will still remain ahead of general IT spending and recover faster. Tabor predicts essentially the same thing.
2009 could therefore see a total decrease in HPC spending, not because users’ budgets are contracting, but mostly due to purchases being placed on hold for a time before they are finally approved. Occasionally, all purchases, whether budgeted or not, are suspended at an organizational level until a future date. Deals slip from this quarter to next quarter, and some eventually topple from 2009 to 2010. The good news is, when the economy eventually turns around (yes, please), sales cycles will shorten again, and we’ll see a positive blip as deals pull back in.
I certainly think this is true in supercomputing, where we have thick layer of insulation between us and both good and bad times. The picture that Tabor Research and other are painting — longer sales cycles, but more delays than outright cancellation — seems reasonable to me.
For my money the companies that focus exclusively on HPC (SGI and Cray are the high end examples) have a “life on the hairy edge” business model, and we have to be worried about these guys not being able to weather any storm, no matter how short. Based on recent debt buy backs and insider stock buys, Cray is betting it can. SGI hasn’t made any such public bets.
Sun, while far from a pure play HPC company, is hurting and may be bowled over by the dip in general IT. If I had to guess, I’d say HP and IBM will get over the hump, although it isn’t clear yet that HPC as a business within these companies will survive. If it doesn’t it will be a religious decision rather than any kind of strong indicator about the viability or otherwise of HPC in this economy — I cannot imagine that HPC drives the bottom line or product decisions at either company in any significant way. Both are about 100 times bigger (in market cap) than the entire HPC market.
What about the Penguins, Rackables, Veraris, and Bulls of the world? They’ll have a harder time mixing it up over cost in the middle market with a less differentiated product or, worse, driving down to the bottom to compete solely over price. Some of these may cease to be, unless the recovery is very swift indeed.