Earlier this month, chip giant Intel (INTC -0.23%) announced that it wouldn't begin mass-producing chips using its 10-nanometer chip manufacturing technology until sometime in 2019, a slip from the company's previous plan of mass production in the second half of 2018.
This technology, Intel has said, is supposed to deliver improvements in the performance and power consumption of its chips, compared to its currently-in-production 14nm technology.
Image source: Intel.
At a recent investor conference, analyst Harlan Sur asked Intel chief engineering officer Murthy Renduchintala about the company's ability to defend its "90%-plus market share" in the data center chip market, in light of the delay in getting 10nm into mass production.
Let's go over what Renduchintala had to say.
Renduchintala is "excited" about the future
"Clearly we haven't disclosed our 10nm server road map in any detail yet, but quite frankly, I'm very excited by the product pipeline that we have going forward in our server road map," he said.
Renduchintala also expects that products built using Intel's older 14nm technology "will be more than capable" of allowing the company to achieve its business goals while it works to get its 10nm technology into good enough shape for mass production.
Is his optimism warranted?
Later this year -- likely in the fourth quarter -- Intel is expected to introduce its next-generation family of data center processors known as Cascade Lake. These will be similar to the current Skylake-SP products that are currently in the market, but they'll be manufactured on an upgraded version of the company's 14nm technology called 14nm++ (the Skylake data center chips are built using 14nm+), and they'll incorporate some minor design changes to further improve performance.
Intel's products tend to last a year in the market, so the earliest we could see Intel transition to its 10nm data center processor family, known as Ice Lake-SP, is in the second half of 2019. If things don't go as well as expected for Intel's 10nm technology, we might see it launch in the first quarter of 2020.
On the other hand, Intel's competitors in the data center processor market could begin rolling out products manufactured using 7nm technologies, from the likes of Taiwan Semiconductor Manufacturing Company (TSM 2.63%), sometime in the second half of 2019.
Intel's 10nm technology and TSMC's 7nm technology are believed to be roughly comparable in terms of performance, power, and area -- the key metrics that determine how competitive a manufacturing technology is.
At best, Intel's 10nm server products could come to market at the same time as competing 7nm technologies, giving Intel no significant advantage in chip manufacturing technology and hurting the competitiveness of its products. At worst, the company may wind up trying to field its Cascade Lake processor family against 7nm parts from the competition, which could mean that Intel's processors wind up delivering much less performance and much higher power consumption.
In either case, I see risk to Intel's market share in its lucrative core market for data center processors. But the magnitude of that risk could be anywhere from moderate (if Intel and the competition are at manufacturing parity) to severe (if Intel is a full generation behind the competition in chip manufacturing technology).
Intel needs to make a better pitch
Right now, Intel hasn't provided a lot of clarity about its future product plans. This makes it hard to determine just what kind of competitive threat the company's delayed 10nm product rollout offers to its data center chip business.
Renduchintala's claims that he's "excited by the product pipeline" in the data center product market are all well and good, but investors have little to no insight into that product pipeline; outsiders have to take his word on the matter. But given that these statements are subjective at best -- and are, of course, colored by the fact that Intel pays Renduchintala's salary, so it's unlikely that he'd be anything but "excited" -- that's not convincing me to invest in the company's shares.
If Intel wants to regain investor confidence -- which I believe was shaken significantly when it announced its latest 10nm product delay -- then it needs to make a more substantive pitch to investors. The company should explain how it intends to defend its market share against aggressive newcomers with manufacturing technology arguably equal to or better than Intel's.
