AMD enjoys massive revenue and profit growth in Q1 2018

In the wake of a solid second-generation Ryzen CPU launch, AMD reported its first-quarter 2018 financial results this afternoon. The company announced revenue of $1.65 billion, up 40% year-on-year, and operating income of $120 million, up a whopping 990% from $11 million a year ago. To wrap up a fine quarter, the company reported net income of $81 million, reversing a course charted by a $33 million net loss a year ago. The company earned eight cents per share, up from a four-cent-per-share loss a year ago. Gross margin was 36%, up four percentage points.

  Q1 2018 Q1 2017 Year-on-year

change

Revenue $1.65 billion $1.18 billion +40%
Gross margin % 36% 32% +4%
Operating expense $477 million $394 million +$83 million
Operating expense % 29% 33% -4%
Operating income $120 million $11 million +$109 million
Net income $81 million -$33 million +$114 million
Earnings per share $0.08 -$0.04 +$0.12

Source: AMD

On a business-unit level, AMD's Computing and Graphics unit took in $1.12 billion in revenue, up 95% year-on-year, thanks to strong sales of Ryzen CPUs and Radeon graphics products. The company says its CPU and graphics-card average selling prices increased thanks to its Ryzen CPUs and new Radeon graphics products. Operating income for the segment was $138 million, compared to an operating loss of $21 million a year ago.

AMD's Enterprise, Embedded, and Semi-Custom revenue was $532 million, down 12% on the year. AMD says that result is thanks to lower semi-custom revenue, but that server and embedded parts helped arrest the slide. The division made $14 million in operating income, compared to an operating profit of $55 million a year ago. The outsize year-ago profit was thanks in part to a licensing payment the company received during that quarter. The All Other operating unit posted an operating loss of $32 million, compared with a $23 million loss a year ago.

For its second quarter of 2018, AMD expects revenue of $1.725 billion, plus or minus $50 million. If those projections hold, the company would increase revenue 50% year-over-year. The company expects a non-GAAP gross margin of 37% next quarter.

Comments closed
    • moose17145
    • 2 years ago

    This is very good news for AMD! And also consumers everywhere. Competition is good for everyone.

    If I were in the market for a new build, I would seriously be looking at either a Ryzen or (more than likely) a 2nd generation thread ripper system (when they eventually come out). Even though the Intel systems may be slightly better for gaming, I just like the overall AMD platform far better. Better IO, and fewer screwball games with the PCI-e lanes and other features, etc.

    I kind of wish TR was able to include some Broadwell-E results in their testing of the Ryzen chips. Would be nice to know how my current 6900K stacks up against the most recent AMD chips.

    • Growler
    • 2 years ago

    Thanks, Scott!

      • ronch
      • 2 years ago

      Still going, eh?

    • sweatshopking
    • 2 years ago

    Hilariously they could have made more money if they had just kept their own GPUs and mined with them. Selling them was a bad move.

      • tay
      • 2 years ago

      This is just so weird and unsustainable. Still waiting for a change…

      • tipoo
      • 2 years ago

      It wasn’t always logical to do so depending on your power cost. The people who made the most in the gold rush were selling the shovels.

    • chuckula
    • 2 years ago

    Good news AMD: 990% growth.

    Bad news AMD: Remember Jim Keller?

    Guess where [url=https://electrek.co/2018/04/25/tesla-autopilot-jim-keller-leaving-chip/<]he's working now[/url<]. THANK YOU JIM KELLER!

      • ronch
      • 2 years ago

      Well, as long as Jim doesn’t work on products that compete with AMD’s stuff, AMD is ok. Then again I bet there are already a bunch of Jimmies working at Intel anyway, those guys who worked on Sandy all the way to Ice Lake. AMD needs not one Jim, but a bunch of Jims.

    • ronch
    • 2 years ago

    With all that money starting to pour in AMD should seriously keep their foot on the pedal and not get complacent again like they did back when the K7 and K8 gave them a taste of how it feels to be on top. Use that money to go full speed ahead on R&D, and not just R&D, R&D with the best heads in the business. They were able to pull off Ryzen on a shoestring budget, they sure as F can do better with more money in the bank if they get real brains to work for them.

      • Anonymous Coward
      • 2 years ago

      AMD wasn’t ever complacent, they just had the wrong ideas. Having the correct objectives is a big deal, which money doesn’t really compensate for.

        • ronch
        • 2 years ago

        I think they’ve been complacent during the K8 years. They settled for a relatively easy objective of simply tweaking the K8 and gluing 4 of them together and adding more cache. And then 2006 came along.

          • Anonymous Coward
          • 2 years ago

          I disagree. Based on the fragments of rumors I’ve heard, they had ideas that didn’t work out, and ended up with K10 because they needed to do [i<]something[/i<]. Then it appears they tried to ship something more related to what they wanted to do instead of K10, and that was dozer. I'm saying the problem was lack of wise leadership in their design team, too much fumbling around and wasted effort. But then they got a wise design lead, and boom, Zen. A great name.

        • IGTrading
        • 2 years ago

        I think they were sabotaged in various different ways.

        After all, from 1999 until 2007, they had the complete technological/performance/efficiency superiority, BUT they were still barely pushing 25% in the server market, 40% in the computer market and even less in mobile.

        Remember those were the years when Intel was bribing and strong-arming OEMs left and right, in multiple countries, on different continents to not design solutions based on AMD chips ?!

        Intel was paying DELL more than 1 billion per year, not to sell AMD solutions. It is obvious why 🙂

        After at least 7 years of being on top, AMD was forced to sell its FABs, just 1 year later.

        That only happens if there’s something else going on, not complacency or “wrong” plans. Even Vishera proved to be a decent design and it would have been even more successful if Intel’s compiler wouldn’t have sabotaged any non-Intel chips or if software designers would have been real professionals and cared to make sure their application is not needlessly nonperforming on some architecture. I’m not even talking about specific optimizations. Just patching the compiler before applying it to the binary would have solved the issue and yielded more performance.

        Intel had multiple concurrent strategies specifically aimed to handicap its competition.

        We can’t really say Intel has put the same effort into optimizing and improving its solutions. Not even remotely. We actually had to go through 10 years of quad cores.

        Moreover, Intel’s off site software optimization engineers have encouraged the partners (software developers) to focus on single threading optimization (because it included the smallest amount of work/effort/expenses) and to avoid optimizing for more than 4 cores (because most market is made of dual & quad cores and to avoid further efforts/spending) and this is something which I happen to have first hand knowledge of.

        Sure the 1/2 ratio of FPUs wasn’t the best idea and the lack of money forced AMD to go for manufacturing technologies that were not the best solution, but remember that once software evolved, we see the FX8370 often beating the 2500K, which was not the case when the former was launched.

        Also, the high power consumption could have been easily solved if AMD would have had access to a good manufacturing process. But they didn’t and money was the main reason for this.

          • chuckula
          • 2 years ago

          See what the hell.

          I go accuse Nvidia of committing crimes that actually happened this century and nobody wants to hear it.

          But the usual copy-n-paste AMD feeling sorry for itself crap?

          Nobody has a problem with that apparently.

          But more on topic: If Intel did anything to keep Bulldozer or Piledriver out of consumer products so that regular people didn’t associate them with AMD, then Lisa Su should be sending Krzanich a fruit basket in thanks.

            • Kretschmer
            • 2 years ago

            We downvote you out of love, not for your opinions.

          • K-L-Waster
          • 2 years ago

          Paying billions more than they should have for ATI didn’t help…

            • Anonymous Coward
            • 2 years ago

            Man, its so annoying to hear people whining about that to this day. Maybe you’ll follow up by saying AMD didn’t need graphics?

            • chuckula
            • 2 years ago

            [quote<]Man, its so annoying to hear people whining about that to this day.[/quote<] I'm sorry, there must be a bug in TR's comments system. Your post isn't showing up as a response to IGTrading.

            • Anonymous Coward
            • 2 years ago

            Yeah there was some buzzing noise about monopolies in a previous age of the earth. I’ll pass on that.

      • Klimax
      • 2 years ago

      Has to be seen. There is fairly good chance they will run into same performance wall Intel seems to have hit.

        • Anonymous Coward
        • 2 years ago

        The performance wall was never in doubt. If I were running AMD, I think I would not invest too heavily in CPU core designs, instead focusing on platform and graphics-related processors.

        There might however be room for throughput-optimized processors, targeting the server space. Heavy on threads, IO and crypto, light on floating point.

          • ronch
          • 2 years ago

          Look what happened with Bulldozer. No, AMD absolutely needs to be competitive in the CPU space otherwise they’ll be more susceptible to pricing pressure from Intel. And having a competitive product does wonders not only to your market share but your mind share as well.

            • Anonymous Coward
            • 2 years ago

            Bulldozer was a very ambitious design by AMD, it wasn’t for lack of effort that they faltered. In my estimation, the CPU market is headed for continued stagnation as far as individual cores are concerned, and again in my estimation, AMD could do fine with a modest R&D budget and wise goal-setting. However a lot is happening in GPU and related fields, that is a place to put some money, I say.

            • ronch
            • 2 years ago

            The only really innovative thing with Bulldozer is the shared resources. Apart from that, each integer core was actually much simpler than what Intel did. Each integer core is quite narrow, therefore requiring a simpler scheduler and everything else. Even the caches were very high latency. The Load/Store units were strong though, with OoO loads and stores, and the FPU was pretty strong as well, supporting every new FP extension out there although AMD decided to give us half the AVX throughout/resources and instead relied on FMA to compensate.

            • Action.de.Parsnip
            • 2 years ago

            Noooo no no no no. Bulldozer was achingly complex. It made it’s awfulness even more stunning. The die size of a module was bizarrely high with a stonking transistor count. Every corner of bulldozer, every critical system was ground-up new and a tech level higher than in K8. 5 years work, prototyped on 45nm, endless massive effort to get it done and out. But it was dreadful and doomed the company to 5 years of utter wilderness and irrelevance.

            Don’t confuse narrow cores to simple cores. The execution units are a tiny pin-prick on any heavyweight cpu die, compare that to say the continent sized decoder or a branch predictor. It’s all the other stuff on the die feeding those units. The other 95% of chip area.

            The L2 was slow because it was shared. The L3 was slow because it was 8mb and 64 way associative.

            Bulldozer was a bloated out of control project. That chip is as advanced and complex as any Sandy Bridge. They made a L1 cache design decision that almost killed the whole corporation. Forget CMT it’s the least of the chips problems.

            • Anonymous Coward
            • 2 years ago

            Awesome description. Do you have a good source of insight here? Sounds plausible, anyway.

            My pet theory is that a CPU core a lot like dozer would be a great server product, given a lot of threads to work with. It seems anyway to parallel some of the choices made by Sun, except without the threads.

            • Action.de.Parsnip
            • 2 years ago

            Hi yes there’s more info out there than you might think but digging required. For the bulldozer overview this is your one stop shop

            [url<]https://www.realworldtech.com/bulldozer/[/url<] The author and techreport used to do podcasts together. In one he reiterated that bulldozers biggest flaw is the L1 cache design, to paraphrase "any performance just dies right there." Beyond that agners fog has a little site about cpu optimisation and using micro benchmarks he's found instruction latencies and quirks for all the famous modern cpus. In short bulldozer has no cache performance, no instruction decode performance (until Steamroller) and not enough instruction fetch performance (Steamroller cannot feed its decoders properly) It's a shocking kludge. You gotta look at the project management. Cmt was a dubious concept but the implementation was far more suspect. Compare that to say the Cell cpu which was semi dubious (hindsight notwithstanding) but was implemented well and had moderate success. Bulldozer could have done at least a little something with a stronger set of leads and some clearer less lofty goals. The concept of cmt had merits and in a many core many thread many socket environment there were the makings of say a good database/virtualisation uarch. The whole cmt idea was of a chip running with all the transistors doing useful work all the time. There's always a worthy use case for that. With a different L1 cache scheme and all the general tightening up seen in Excavator and a competent L3 then you've got something capable of useful work.

        • ronch
        • 2 years ago

        Yes they will. Sooner or later everyone catches up. I’ve said this before, and AMD just needs to come up with a wide core like Intel and they’ll be ok, and that’s what happened.

      • kvndoom
      • 2 years ago

      Hector’s long gone. I don’t think they’ll repeat the idiotic mistakes of the past.

        • ronch
        • 2 years ago

        Dirk Meyer was fired because of Bulldozer, I reckon. It’s not just Hector. But yeah Dirk’s gone too so there’s that.

      • Mirage8008
      • 2 years ago

      What company do you run? I presume it is doing better than AMD and we’ve all heard of it.

    • ronch
    • 2 years ago

    Eat your heart out, CTS!!! >:-)

    • Krogoth
    • 2 years ago

    Hopefully, Epyc platform pick-ups on the enterprise and SMB world. Ryzen APUs might be able to snag a couple of mainstream OEM sales too.

      • Anonymous Coward
      • 2 years ago

      Its surprising to me that there isn’t a much bigger slice of server sales.

        • Klimax
        • 2 years ago

        Intel covers almost every single part of that space and list prices are not even remotely near to real prices. While number of SKUs is insane it also provides near perfect selection for any particular need. (more or less cause for insane number o SKUs)

        Although I don’t know to what extent but there are quite some enterprise solutions where AVX(x) can be used to speed up processing. (e.g. MS SQL can use it)

          • Anonymous Coward
          • 2 years ago

          Intel certainly has a good product lineup, but AMD also has a solid platform with superior IO, and at the end of the day, price does matter, Intel isn’t giving their stuff away.

            • Klimax
            • 2 years ago

            Correct. That’s why I mentioned that list price is no where near real price. (90+% of servers are through OEM and thus with OEM prices on input)

            • Usacomp2k3
            • 2 years ago

            The CPU cost is usually a very small portion of an IT budget…

            • Anonymous Coward
            • 2 years ago

            So is basically any other component of the operation, taken individually.

            • Krogoth
            • 2 years ago

            For 1P/2P platforms this is quite true. It starts becoming muddled with 4P and beyond.

        • Zizy
        • 2 years ago

        Just big ones buy directly, the rest buy through OEMs and the tiny ones even through local resellers. Majority of OEMs are only slowly starting to offer these chips, HP announced theirs not long ago. Validation and stuff takes time. Plus everyone more or less knows how Intel parts will perform, while AMD ones are a wildcard – better in some tasks, worse in other. Now are you really going to sign the purchase until you can test the parts with your code?

        Server revenue should keep slowly growing, chips are good enough for that. But I don’t think it can become huge before 7nm parts (and not because current ones aren’t good enough – it just takes time in this space).

          • Anonymous Coward
          • 2 years ago

          AWS didn’t even get a *lake Xeon out before the end of 2017, thats slow movement. A lot of the stuff we use there isn’t even on 14nm.

    • Unknown-Error
    • 2 years ago

    See! It is not about bias. It is about having good products. During the Bulldozer days, the products were simply horrible. That why we all shouted AMD SUCKS, or said things like AMD is [b<][u<]Al[/u<][/b<]ways [b<][u<]M[/u<][/b<]ediocre [b<][u<]D[/u<][/b<]evices'. Market responds to good products. You don't have to beat Intel, you just have to provide a viable alternaive.

    • flip-mode
    • 2 years ago

    If you build it, they will come….

    if it doesn’t suck.

      • DPete27
      • 2 years ago

      And if they’re all sold out, it doesn’t matter if they suck or not.

      • ronch
      • 2 years ago

      But what if they’re making vacuum cleaners that REALLY SUCK HARD??

    • WhatMeWorry
    • 2 years ago

    I’m glad and all. Competition makes everyone better. But they have a pretty big hole (debt) to climb out of.

    • PrincipalSkinner
    • 2 years ago

    Excellent news for AMD. Now, let’s put some of those coins into GPU R&D.

    • dpaus
    • 2 years ago

    Would it be rude of me to ask Jeff if there isn’t, maybe, something else he should be working on…?

      • Eggrenade
      • 2 years ago

      This article probably took all of 45 minutes to research and write, so it’s probably not a big deal. But, yeah, I would have expected Jeff to get one of his underlings to write this.

      • odizzido
      • 2 years ago

      Yeesh he has probably been slaving away for ages on it. At most he took a break from it to do other work on this site. Maybe he is currently running a test that takes time and he did this while that was going?

      I think I’ve maybe pressed the down vote once, if that. But I am hitting that button now. Give the guy a break.

      • Srsly_Bro
      • 2 years ago

      I LOLed so hard at this. Thank you, dpaus.

      • derFunkenstein
      • 2 years ago

      if you follow him on Twitter you’ll see updates on that progress. Even if you don’t follow him, his twitter is on the bottom of the front page.

    • bthylafh
    • 2 years ago

    I’m glad for them, but I wish it wasn’t mostly from the perkeleen cryptominers.

      • Pancake
      • 2 years ago

      Ah-ha-ha-ha. It WAS mostly from cryptomining. May that evil waste of resources end.

    • chuckula
    • 2 years ago

    I bet that if you put Intel’s and Ngreedia’s profits TOGETHER they still won’t come close to a 990% gain!

    Especially when Intel won’t even report a profit and Ngreedia only makes a tiny profit due to its illegal manipulation of the coin mining market.

      • dodozoid
      • 2 years ago

      You sure we don’t have any real AMD shills left so you have to cover the role yourself?

      Edit: badly…

        • chuckula
        • 2 years ago

        What are you talking about, I insulted Intel and called Nvidia “Ngreedia”.

        I also accused at least one of them of committing a crime.

        Are you upset that I didn’t copy-n-paste the typical whine about Intel “illegally” keeping AMD out of Dells in 1999? Because [url=https://techreport.com/discussion/33531/amd-ryzen-7-2700x-and-ryzen-5-2600x-cpus-reviewed?post=1075725<]one of those just got a +11 upvote the other day[/url<] and I think I deserve better treatment for at least keeping the allegations of criminal behavior in this century thank you very much!

          • dodozoid
          • 2 years ago

          Ok, you got all the cliches right, but it lacks the spirit, the sincerity, raw enthusiasm…

            • mtcn77
            • 2 years ago

            Internet troll romanticism is a thing.

          • cegras
          • 2 years ago

          Are you ok?

        • tay
        • 2 years ago

        Chuck is like a one person show. He needs to play the AMD troll, the NGreedia troll, and the intel troll at the same time.

    • techguy
    • 2 years ago

    That’s great news! Way to go AMD. Congrats to all those that worked hard to turn the company around over the past several years. More competition is good for every one of us.

      • blastdoor
      • 2 years ago

      It’s a ding dang miracle is what it is. I really don’t understand how AMD stayed in business, let alone rising from the ashes to compete with Intel again. Maybe someday somebody can write a book about it.

        • Pancake
        • 2 years ago

        Debt. The miracle of an extended period of very low interest rates. The challenge for AMD is to reduce their billion+ debt before interest rates rise as they inevitably will. Now is a way scary time to have any significant debt exposure.

          • w76
          • 2 years ago

          Yes, you’re right. This quarter is good, but it doesn’t erase the years of losses and shrinking revenue or do much at all to offset the wave of debt maturities approaching. Interest rates seem set to continue rising, at least for now, so even if the market allowed AMD to roll over it’s debt, their interest costs could soar.

          Just my opinion, but reducing debt should be a top risk-reduction activity for them, so they can’t afford to get lazy and let costs rise.

          • derFunkenstein
          • 2 years ago

          what’s amazing is that they can actually afford their debt now. Can’t turn a profit if you can’t pay the interest.

            • blastdoor
            • 2 years ago

            I wonder if we will discover one day that AMD actually invented BitCoin….

          • blastdoor
          • 2 years ago

          I might be misreading this, but according to here:

          [url<]https://finance.yahoo.com/quote/AMD/key-statistics?p=AMD[/url<] (by way of contrast, take a seat and a stiff drink and look at this: [url<]https://finance.yahoo.com/quote/TSLA/key-statistics?p=TSLA)[/url<] AMD's debt doesn't seem so bad. They have about $1 billion in cash and about $1.4 billion in debt, so net debt of about $400 million. If they were piling up losses then this would suggest they have a big problem, but so long as they can keep turning quarterly profits for the next few years they should be fine. The thing that I don't understand is how they don't have far more debt than this. I guess I don't remember the particulars of the fab spinoff to GloFo -- maybe that gave them more financial cushion than I remember. And of course they've had a lot of layoffs. But still, even with all of that, it's amazing.

            • Beahmont
            • 2 years ago

            That is not how accounting, especially Non-GAAP accounting, works. Nor is it how balloon loan and their payments work.

            When that debt comes due, I believe it’s in 2020, AMD either has the 1.4 billion on hand or it dies and has it’s assets sold off and liquidated to cover what it didn’t have. The company litterally can’t run with zero cash on hand, let alone a negative cash on hand.

            Now, that’s if AMD can’t restructure the payments and extend the balloon date once again. But $81 million a quarter in profit isn’t going to cut it if AMD wants to survive past 2020.

            It’s also very easy to not be able to pay interest on loans and show a profit if you structure the debt properly. In fact that’s why GAAP has been slowly updating in recent years to prevent certain kinds of accounting tricks that hide debt and unprofitability.

            • Anonymous Coward
            • 2 years ago

            I might be ignorant, but it seems to me that 81 million a quarter isn’t a bad chunk compared to 1.4 billion in debt. Surely bankruptcy would be waste. (I’m not sure they can be expected to keep that level of profit though.)

            • blastdoor
            • 2 years ago

            It’s not bad at all.

            I suspect the key is whether Intel will tolerate some marketshare loss in order to keep their margins high. If Intel will do that, then AMD could maintain or even increase profit.

            The risk for AMD is if Intel decides to get ruthless.

            • Beahmont
            • 2 years ago

            The problem is that $81 million in profit is before reinvestment and various other things that are going to chip it down in size. They are probably increasing their cash on hand no more than $20 million, and probably closer to $10 million.

            They have less than two years at this point to come up with another $400 million in cash on hand just to repay their debt. They also have to find even more money to reinvest in the company to just keep pace with where they are now in comparison to Intel. They also have to find the extra money to have enough cash on hand to continue operations. And if they really want to not go the way of the Dodo, they need to increase their cash on hand enough to survive a disaster or two.

            And that’s all if something doesn’t happen to the US or Global economy within the next two years that makes them repaying their balloon debt impossible. And the way things are going now, it’s not unreasonable at all to believe the economy is going to drop off a cliff again soon.

            AMD has made an amazing recovery, but they are still in critical condition. But there is a hard deadline for their recovery, and it’s still a big ‘IF’ that they will make that deadline.

      • Waco
      • 2 years ago

      No kidding! I’m glad to seen the red team not in the red!

    • dpaus
    • 2 years ago

    First! report of ‘massive revenue growth’ for AMD in a long, long time. Good for them.

      • dodozoid
      • 2 years ago

      From zero, ANY growth is massive in terms of percentage…

      • RDFSteve
      • 2 years ago

      Two consecutive ‘first’ posts?? Must be a very slow day. Maybe everyone is sitting on the 2700X review page hitting ‘refresh’ constantly as they wait for the rest of the review?

Pin It on Pinterest

Share This