When AMD released its Q4 revenue figures for 2018, it said it would wait for additional data on market share to come in from the analyst firms that track this sort of information. Now, Mercury Research has released data on AMD’s overall performance throughout the quarter. It’s all good news for AMD, though Mercury didn’t find that the company hit its goal of midrange single-digit market share for 2018. on that in a moment.
From Q4 2017 to Q4 2018, AMD expanded its share of the desktop market by 1.32x, a gain of 3.9 share points. Notebook sales grew at a faster clip, with AMD moving from a 6.9 percent movement share in Q4 2017 (just after the launch of Ryzen Mobile) to a 12.1 percent share in Q4 2018. That’s a 1.75x improvement in shipments, and a gain of 5.3 share points year-on-year.
Finally, there’s server. According to Mercury Research, AMD started in Q4 2017 with a whopping 0.8 percent market share. In Q4 2018, AMD stands at 3.2 percent market share, a 4x improvement in just one year, and a gain of 2.4 share points.
AMD, however, has already stated that it believes it hit its target of mid-single-digit market share for Epyc in 2018. How to square these two statements? The company included some additional information that sheds light on how it sees the situation.
Mercury Research captures all x86 server class processors in their server unit estimate, regardless of device (server, network or storage), whereas the estimated 1P and 2P TAM provided by IDC only includes traditional servers. We used IDC’s server forecast of the 1P and 2P server TAM of roughly 5M units to compute our server market share estimates. We believe that in Q4 2018 we achieved ~5% unit share of the 1P and 2P server market addressed by our EPYC processors (as defined by IDC).
AMD, in other words, believes it met a version of this goal, barely. But this is a point where I’d avoid reading too much into the server ramp from either direction. It’s not unusual for server ramps to be slow. Server customers are conservative and they want to see that a company can deliver a series of parts across multiple generations before diving in. It took AMD multiple years to begin building server share when it took this fight on the first time. Not hitting the 6-8 percent mark according to all metrics is a bit of a miss compared with where AMD wanted to be, but it may not be a particularly consequential one.
Intel isn’t in the same position it occupied back when AMD’s Opteron first rose to prominence. Back then, Intel was struggling with Prescott and Smithfield, while the Athlon 64 and Opteron were positively tearing up the charts. AMD’s “glueless” architecture for Opteron offered additional scaling potential that the Xeons of the day struggled to match. Even so, as the graph below shows, it took several years for Opteron to ramp successfully.
In other words, there were strategic and competitive factors in play 13-14 years ago that aren’t pushing the market in the same fashion now. This is not to say that AMD won’t continue to gain server share. But 7nm Epyc is where I’d expect to see adoption accelerate, provided that part proves to be competitively positioned against Intel’s equivalent Xeon stack.
These numbers show strong growth for AMD in multiple markets, emphasizing that the company’s financial improvement wasn’t simply driven by crypto sales. Intel obviously remains the dominant player in all three spaces, but the market is clearly responding to Ryzen’s value proposition across the entire computing space. Increased competition and better product options are always a good thing for consumers, regardless of what kind of hardware they’re looking to buy.