Send me real-time posts from this site at my email
Alex Cho

Intel's Position in Servers Remains Strong Despite Competition

Analysts continue to back Intel going into Q1’16 earnings despite the recent departure of executives and new entrants into the server CPU space. For the most part, I agree with analysts on the insurmountable moat that Intel has built for itself with x86 architecture.

Here’s the latest commentary from Credit Suisse:

Our TP of $40.00 represents ~17x our CY16 EPS estimate of $2.35 but ~10x what we view as the Company's medium-term earnings power of $4.00+. On INTC, our thesis remains intact as the Company begins to show operating leverage from prior investments, a scarcity value as most large cap tech companies have been choking investment in favor of financial engineering. In addition, INTC remains well levered to Data Growth not only in DCG but also in its IoT and NVM businesses. We believe a structural re-rating of the Company is warranted given our view that the Company's business model/IP has significant barriers to entry (evidenced by one of the highest "R&D + CapEx to Revenue" ratio), and we continue to argue that Moore’s Law has been the cornerstone of Tech economics for 40 years and that INTC will be the Last-man-standing on Moore’s Law, leading to market share gains in Compute (including Mobile), as well as Foundry.

For the most part, they make some pretty compelling points, as Intel has demonstrated continued leadership in foundry technologies. While IBM has mentioned that they’re on track to complete 7nm before Intel, the company’s patchy track record plus lack of confirmed wafer equipment investment raises a lot of questions as to whether IBM can rebuild its datacenter platform through its own hardware.

Bloomberg recently reported that Google is accelerating development of software that can run on IBM CPU architecture:

Google, the main division of Alphabet Inc., said Wednesday it has built software that lets it use OpenPower chips from International Business Machine Corp. to run some of the computer servers in its many massive data centers. Google also said it’s developing a data center server with cloud-computing company Rackspace Hosting Inc. that runs on a new IBM OpenPower chip called Power9, rather than Intel processors that go into most servers. The final design will be given away through Facebook Inc.’s Open Compute Project, so other companies can build their data center servers this way, too.

While Google Cloud Platform is expected to grow rapidly, it’s highly unlikely that Google will get significant buy-in from its customers if workloads were centered around IBM’s OpenPower chip. The main advantage to X86 architecture is the openness of the ecosystem, as both Linux and Windows Server kernels operate efficiently with very minimal middleware. Furthermore, Intel has tackled optimization at both the hardware design and lower software layers to maximize its TCO advantage. I could go on about this for hours, but basically it’s Google’s desperate attempt at negotiating lower prices as Intel continues to build out the rest of its platform with a fully bundled package of hardware accelerators, FPGAs, storage, silicon photonics and Xeon-series CPUs.

Welcome! Is it your First time here?

What are you looking for? Select your points of interest to improve your first-time experience:

Apply & Continue