Coming off a record quarter, Intel has reason to celebrate, as well as be optimistic about where it's headed. It collected $19.2 billion in the third quarter of 2019, largely without the benefit of volume 10-nanometer CPU shipments. The prosperous quarter also brought with it some interesting introspection. In a subsequent chat with investors, Intel CEO Bob Swan alluded to AMD racing past the company in process technology.
He didn't actually mention AMD by name, but did say Intel is "investing to recapture process leadership going forward." That's a clear reference to AMD and its 7nm lineup, which now extends across both its latest generation CPUs (Ryzen 3000 series) and GPUs (Radeon RX 5000 series).
The comment sticks out because of the tough road Intel has traveled to 10nm. Cannon Lake was supposed to ship a long time ago, and instead we are just recently seeing 10nm Ice Lake processors for laptops, with 10nm desktop CPUs yet to be released.
Without taking a deep dive in architectures and process nodes, it's fair to say Intel's 10nm technology is roughly on par with AMD's 7nm tech. However, it's been a long and frustrating journey to 10nm for Intel, which instead has continually refined its 14nm node (with even more 14nm processors on the horizon).
The saving grace for Intel has been performance. While it's 14nm product line is no longer the most advanced, they are still stout chips; the Core i9-9900K is the best CPU for gaming.
What's more encouraging, though, is Intel's belief it's finally back on track. At long last, 10nm parts are finally shipping in volume, even if only in laptops for the time being, and Intel is already looking ahead to 7nm and even 5nm.
"We're ramping a multitude of products. We have increased confidence in 5-nanometer. And as we mentioned for 7 and 5 getting back to a two-and-a-half, two year cadence is what we're focused on and we're confident in the future," Swan said.
That two-year cadence went out the window with all the challenges associated with Cannon Lake. Earlier in the year, Swan admitted Intel had set a performance goal that was simply too aggressive, and erred when it "prioritized performance at a time when predictability was really important."
It's also worth noting that Intel's original 10nm tech was based on a different, more challenging type of lithography. It's 7nm tech is a separate effort by a separate team, and is based on a less problematic extreme ultraviolet lithography (EUV), which is probably where Intel's newfound confidence is coming from.
Intel still has to execute, of course, but at least now there appears to be light at the end of the tunnel. Fortunately for AMD, it's already there, and looks to be in good shape going forward as well.