Today AMD released their Q1 FY 2015 financial results, and the company reported revenue of $1.03 billion for the quarter. This is a 16.9% decrease as compared to Q4 2014, and a 26.4% decrease from the $1.40 billion recorded in Q1 2014. Operating income based on GAAP numbers was an operating loss of $137 million, which is a substantial decrease in loss as compared to Q4 2014, where they had an operating loss of $330 million, however in Q1 2014 they had a small operating income of $49 million, so although they have improved quarter-over-quarter, that is a significant reduction year-over-year. Net loss for the quarter was $180 million, or $0.23 per share, which once again is better than Q4 2014 where there was a $364 million ($0.47/share) loss, but much worse than the $20 million ($0.03/share) loss in Q1 2014.

AMD Q1 2015 Financial Results (GAAP)
  Q1'2015 Q4'2014 Q1'2014
Revenue $1.03B $1.24B $1.40B
Gross Margin 32% 29% 35%
Operating Income -$137M -$330M $49M
Net Income -$180M -$364M -$20M
Earnings Per Share -$0.23 -$0.47 -$0.03M

Part of these losses are due to the ongoing restructuring at AMD, which has contributed heavily to these numbers. One of the new restructuring fees is due to the exit from the Seamicro branded dense server business, which has cost them an additional $75 million this quarter, including $7 million in cash. Due to these hits, AMD also provides Non-GAAP results which exclude these numbers. On a Non-GAAP basis, AMD’s operating loss is just $30 million, however that is still down significantly from the $89 million operating income in Q1 2014, and the $52 million operating income from last quarter. Net loss on a Non-GAAP basis is $73 million, or $0.09 per share. This is a decline from Q4 2014 where there was a net income of $18 million ($0.02/share) and Q1 2014 where they were able to achieve a net income of $35 million ($0.05/share).

AMD Q1 2015 Financial Results (Non-GAAP)
  Q1'2015 Q4'2014 Q1'2014
Revenue $1.03B $1.24B $1.40B
Gross Margin 32% 34% 35%
Operating Income -$30M $52M $89M
Net Income -$73M $18M $35M
Earnings Per Share -$0.09 $0.02 $0.05M

AMD has also entered into a fifth amendment of their agreement with GlobalFoundries, and AMD is expecting to purchase about $1 billion in wafers in 2015.

Breaking down their product segments, the Computing and Graphics segment had a 20% decline in revenue quarter-over-quarter, and a 38% decrease year-over-year, with Q1 having net revenues of $532 million. The quarterly decrease was due to lower desktop and notebook processor sales, whereas the yearly decrease was due to lower desktop processor sales and GPU channel sales. The division had an operating loss of $75 million for the quarter, which is a significant change from the $56 million loss last quarter and the $3 million income in Q1 2014. The loss was partially offset by lower operating expenses, but clearly more work is needed. AMD is hoping for better success with their new APU, Carrizo, which they are expecting to deliver double digit performance increases and much better energy efficiency compared to Kaveri, which is the current APU.

AMD Q1 2015 Computing and Graphics
  Q1'2015 Q4'2014 Q1'2014
Revenue $532M $662M $861M
Operating Income -$75M -$56M $3M

The Enterprise, Embedded and Semi-Custom segment had a year-over-year revenue decrease of 7%, and a quarter-over-quarter decrease of 14%, with Q1 2015 coming in at $498 million. The quarterly drop is due to a seasonal decrease in semi-custom SoC sales (read: Consoles had a ramp up for the holidays and are now back to lower sales) and the yearly decrease is due to lower numbers of server processors being sold. However this segment did have an operating income to report of $45 million for the quarter, but this is down from the $109 million in Q4 2014 and $85 million in Q1 2014.

AMD Q1 2015 Enterprise, Embedded and Semi-Custom
  Q1'2015 Q4'2014 Q1'2014
Revenue $498M $577M $536M
Operating Income $45M $109M $85M

The “All Other” segment had an operating loss of $107 million. As compared to Q1 2014, this is $68 million more operating loss, which is primarily due to the $75 million hit for exiting the dense server business. In Q4 2014 this segment had a $383 million loss.

For Q2, AMD is forecasting revenue being down an additional 3%, plus or minus 3%, and non-GAAP Gross Margin to remain flat at 32%.

AMD is certainly not in a great position right now, and the new CEO Dr. Lisa Su has some work to do in order to get AMD back to a financially viable state. Part of that is diversifying revenues, especially with the PC market slowing again. AMD has not had a significant product launch in a few quarters, which has not helped either.

Source: AMD Investor Relations

POST A COMMENT

140 Comments

View All Comments

  • lordken - Friday, April 17, 2015 - link

    Well mantle works on HD7000 series also (should work on all GCN). As for freesync
    I'm not sure, it looks it needs something else just driver update - as not even whole R9 series support it. Though I would love if freesync would work on my 7950
    Reply
  • silverblue - Saturday, April 18, 2015 - link

    That's exactly what I said. NVIDIA fixed their issues two years before AMD fixed theirs. FCAT was a method of showing AMD up in front of its customers for not listening, and it worked a treat.

    Your opinion on G-Sync is partly clouded by owning a monitor with said module inside, and the fact that the first few FreeSync monitors have been wide of the mark. The drivers can be tweaked to resend frames to reduce low FPS tearing, sure, but it would be incorrect to place the blame at FreeSync's feet as the standard isn't at fault for the tearing or flickering.

    NVIDIA's last generation was the 7xx series; VLIW4 was AMD's last pre-GCN series. It's hardly the same thing. As it is, the 780 Ti should be beating the 290X universally, but if you take a look at DX11 gaming results in AT's GPU 2015 bench section, the 290X goes from a small loss (on average) with older titles to a small win (on average) in newer ones. Shouldn't the 780 Ti be winning by 15-20%? The 7xx series is only JUST the last generation; in AMD's defence here, they've had the 2xx series out for a while, however they should've done better with pushing those advancements back to GCN1.0. I suspect the lack of the XDMA bridge is mainly the reason why the 7xxx series doesn't get as much love. FreeSync is GCN1.1 and up due to their DisplayPort support, hence the reason why GCN1.0 won't get it, though you could implement an algorithm to send extra frames if the rate doesn't hit the target.

    You haven't mentioned TressFX... I'm not sure what AMD were thinking here. Who cares about Lara's hair?
    Reply
  • chizow - Saturday, April 18, 2015 - link

    No, Nvidia knew about framepacing and had a software fix for it since G80 and a hardware fix for it since Fermi (source: Tom Petersen interviews). Inventing the current implementation of multi-GPU/AFR, debugging, and providing a solution before the competition even knows about it does not fit your characterization of "Nvidia also having framepacing issues".

    My opinion of G-Sync isn't clouded at all, its *ENLIGHTENED*, there is a difference. If Nvidia invents a Variable Refresh Rate solution and claims it solves all of the issues that made V-Sync On/Off undesirable, and their working product actually does all they said it does, that becomes the STANDARD for what you would expect from such a solution. When the competition fails to meet that standard, obviously G-Sync remains the superior solution as we have seen with this initial introduction to FreeSync.

    Also, you should be careful about your claims about what can and cannot be accomplished via "driver tweaks" lest you add to and perpetuate the misinformation regarding FreeSync fixes. If it it was so easy, why hasn't AMD fixed it yet? There were already rumors and concerns about ghosting at CES, why wasn't it resolved in the 3 full months since then? Why hasn't it been fixed in the month since then if it is just as simple as tweaking drivers? Maybe there was some truth to what Nvidia said, that getting VRR right is hard, and that that their ultimate conclusion that additional hardware in the form of a scaler that directly communicated to both the GPU and TCONS with its own local frame buffer, was the more sound approach?

    It is clearly the fault of the spec/standard when the flickering and ghosting does not exhibit itself when FreeSync is disabled. I'm not sure how this is hard to understand, AMD came up with a very simple solution of relying on Vblank but didn't work out the intricacies of other fixed frequency functions, like overdrive and backlight pulse. Nvidia did. The end result is evident, FreeSync is a decent attempt at copying G-Sync, but it still falls short in a number of key areas which ultimately means it fails at solving the key issues with V-Sync, which is the fundamental reason for using VRR to begin with.

    When AMD abandoned VLIW4/5 and went to GCN there were still VERY relevant parts on the market with the 5850/5870 and their last-gen was VLIW5 6000 series. But even within GCN arch AMD has already made clear distinctions in support levels with FreeSync support, FramePacing and even Mantle support. And CF? Good luck, its either still broken with many of these features or in constant states of half-support.

    And title support? It was always the 780Ti winning at lower resolutions up to 2560 and the 290X pulling ahead at higher resolutions. This can mainly be attributed to differences in ROP/mem controllers and Nvidia clearly fixed their ratios, re-emphasizing ROP to TMU with Maxwell. Look at recent titles like GTA5 for example. Everyone panned the GTX 680/770 for falling out of support and being abandoned and getting trounced by the 7970/280X yet there it is, often pulling ahead of the Tahiti-based parts. It just depends on the game engine but clearly Kepler still has legs and continues to be supported which is still certainly better than the support level you can expect from Tahiti vis-à-vis AMD's newer technologies.
    Reply
  • TheJian - Monday, April 20, 2015 - link

    http://www.pcper.com/reviews/Displays/AMD-FreeSync...
    "For its part, AMD says that ghosting is an issue it is hoping to lessen on FreeSync monitors by helping partners pick the right components (Tcon, scalars, etc.) and to drive a “fast evolution” in this area."

    Apparently you haven't read about ghosting etc (even anandtech had to mention it after PCper's 19th article). Hoping to LESSEN? So not hoping to TOTALLY STOP IT like NV? Stop reading so much anandtech where they minimize amd's problems and get the rest of the facts elsewhere. ;) AMD currently blames the hardware, but you can't fix current monitors with a driver. Even AMD says they'll have to help monitor makers pick proper parts to improve the experience, and they can't seem to force that. On top of that BS, there is no proof it's PURELY about components chosen by the monitor people. Until we see a unit work without ghosting etc, it is AMD vapor-logic IMHO.

    You certainly wouldn't want to be one of the buyers of beta 1 versions of freesync :( I call it beta until we see it match or beat NV's gsync (with this special AMD picked components...LOL), and certainly beta while AMD is claiming it's the parts in the monitors that are at fault. I mean you expect me to believe they couldn't tell monitor makers "don't put this in it, as it will give our tech a bad name for launch monitors, use that junk on later models if desired please". Really? I would have thought you would have at least funded ONE monitor with correctly picked parts if that was the case right? They're hoping to drive a fast evolution of monitors? ROFL. So you'd be wise to NOT BUY FREESYNC until it's EVOLVED then right? It's comic you don't read AMD's comment on anandtech ;) (unless i missed it). It would appear they've cut costs by going crappy or they're just not telling the truth.

    http://www.anandtech.com/show/9097/the-amd-freesyn...
    "There may be some additional costs associated with making a FreeSync display, though mostly these costs come in the way of using higher quality components. The major scaler companies – Realtek, Novatek, and MStar – have all built FreeSync (DisplayPort Adaptive Sync) into their latest products, and since most displays require a scaler anyway there’s no significant price increase."

    Well I guess if its not working properly with current parts, and as stated better parts cost more (well duh), freesync isn't so free ;) How much will it cost to get it right? As usual anandtech sugar coats AMD problems as much as possible. "No significant price increase" for the JUNK models. But how much for "works like gsync" models? They end claiming freesync is winning (by numbers anyway), but forget to say why would you want ghosting crap etc? I call that losing. The reason you buy one of these techs is to FIX the problems, not still deal with them.
    Reply
  • FlushedBubblyJock - Saturday, April 25, 2015 - link

    silverblue, there are so many additional nVidia features that AMD doesn't have it's absolutely ridiculous.

    I think the count is likely beyond 10 by now. Of course none of them matter to amd fans such as yourself, that's why AMD is almost defunct.

    I'm sure the AMD monitors come by and see what their little fanboys say, and they constantly see they just don't care for all the many extra nVidia features, so why would AMD do a dang thing to improve your end user experience ?
    None of you want any of it.
    No game day drivers, who cares
    No PhysX, who cares
    No auto config card settings for hundreds of the most popular games, who cares
    No adaptive sync, who cares
    No twitch in game broadcasting and recording, who cares

    It's one after the other, it's on and on and on .... but amd fans just want cheap barely higher and not worthwhile fps, that's all....

    It's no wonder AMD is in a deep hole it cannot dig itself out of - it's fanboys are frankly a detriment!
    Reply
  • TheJian - Monday, April 20, 2015 - link

    If NV was gouging they'd be making as much as 2007, which they haven't (even if you could say they were then at ~800mil/yr income). If you take out Intel's 266mil/yr right now you end up with NV making ~200mil/yr, which is 1/4 of what they were making in 2007. Gouging? LOL. Making peanuts because AMD insists on pricing themselves to death while trying to price NV to death. Very stupid management over there at AMD. You can't price someone to death when you're the guy with debt, and the other guy has 3B+ cash.

    IF NV was gouging titanx wouldn't be selling for 20%+ over retail on ebay daily. Clearly they could have put a $1200 sticker price on it and sold the same. TitanX vs. 295x2, you likely need a PSU with AMD here, and your electric bill over 4-5yrs (longer when buying this high end?) will cost you $200-400 depending on where you live and how much you game. I think its safe to assume someone spending this much games a fair amount. My assumption is at ~12c/kwh and 21hrs a week gaming. IF you have a kid or two this can be completely blown up (especially in summer, or anyone playing world of warcraft etc...LOL), and if you live in any of the 15 states that are over 15c it also blows up. No idea how bad it is out of USA, but it can't be much better in many places.

    People need to consider the TCO when buying a vid card. Also 295x2 does not win every benchmark vs. titan, not to mention you can OC it, be far less watts and beat them in a LOT of games. Most review sites say its got far better gameplay even if fps say otherwise in some games (fluid). You are counting on AMD to make profile after profile for years for it, and they don't have the money to even get that done for all the games tested at a dozen review sites (about 30 games that don't change for ages in testbeds). So what are the odds they'll hit games that aren't tested if they can't get ~30 done?
    Reply
  • Impulses - Saturday, April 18, 2015 - link

    Yeah I dunno why many are seemingly all too happy to start dancing on AMD's grave, NV isn't Intel, they don't have big client demands or outside forces like ARM closing in (as far as GPUs goes). The desktop GPU market will go to hell in a handbasket if AMD goes poof. Maybe not immediately, but it'll happen. Reply
  • mgilbert - Friday, April 17, 2015 - link

    Intel and Nvidia products are going to get very expensive if AMD fails... Reply
  • TallestJon96 - Friday, April 17, 2015 - link

    Yeah, the only AMD product I've bought is xbone and ps4, but I want them in business to compete with Intel and Nvidia. Intel barely does anything for CPUs anymore, imagine what we'd get without AMD chasing after performance. Reply
  • FlushedBubblyJock - Saturday, April 25, 2015 - link

    What ? Have you tried a G3258 intel haswell for 50 bucks ?

    AMD can't beat a 50 dollar haswell, but it's intel who does nothing... R O F L

    Tell me lies, tell me sweet fanboy lies, ya tell me lies won't you tell me lies
    Reply

Log in

Don't have an account? Sign up now