End-of-Support for Windows 2003 Yields Unexpected Market Results
I think Benjamin Franklin said it best, “the only two certainties in life are death and taxes.” This is also a precept or tenet for anyone conducting foresight analysis; nothing is ever certain and blind spots are always be lurking. Case in point: many analysts were optimistic that Microsoft’s sun-setting of support for Windows Server 2003 would be a catalyst for longer-term growth, but this simply hasn’t happened.
In fact, in our most recent press release on the server market where we measured June 2015 to August 2015 versus the prior year period, we saw 6.6 percent unit growth at approximately 125,000 unit shipments over the previous year’s approximate 117,000 unit shipments. And although there was a bit of a spike in June (pre-cutoff), growth rates for July and August were less than stellar and buoyed by pretty hefty ASP declines versus prior years. For example, July’s average selling price declined by 12.3 percent, indicating there was more supply than demand and pricing was the lever used to move inventory as opposed to the Windows Server 2003 end-of-support.
With the lack of a significant sales impact to the server market from the Windows Server 2003 end of support, we are left to industry dynamics to analyze where the market goes from here. So, what are these dynamics at play?
First, in an effort to prune capital expenses coupled with operating expenses, many firms have virtualized their infrastructure, which has left a glut of comatose or unused servers. In fact, the Uptime Institute estimates that approximately 30 percent of the server installed base is in a comatose state. For firms that have not jettisoned their unused servers, this offers them pre-existing infrastructure to leverage. For those companies that have disposed of their unused servers, many are also being reconditioned by firms in the ITAD (IT asset disposition) market and sent right back out to budget-conscious businesses.
Second, it’s tough to get an accurate picture of how many servers have been displaced by firms embracing cloud computing, but we do know for certain that this trend is happening, and it’s impacting the server market in a big way as those firms that do embrace cloud computing often forgo adopting a server.
Third, many companies have chosen to outsource portions of their operations such as call centers, analytics and application programming since the cost of labor is much cheaper in countries such as China and India. Therefore, this also contributes to the dilution of a portion of the server installed-based for U.S. companies that provide these services.
Fourth, there is a lot of vendor-direct activity from firms (e.g., StackVelocity) building customized systems for large hyperscale providers (e.g., Facebook) based on the OpenCompute standard. And this vendor-direct activity tends to bypass the indirect channel with few exceptions.
To wrap this up, let’s emphasize a few points. First, always remember Ben Franklin’s rule – not everything is certain. Second, variables such as the further adoption of cloud computing, firms virtualizing their infrastructure, and the continual outsourcing of functions to regions offering lower-cost alternatives are combining to make an impact on the server market now, and moving into the future.