Google, IBM, ARM And AMD Begin The Assault On Intel's Server Franchise

| About: Intel Corporation (INTC)


Intel’s Data Center Group (server chips) is growing at 15%/yr. and provides more than 50% of Intel’s profit.

Google recently announced it is developing apps for both IBM Power and ARM chips.

AMD’s Zen chip will compete with Intel’s Windows x86 franchise.


Intel's (NASDAQ:INTC) stock hit a 12-month high of $35.44 on Dec. 29, 2015. Since then, it has been range bound between about $28 and $32. The reason for this is the sluggish sales of desktops (CCG - Client Computing Group), integrating the Altera acquisition, the miniscule or non-existent sales of 3D NAND, and the lack of growth in the highly touted IoT Products Group. Offsetting all this bad news is the robust sales of the DCG (Data Center Group) as enterprise users gobble up millions of server chips as they try to keep up with the ever increasing demand for cloud storage and data delivery.

The biggest problem is the DCG's virtual monopoly (99% market share) on server chips has come into the view of Google (NASDAQ:GOOG) (NASDAQ:GOOGL), IBM (NYSE:IBM), ARM (NASDAQ:ARMH) and AMD (NASDAQ:AMD), all of whom would like to see that 99% number come down substantially. This is the continuation of a theme I outlined here "Intel Versus Qualcomm: Who's The Chip Champ Going Forward?" here "More Woes For Intel: Chip, Chip, Chipping Away At Chipzilla," and here "AMD: Zen's Release Will Be Taking Names And Kicking Bandwidth Sooner Rather Than Later."

In this article, I will examine whether Intel can respond to the challenges it faces in the all-important server chip markets.

"No problem can withstand the assault of sustained thinking." Voltaire

Google has announced it is writing apps for the IBM Power series and ARM server chips.

"The Power architecture is now fully supported across our toolchain," said Maire Mahony, a hardware engineering manager at Google and director of the OpenPower Foundation.

This is an important turning point for non-Intel chip manufacturers because whatever Google does is likely to also be done by all the other server chip devouring companies like Facebook (NASDAQ:FB), Amazon (NASDAQ:AMZN), cloud servers and maybe even Microsoft (NASDAQ:MSFT) which uses ARM chips in its phones.

Another reason why this development might be worrisome to Intel is the ever increasing use of Linux on the enterprise. Linux is the operating system of choice for cloud servers such as Google, Amazon and Facebook, and Linux does not need Intel's x86 instruction set required by Windows servers. Keep in mind, Google's Android system is Linux based.

Here's an article by titled "Google taps IBM, Rackspace to Dent Intel's Hold on Server chips." The title pretty much describes Intel's problem. This was the most interesting comment:

"Every line of software code that Google writes is now tested to make sure it can run on OpenPower or ARM-based chips".

That sure sounds like a change is coming.

Intel's response will be to bundle its Xeon chips with Altera's FPGA (Field Programmable Gate Arrays) hopefully giving them a performance advantage over the newcomers. The problem is IBM and ARM can add FPGA chips too.

And Intel is not home free in the Windows server market either as AMD plans to release its new Zen chip later this year and a server version in 2017. Intel has for all practical purposes a 100% market share in Windows servers, and while AMD is too small to have a big effect, it may require Intel to lower prices (and subsequently margins) to maintain a 90% plus share. This is important as you can see from the following graphic that shows operating cash flow (in billion dollars) shrinking over the last several years.

The problem is Intel doesn't have anything to replace server chip sales and profits.

One of the mistakes investors make concerning Intel is forgetting how big it is. It is a behemoth with $55 billion in sales in 2015 and $49 billion of that being CCG and DCG. That means if CCG continues to falter and/or DCG finally faces some real competition, there is nothing on the bench to replace the profits of the two superstars. The software group and IoT combined had less than 5% of CCG and DCG's profit, so a little downturn in the big two will overwhelm any increase in the two little guys.

To give you an idea how daunting this task is, I have done a spreadsheet showing 3D NAND and 3D Xpoint both surging to 10% of sales or $5.5 billion each. That is ignoring the probability that 3D NAND will replace virtually all of the $2.5 billion in 2D NAND sales Intel had in 2015. Assuming also that profit margins are the same as the semi-monopoly x86 products, another unlikely possibility in my estimation, EPS grows to about $3 per share. The question then becomes which year will it be that Intel's 3D products generate $11 billion in sales? Or to put it another way when do Intel's 3D products grow from zero to almost the size of Texas Instruments (NYSE:TXN) ($13 billion in sales)? When that happens, Intel will make $3 per share. Is that 2017, 2018 or what year?

And if you believe that big sales are on the way for Intel's 3D products, why not invest in 50% Joint Venture partner Micron (NASDAQ:MU) instead? After all, Micron already sells more NAND than Intel, and if the supernova Texas Instruments size sales are coming for 3D, then Micron, at less than $11 per share, is a much better investment choice because the effect on its sales and earnings would be spectacular at about $4.60 per share. I think Micron is the choice as I outlined here "Micron: At Under $10 A Share Even I Would Be A Buyer," and here "Sometimes An Ugly Duckling Turns Into A Beautiful Swan - Buy Micron Under $11."

Of course, adding those two aggressive sales targets together and you are looking at going from zero to almost Qualcomm's size ($25 billion). That certainly gives pause when you think about it that way.


Intel's huge cash flow depends upon the continued growth of its two stalwarts, desktop and server chips. The semi-monopoly status of these dominant groups generates huge margins that are not possible in any of their other groups. Therefore, if either of these two falters, there is nowhere to go for help. None of the other groups have anywhere near the sales let alone the margins to make up for anything but minor hiccups, if that. True, there is 3D NAND and 3D Xpoint coming down the road and they may be able to supply some ammunition to the beleaguered Big 2, but high volume sales and high productive margins are somewhere down the road. Maybe late 2017, 3D NAND can help and maybe in 2018, 3D Xpoint will kick in some big profits. But until then, things look meager indeed.

Intel is at best a hold, though its dividend does provide some downside protection. I would remain on the sidelines until I can see a more certain source of sales and especially profits falling to the bottom line. Until then, it is a wait and see. Micron is a buy here because of its low price which is currently below book value. I am long Micron.

Disclosure: I am/we are long MU.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.