Intel (INTC 4.02%) essentially owns the server chip market. Over the past few years, Intel has pushed Advanced Micro Devices almost entirely out of the market, and the complete lack of competition has allowed the company to extract enormous profits. While the PC market still accounts for the bulk of Intel's revenue and profit, the server market is growing faster and carries higher margins for Intel.
Competition is coming
The days of no competition in the server market are coming to an end, however. Various companies, including AMD and Qualcomm, are pushing ARM-based server processors as an alternative to Intel's products, particularly at the low end of the market. At the high end of the market, IBM has joined forces with companies like NVIDIA and Google, opening its Power architecture in an attempt to build a competing ecosystem.
Intel does have some significant advantages that will make it difficult for either ARM or Power to gain much traction. Unlike most of its competitors, which rely on external foundries to produce their chips, Intel manufactures its own products, and it currently has a meaningful technological lead over the foundries. Intel's x86 architecture is also the standard in both the PC and server markets, meaning that software needs to be optimized for other platforms, introducing switching costs in the process.
There's certainly an incentive for the big web companies, like Google and Microsoft, to encourage alternatives, given that both companies spend enormous sums of money each year on Intel chips as they build out their cloud data centers. Google is a member of the OpenPOWER Foundation, and Microsoft is reportedly working on an ARM version of Windows Server, software that powers roughly a third of all web servers.
The real problem for Intel
I doubt either ARM or Power will gain all that much market share in the short term, given the strength of Intel's products, but the real threat for Intel is not losing market share, but being forced to cut prices to stave off the competition. Intel's data center operating margin was around 50% during the third quarter of this year, compared to just 35% in fiscal 2009.
Over the past 12 months, the data center group has generated an operating profit of $6.5 billion, about 45% of the company's total, although this percentage is skewed by the multibillion-dollar losses associated with Intel's mobile business. Regardless, server chips represent a significant chunk of Intel's profits, and any decline in profitability would have a severe negative effect on the bottom line. Based on data center revenue over the past 12 months, here's what would happen if the data center operating margin declined.
|
Data Center Operating Margin |
Data Center Operating Profit |
|---|---|
|
48.9% (TTM value) |
$6.51 billion |
|
45% |
$5.98 billion |
|
40% |
$5.32 billion |
|
35% |
$4.65 billion |
How much pressure Intel will actually face from its newfound competition remains to be seen. It's certainly possible Intel will manage to maintain both its market share and its margins, given the advantages it enjoys. But Intel's sky-high margins gives competitors an awful lot of room to offer a better value, and it's not unreasonable to think Intel may not be able to maintain 50% operating margins in its data center segment in the long run.
If the operating margin fell to 2009 levels, nearly $2 billion would be shaved off of Intel's total operating profit. That's an extreme drop, but even a more modest drop could lead to declining segment profits, even if Intel still manages to grow revenue. This is the problem with being so utterly dominant: It's hard to imagine margins going much higher, and any decline would be terrible news for the bottom line. Even if Intel remains the dominant market leader in the data center, which it likely will, its profitability may never be the same.
