Today AMD announced their third quarter earnings for the 2019 fiscal year, and AMD has not seen revenue like this for a long time – in fact this is the highest quarterly revenue since 2005 for the company. AMD’s revenue jumped 9% year-over-year to $1.8 billion, and at least as importantly, AMD had gross margins of 43%, which is up 3% over last year, and the highest margins they’ve seen since 2012. Operating income was up 24% to $186 million, and net income was up 18% to $120 million. This resulted in earnings-per-share of $0.11, up 22% from Q3 2018.

AMD Q3 2019 Financial Results (GAAP)
  Q3'2019 Q2'2019 Q3'2018
Revenue $1801M $1531M $1653M
Gross Margin 43% 41% 40%
Operating Income $186M $59M $150M
Net Income $120M $35M $102M
Earnings Per Share $0.11 $0.03 $0.09

This is the first full quarter for AMD since the launch of their 7 nm Zen 2 processor, and AMD attributes the revenue growth to the Computing and Graphics, but offset by lower revenue in Enterprise, Embedded, and Semi-Custom. Revenue for the Computing and Graphics segment was up 36% year-over-year to $1.28 billion, thanks to both increased volume and Average Selling Price (ASP) for Ryzen on the desktop. GPU ASP also increased year-over-year thanks to higher channel sales. The Computing and Graphics segment had operating income of $179 million, which is up 79% from a year ago.

AMD Q3 2019 Computing and Graphics
  Q3'2019 Q2'2019 Q3'2018
Revenue $1276M $940M $938M
Operating Income $179M $22M $100M

Enterprise, Embedded, and Semi-Custom had revenue of $525 million for the quarter, down 27% year-over-year, mostly attributed to semi-custom sales, which makes sense since a large chunk of that is for the AMD APU powering both the Sony PlayStation 4 and Microsoft Xbox One, both of which are scheduled for new models in the next calendar year. Offsetting this was higher EPYC processor sales, although not enough of an offset to cover the semi-custom drop. Operating income for this segment was $61 million, down 29% from a year ago.

AMD Q3 2019 Enterprise, Embedded and Semi-Custom
  Q3'2019 Q2'2019 Q3'2018
Revenue $525M $591M $715M
Operating Income $61M $89M $86M

Finally, AMD’s All Other category reported an operating loss of $54 million, which is a 50% larger loss than a year ago.

AMD had some big news in Q3, with multiple design wins for both Ryzen and EPYC, including Cray’s Shasta supercomputer leveraging 2nd Generation EPYC, and AMD getting a big design win in the PC space with the Microsoft Surface Laptop 3.

Looking ahead to Q4, AMD is expecting revenue of $2.1 billion, plus or minus $50 million, with a Non-GAAP gross margin of approximately 44%.

Source: AMD Investor Relations

Comments Locked


View All Comments

  • Spunjji - Wednesday, October 30, 2019 - link

    This. they won't make significant inroads until their platforms have been available for a few years and people know what to expect.
  • rahvin - Wednesday, October 30, 2019 - link

    As AMD CEO has said, they only way they can succeed in the Server space is to continue to execute. They have to show they are in this for the long haul and not a flash in the pan. They are taking a significant chunk of the Cloud servers already but it's going to take time to build marketshare.
  • FreckledTrout - Wednesday, October 30, 2019 - link

    Honestly the market share is where I expected. Single digit for first gen and low double digits for second gen EPYC. I expect higher double digit on third generation EPYC like 20-25% share especially if they beat Intel to the punch with Zen3. It's like turning a battleship, you turn the steering wheel and 5 minutes later you have turned around.
  • Tams80 - Wednesday, October 30, 2019 - link

    I mean, AMD have had to convince server administrators that their products are even worth considering. Then from those, they've had to get those with resources to spare/play with to buy products. And then, they've had to show that they can support their products over a long period of time.
    Only then will anyone significant meaningfully put any of their eggs in AMD's basket. AMD are only just at that point. However, consideriing how intrangent the server space is and how slowly it changes, it's looking very good for AMD.
  • rahvin - Wednesday, October 30, 2019 - link

    Epyc 1 revenue is in there. Epyc 2 hasn't been out long enough to effect the prior quarter's numbers.

    AMD is doing very well for the competitive position they are in with Intel's market manipulations. AMD is beating doing better in servers than they'd hoped. Servers is an area where it take a long time to build up market share, it can take almost 5 times longer to build marketshare in the server market than it takes in the retail side.

    Intel is easily 10 times bigger than AMD so them taking even a double digit in server sales is a major achievement considering the dollars Intel can dump at marketing and incentives. In the past Intel used these marketing slush funds to keep AMD locked out of the OEM marketplace. Lets not forget, when Athlon was king Intel payed Dell almost $5 Biillion to keep them from buying AMD chips. Intel plays dirty so AMD taking marketshare means they are providing much better products for the price.
  • BurntMyBacon - Thursday, October 31, 2019 - link

    It is not just this. Intel is also struggling to produce enough chips to keep up with demand due to 10nm struggles simultaneously lowering the total number of chips produced and forcing more concurrent products to use their 14nm node than they planned for. From a business standpoint, there is a lack of confidence when your supplier doesn't have enough supply to go around. When Intel finally gets 10nm manufacturing up to the point that they are comfortably meeting both 14nm and 10nm product demand, then this confidence issue will disappear. Hopefully, AMD will have sufficiently proven their merit to the enterprise customers by then. Otherwise, it will be hard to hold on to the market share they've earned since producing Zen. I'm cautiously optimistic about their outlook here.
  • FreckledTrout - Wednesday, October 30, 2019 - link

    AMD is doing everything right. They really need some decent marketing like "Intel Inside" next level marketing.
  • Targon - Wednesday, October 30, 2019 - link

    How about "Intel not inside" as a good thing, since Intel chips have so many security holes, most people who are concerned about security should avoid them.
  • yeeeeman - Friday, November 1, 2019 - link

    Most of those security holes are fixed in Ice Lake. Also, those security holes were pushed by several, not very good intended people, to create bad press for Intel and lead people like you to believe Intel doesn't give a shit about security. You are very wrong. All pieces of hardware have security holes, AMD does, ARM does, nobody escapes from it. The fact that Intel took some wrong decisions when designing its uArch, doesn't mean they did it on purpose. So lets stop with this BS and move on.
  • TechIsFun - Wednesday, October 30, 2019 - link

    EPYC is doing fine, as AMD had essentially 0 presence in the DC 5 years ago. That means they had little to no validation in place, which is important in enterprise spaces. The EPYC sales you see so far are for DEV environments, so businesses can validate their software stacks for production. From what I've seen, AMD has been pretty good with working out firmware/software issues between hardware/software vendors. The change in NUMA from Naples to Rome is a good example of that. Naples let you carve out really nice 4-8 core VMs, where Rome should handle 32 vCPU VMs with little overhead. The future is virtualization, so Intel is pretty hamstrung until they can cram 64+ real cores with good IPC and decent clocks (at decent power levels) into the same socket. Once you add in the lack of arbitrary segmentation, there is little reason to buy Intel at the moment unless you simply MUST scale production RIGHT NOW.

    As for AMD's enterprise revenue being only a tiny portion of the overall, see the above statement about their presence in enterprise. They know they have to grow this market, and I think it shows in how they have allocated silicon in their Zen launches. That's a whole other game on top of building an architecture, making sure you can make enough to sell enough to keep on trucking. You really want a shock? Compare these financials to similar from Intel and figure out how AMD is even being considered an option.

    The near future for AMD is desktop, next-gen consoles, and mobile. Looking at what AMD did with the new Surface model, I can't wait to see what they can do with Zen2 in mobile. I think the coming fight in Mobile between AMD and Intel will be pretty exciting. Thin and light laptops that can pull off AAA 1080P gaming sound cool to me. Heck, even ARM is making moves.

    The real question is if AMD can keep it up. They are in sweet spot right now, having caught the dragon sleeping. What kind of traction can they build before Intel sorts itself out and achieves good volumes with SunyCove/10nm? I mean, Intel will be fine, with their many verticals and established contracts. But, they can't stand another 2-3 years of not being the defacto choice without opening the door to actual, long-running competition. Sunny Cove shows some decent IPC gains, but that was a long time coming, and resulted in lower clocks. Given how long it took to come to market, I don't think we can make the usual assumption that Intel will have a newerfasterbetter version out any time soon. I think they will be busy trying to push it up into DT/HEDT/ENT. Can AMD keep ramping clocks AND IPC? They have pretty much saturated core count for the near future as far as I'm concerned. 16 cores on desktop, 32 cores on HEDT, and 64 cores per socket in servers sounds like a nice stack compared to Intel. Can AMD keep pushing clocks, can they keep improving IPC, can they do both? How fast can Intel counter? Isn't Keller still working there?

    It's fun being excited about CPUs again. I hope the coming GPU wars are as much fun.

    As for GPUs, the 5700 XT has turned out to be a proper foil for the 2070 Super launch. Most people I know, even avid PC gamers, simply won't spend more then $400+shipping on a GPU. And of the rest, most won't go over $300. AMD has this whole market tied up, with the exception of those with G-Sync displays and those that just have to have RT. The thing is most with G-Sync monitors likely shop higher anyways (they got the moneys) and RT is pretty much crap on anything below a 2080/2070s. Honestly, the 5700 XT is the $400 upgrade from the GTX 1070 everyone has been whining for. If they can scale it up and keep improving, we could have actual competition for high-end GPUs.

    Like I said earlier for Intel, Nvidia will be fine. The question is, "Can AMD make this a real fight or just another decent round where they get in a few more hits but gain no meaningful ground?". I like the spending on R&D, we just have to see what it's buying.

Log in

Don't have an account? Sign up now