Last quarter, Intel's (NASDAQ:INTC) data-center group (DCG), which sells processors and other components that power most of the world's data centers, saw revenue grow 7% year over year. That growth was driven by a roughly 7% sales increase in platforms as well as an approximately 16% increase in sales of non-processor/chipset components -- what the company refers to as "adjacencies."
Though a 7% increase in sales isn't awful, it's below Intel's "double-digit" long-term revenue growth target for the business. Moreover, the relatively anemic growth comes hot on the heels of the introduction of Intel's new Xeon Scalable processor family, which the company touted as "the industry's biggest platform advancement in a decade."
Analyst Stacy Rasgon pointed out during the company's earnings conference call that despite the launch of these new Xeon Scalable chips, Intel's DCG platform revenue didn't accelerate. The analyst, considering this, asked management to explain why the new platform isn't driving such an acceleration.
Here's what Intel CEO Brian Krzanich had to say.
Early innings of the product ramp-up
Krzanich said it's still early in the production ramp-up of this new platform. The executive went on to explain that different data-center customers should adopt the new platform at different times. The first customers for the new platform are the major cloud service providers, commonly referred to as the tier-1 providers. After that, the smaller providers, which Krzanich referred to as "tier-2," will hop on board.
After that, the executive indicated, communications service providers -- telecoms -- and then finally enterprise data-center customers will migrate to the Xeon Scalable chips.
"This is going to take a year or more," Krzanich said.
Some additional data
Later during the conference call, Intel CFO Robert Swan told another analyst that the company is still seeing a sales transition from data-center products based on its Haswell architecture, which launched in the second half of 2014, to products based on Broadwell, which launched in the first half of 2016.
Haswell was Intel's second-generation processor family manufactured on the company's older 22nm manufacturing technology, and Broadwell is the company's first-generation 14nm product. The new Xeon Scalable processors, on the other hand, are based on the company's Skylake Server architecture and are built using the company's second-generation 14nm technology, known officially as 14nm+.
It's clear, then, that it takes time for data-center customers to transition to new Intel platforms.
Will Intel see a sales inflection later?
What seems to be implied in this discussion is that as more customers transition to the Xeon Scalable chips, Intel should see increases in average selling prices. Swan indicated that the ongoing transition from Haswell to Broadwell and from Broadwell to Xeon Scalable do have "positive [average selling price] implications."
However, considering that these transitions seem to be continuous in nature, in that customers are constantly in the process of moving to new platforms, I'm not sure if stockholders should get too excited about a processor-related inflection in revenue growth over the course of the next year.
What I think will be more likely to drive an acceleration in Intel's DCG revenue will be the launch and subsequent ramp-up in shipments of the company's upcoming 3D XPoint-based memory modules next year. These products will represent the entry into a new product category for Intel and, if successful, could accelerate Intel's revenue growth.
In addition, since these 3D XPoint-based memory modules will require Intel's upcoming Cascade Lake processors -- which are updated, higher-performance versions of the current Xeon Scalable parts -- Intel could see an acceleration of the adoption of its newest platform next year.