Intel Riding High in Data Center Market as AMD Threat Looms
via Shutterstock
In the choppy seas of public and investor opinion on which Intel currently sails, blustery and changeable winds are both besetting or boosting the company – though, to be sure, the bulk of them are headwinds.
An article in yesterday’s Wall Street Journal declared that AMD is “set to crack Intel’s lock on data centers” and “poised to take market share” with its EPYC chips. “Far bigger rival Intel has long accounted for about 99 percent of server processors sold. That will almost certainly change in the coming year…”
The data center market is in hyper-growth mode driven in significant degree by the FANG tech giants, with Intel garnering more than $20 billion from the sector over the past year. This has counterbalanced flatness in Intel’s PC processor product line. But AMD has an opportunity in the data center business because of Intel’s challenges in moving to a new 10 nanometer chip manufacturing process from 14nm, available since 2014. AMD expects to deliver its next-gen chips later this year while Intel’s 10nm chip, code named Ice Lake, has been delayed until 2020, as reported last week.
One analyst, Handel Jones of International Business Strategies, estimates AMD’s lead in the nanometer race could cut Intel’s data center processor market share by up to 5 percent.
Even Intel doesn’t believe its near total ownership of the data center market can last. The company conceded two months ago that AMD will take market share -- Intel's goal is to limit AMD’s share to the 10 percent range.
These factors have major stock impacts – Intel stock, which over the past 12 months peaked at $57 on June 1, is at $47 today. AMD stock, which hit a low over the past year of $9.50 in early April, is now at nearly $20.
And of course, this Intel vs. AMD discussion leaves out the competitive force of Nvidia and its industry leading GPU processors, which power AI, machine learning and deep learning applications, the strategic future of high end data center server processing.
Nor does it take into account recently identified Intel chip vulnerabilities, malware called L1 Terminal Fault, which we covered earlier this week, similar to the Meltdown and Spectre malware revealed late last year.
OK, that’s the bad news.
On the other hand and over the long term (Intel is celebrating its 50th year) it’s been a bad idea to bet against Intel. The company has weathered major challenges from AMD over the past two and three decades – the challenger launched server processors early in the 2000s and within six years had won about 25 percent of the server processor market. But Intel successfully beat back AMD due in part to its tremendous financial reserves that enable it to acquire key technologies and to fund enormous R&D activity.
Putting Intel’s powers of resilience in perspective, last year the company spent $13.1 billion on R&D company-wide. Total AMD revenue for 2017: $5.33 billion.
And a recent post on the Motley Fool site from Ashraf Eassa points up Intel’s current, explosive data center revenue growth, with revenue last quarter hitting $5.5 billion, a 27 percent YoY jump, and operating income of $2.74 billion, “an eye-popping 64.8 percent surge from a year ago,” Eassa said.
This growth reflects the fast rising tide of the high end data center market, avid for Intel’s Xeon chips for data center servers. Intel said that sales by its Data Center Group to cloud service providers, enterprise and government, and communications service providers jumped more than 40 percent, 10 percent and 30 percent, respectively. “In recent years, the company's enterprise and government sub-segment has struggled, declining at a 4 percent compound annual rate from 2014 to 2017, according to Intel's Rajeeb Hazra, so the 10 percent growth last quarter (accelerating from 3 percent growth in the first quarter of 2018) was a sharp improvement.”
Another Motley Fool article, this one by Harsh Chaurhan, spotlights Intel’s growing success with AI-enabling FPGA processors and its looming threat to FPGA market leader Xilinx, which currently controls almost 60 percent of the market, Intel controlling most of the rest.
Intel's Programmable Solutions Group revenue jumped nearly 20 percent YoY in the second quarter to $517 million, and Intel said PSG revenues have grown by more than 100 percent two quarters in a row, a faster growth rate than Xilinx’s. “This seems a bit surprising as Xilinx reportedly enjoys an 18-month technology lead over Intel in FPGAs,” wrote Chaurhan, “but the latter has been outgrowing its nemesis of late.”
The reason, Chaurhan said, is due to Intel’s recent acquisition of semiconductor manufacturer eASIC, maker of customizable chips for cloud computing and other wireless environments.
“This move is expected to help Intel address a major problem with FPGAs,” said Chaurhan. “The fact that FPGAs can be customized to perform specific tasks has boosted their popularity as they are more efficient at executing AI applications. But they need to be transformed into ASICs once an application doesn't need customization… once an application has stopped changing, it has to be run using chips that are more structured in nature to deliver better performance and consume lower power. ASICs are well-equipped to handle the post FPGA stage because of their power efficiency and ability to be configured quickly.”
The result: Intel can deliver “an end-to-end solution to customers that begin with FPGAs and finally use ASICs to tackle data-heavy applications in the cloud or aid the deployment of faster wireless networks.”
Intel’s potential success in the FPGA market, and therefore in a portion of the exploding AI market, is a demonstration of a company diversifying beyond its old x86 warhorse and leveraging new technologies that address strategic markets. The question is whether the company will make the right moves, the right acquisitions and focus its R&D on the right market opportunities. It’s impossible to say if that will happen, but it’s accurate to say that until now, the company nearly always has.
