World Server Forecast – Revenue Will Decline 26% In 2009 With No Growth Til H2 2010

Having spent 23 of the last 26 years researching the server market (even if we did call them ‘Multi-User Computers’ in the 1980s), I think it’s about time to take look at this ITC offering and suggest a forecast for the next few years. As with all of my forecasts I’m looking solely at vendor revenues. Adding channel margins will make the total value of the market higher than I show here. Let me know if you’re interested in discussing the methodology.

The World Server Market Will Decline By 26% In 2009

The world server market has suffered an unprecedented fall in value in the last 2 quarters (-15% in Q408 and –23% in Q109). A view of the development of revenues is shown in Figure 1.

Servers are only used by business and the credit crunch has been the major cause of the slump. However server and other ITC suppliers have added to the problems by pursuing strategies that have a negative affect on these products. In particular:

  • Dis-aggregation– the move towards industry standard operating systems (mainly Microsoft Windows, but also Linux from Novell and Red Hat, as well as Sun Solaris), third party maintenance, third party infrastructure software and third party hardware in the networking and storage systems areas have restricted the amount of money being paid on server hardware. Less ‘Aggregated systems’ such as IBM’s zSeries and pSeries, HP’s HP9000 and Non-Stop Series, Sun’s Sparc are also being disaggregated, as users demand better integration and cheaper prices.
  • Consolidation – Virtualisation, Green IT, Dynamic Data Centres and Cloud Computing projects all begin with physical consolidation. Successfully increasing server utilisation from – say – 15% to 80% will inevitably reduce the number of servers used by a large organisation. It also leads to money being spent on integrators rather than server suppliers – even if they work for the same company.
  • Outsourcing– vendors typically count the revenue from outsourced hardware as IT Services rather than server hardware, although IBM attempts to measure its ‘Blue dollars’ (revenue from the service provision of zSeries machines) in terms of hardware spend.

Servers will continue to diminish as a part of the ITC market until their suppliers find new roles for them. Certainly the massive increase in digital content and automation will require more servers, as will the move to virtual PCs. Personally I believe there will come a time when consumers begin to purchase ‘home servers’ to handle security, media streaming and data management. However it’s likely that such machines will be named something other than ‘servers’.
I expect the server market to fare worse than most other ITC offerings in coming quarters. Click on these highlighted links for my forecasts for the over all ITC market worldwide and in EMEA, telecoms, PCs, disk drives and software. A picture of the comparison of some of these is shown in Figure 2.

The good news is that the server market is likely to reach positive growth ahead of software; however I don’t expect it to see positive growth until the second half of 2010.

IBM And HP Battle It Out For Server Market Share

Unlike IT Service, Telecom service, PC and Software areas, the server market is owned by a small number of suppliers. I’ve given my estimate of market share for the leading vendors in Figure 3, calculating revenue from company financial results for 2008. In total vendors received $54.4 billion from users for server hardware – which was small in comparison with PCs at $450 billion, but still a major component of the over all ITC market.

If we consider the development of each major vendor’s revenues from servers over a lengthy period, we can see that IBM’s lead over HP has been narrowing. Sun, NCR, Fujitsu and Hitachi have failed to grow their business strongly. Dell posted strong growth for servers for all quarters leading to Q408, when (like all the others) it started to suffer steep declines. A picture of the development of server hardware revenues is shown in Figure 4. I’ve added the 4 trailing quarters in each period.

Conclusions – Users Should Expect Server Suppliers To Be More Flexible

I don’t want to be too depressing about the prospects for server vendors. After all the managers running these businesses have proved in the past to be adept at managing in a downturn. There are a number of positive strategies I expect them to pursue. In particular:

  • Cloud Computing (again) – in the last recession vendors proposed a number of flexible offerings to help users. Does anyone else remember ‘on-off-upgrade-on-demand’ and other similar? This time there’s a stronger requirement for users to move to a subscription (as oppose to outright purchase, or lease) model. While most of the focus is currently on the Software market, Cloud Computing can offer compute-by-hour services and help vendors utilise unsold hardware in the bargain.
  • Get Competitive – downturns are a great time to upset traditional supplier relationships. Vendors are already spending money on competitive swap out deals and defensive marketing. I expect these activities to become even stronger in the next 6 months.
  • Bundle Solutions– while the server hardware market has suffered from dis-aggregation, many users are confused by the large number of relationships they have to handle with the resultant group of suppliers. By re-aggregating solutions for Small and Medium Businesses (SMBs) vendors can help and make money.

Of course all of the bad news for suppliers has a positive effect of making servers cheaper for users. If you’re finding it difficult to finance your next major server purchase you should talk to your supplier. You may find that market conditions are making them increasingly flexible.