In a question-and-answer session with Diane Bryant -- the individual who runs Intel's (INTC +2.91%) Data Center Group (or "DCG") -- during the company's annual investor meeting, one analyst asked Bryant about her view of the competitive landscape in the coming years. In particular, the analyst seemed to be interested in what Intel is assuming with respect to competition in its growth projections.
Her answer was both detailed and informative, and I'd like to go over what she had to say, as I think it should be of interest to just about anybody interested in Intel stock. Without further ado, let's get to it.
Who are Intel's competitors here?
According to Bryant, Intel faces competition against two major processor architectures. The first is IBM's (IBM +0.56%) POWER architecture, particularly through its OpenPOWER initiative by allowing others to build processors based on IBM's processor designs and architecture.
The second, and perhaps more talked about, threat Bryant discussed is that there are "about 16" chip companies trying to build systems around the ARM (ARMH +0.00%) instruction set architecture. Such companies include Qualcomm, Cavium Networks, and Applied Micro.
Bryant is paranoid but confident
Bryant pointed out that at ARM's most recent analyst day, the company projected that by 2020 it expects its licensees to collectively capture about 25% of the server processor market (up from prior projections of 20%). Bryant also noted that ARM believes its licensees will be able to capture around 45% of the market for networking processors (a market that Intel has around 9.5% market share in today).
"We take all competition very seriously," Bryant stressed as she addressed the analyst's question. "If [former Intel CEO] Andy Grove taught me one thing it's 'be paranoid' and [Intel Chairman] Andy Bryant reinforces it every day."
She addressed ARM's networking claim with a simple but interesting argument, claiming that four of the company's biggest customers (AT&T, South Korea Telecom, Verizon, and Vodafone) are already committed to transitioning to Intel Architecture (or, more colloquially, x86). Bryant claims this is due to the fact that x86 is the "only architecture that is virtualized today" and that there is no virtualized ARM solution today.
As far as the networking chip space goes, Bryant said Intel will ultimately "be in battle with [the ARM vendors] over the years" but ultimately seemed confident, particularly due to the "vote of confidence" from the aforementioned four large customers.
Moving on to the server market, where Intel has overwhelmingly dominant share, Bryant said Intel actually went out and counted the number of ARM-based servers deployed today to try to estimate ARM's current market share. This figure, she noted, ultimately "less than one half of one tenth of a percent."
"It's a high bar, it's hard to get in," Bryant said. "Where we see it -- and I don't want to underestimate it -- is a tool for [Intel's] customers to use in pricing negotiations. They want us to know that there's always an option for them if they want to port their code and make that investment, they could do it."
How should investors view this?
I believe that part of the reason ARM's market share in the server market is so low is simply due to the fact that none of the parts that have been fielded from ARM's various licensees have been particularly interesting/compelling thus far.
I do think that parts from competitors -- the ones that are left standing, anyway, as the investments required to be successful here are quite large and only getting larger -- will become more interesting over the next several years if their public product plans are to be believed.
That being said, Intel isn't standing still either and its products, too, should also get better. Indeed, if we take a look at this slide from Intel CFO Stacy Smith's presentation, we see that the area where the company is making its largest incremental investment is in DCG:
Source: Intel.
It's clear Intel is willing to spend whatever it feels it needs to in order to try to maintain its dominant positioning in the market.
Whether the ARM vendors (or members of the IBM OpenPOWER alliance, but Intel seems more focused on the ARM vendors than on IBM) are ultimately able to deliver parts that are fundamentally competitive from a total cost of ownership perspective remains to be seen. That said, given Intel's strong execution and very heavy investments in this market, I tend to agree with the notion that it will be quite hard for competitors -- particularly smaller ones -- to gain material share in servers anytime soon.
