Intel (NASDAQ:INTC) is the dominant vendor of the microprocessors that power the world's data center servers. In 2017, Intel's data center group raked in $19.1 billion in sales, $17.44 billion of which came from sales of its Xeon processors and accompanying platform controller hub (PCH) chips.
Since the company's last financial analyst day, which took place in February 2017, the company has indicated that its data center growth will increasingly be fueled by sales of so-called "adjacent" components.
For some context, data center servers incorporate a lot of other technologies beyond Intel's processors -- they include memory, storage, network connectivity, and other components. The idea here is that if Intel can broaden the portfolio of products that it offers into this market, it can accelerate its revenue growth and, potentially, widen the moat around its main processor product lines.
Nevertheless, while Intel's strategy to broaden its data center product portfolio is reasonable, the company's messaging seems a bit too focused on everything but its core processor products. Here's why that's a red flag to me.
Trying to hide share loss?
Shortly before his ousting on June 21, former Intel CEO Brian Krzanich apparently told analysts with stock market research company Instinet that the company expected to lose market segment share to rival Advanced Micro Devices during the second half of 2018.
"Mr. Krzanich was very matter-of-fact in saying that Intel would lose server share to AMD in the second half of the year," the report read.
What this looks like to me, then, is that Intel is trying to redefine its total addressable market in the data center to include a wide range of other products and technologies so that it can tell its investors that it has a big opportunity for overall data center product share gains even as it loses share in its core CPU business.
In fact, take a look at the official statement that Intel provided me when I asked for comment around Krzanich's remarks to the Instinet analysts (emphasis added):
We see significant opportunities for growth in the data center -- an estimated $70 billion market opportunity by 2021 where we have an opportunity to grow our total silicon datacenter market segment share from where we are today. To win, we will continue our history of CPU leadership and deliver the broadest portfolio of products that, when combined, change the basis of competition in the data center.
While we are prepared for a more competitive environment as we move through 2018, we've already factored that into our financial forecast and we're in a great position to compete. We remain very confident in our products, our roadmap and our competitive position. For example, Intel Xeon processor Scalable family represents the biggest advancements in platform capabilities in a decade, and later this year we'll introduce breakthrough new Intel Optane DC persistent memory and storage technology architected specifically for the data center.
While Intel does make a token mention of its CPU competitiveness, claiming that it will "continue [its] history of CPU leadership," the focus of the statement is clearly around everything but its core CPU offerings.
This seems strange considering that less than 10% of Intel's data center group revenue in 2017 came from these non-CPU areas.
Ultimately, I think part of the reason that Intel's stock is down significantly from its peak despite continued strong financial performance is that investors are becoming increasingly worried that the chip giant's core data center CPU business is on the verge of being disrupted.
To make matters worse, Intel has been, at best, cagey on its future CPU plans, only offering vague statements about its "confidence" in its product strategy, with little to nothing in the way of specific plans.
Until Intel can provide some additional clarity on how it can defend its position in the data center processor market and give investors reason to be confident that a serious disruption isn't imminent, I think the upside to the shares will be limited and the downside could be fairly significant.