The only quarters where AMD is "profitable" are the product of exaggerating loses slightly in other quarters. AMD has lost billions the last decade and there is no outlook it will ever get any better.
So the big question is - why bother running it? The only thing AMDs existence achieves is to theoretically undermine Intel's de-facto monopoly. It only makes Intel look less of a monopolist, but it doesn't push Intel to innovate or improve the value of its products.
The last time AMD stood a chance to turn into a profitable company, it was destroyed from within, overpaying for ATi, just to have AMD+ATi worth a small fraction of what it paid for ATi, which led to losing its fabs, further decreasing profit margins. And now AMD is stuck forever right where Intel wants it to be - an impotent pseudo competitor.
"The only thing AMDs existence achieves is to theoretically undermine Intel's de-facto monopoly. It only makes Intel look less of a monopolist, but it doesn't push Intel to innovate or improve the value of its products."
To a lesser extent, the same could be said about their relationship with Nvidia.
I find this strange, there is nothing in the product lines of both companies to justify those figures. I myself mostly run radeons, but I don't really care about gaming, radeons offer better FP64 performance for the $.
Pretty simple, nVIDIA released Maxwell 2 before AMD released Fiji. Maxwell 2, in particular GTX 970, was a killer price/performance point for people who had been waiting to upgrade. By the time Fiji arrived, not only was it too expensive for what it brought to the table, its market had pretty much dried up. Oh, and there's the fact that nVIDIA has a competent marketing department.
AMD's consistent recycling of old products instead of introducing new ones, and hyping their new products while lying about their features (see: "Fury X is a great overclocker") doesn't endear consumers to them.
It's interesting that the GTX 970 is so popular when the 290 has been a superior price/performance pick for the majority of the time since the 970 dropped. And to think the 960 actually sold when there was usually a $250 290. People are bad at doing their research and all too willing to do companies' marketing for them.
You have to remember, that the Steam hardware survey does not represent a random group of samples from the overall market, but rather those who are biased toward gaming (by virtue of having Steam installed). Also, IIRC, participation in the Steam survey is voluntary - this further narrows the group by essentially filtering out those who do not care so much about the PC hardware market, to want to be part of the statistics and let Valve scan their PCs.
So the only conclusion you can really draw from that, is that among people who are gamers and knowledgeable about tech, AMD has a slightly higher approval rating than in the overall market. This would fit well with AMDs lack of effective marketing, and with the (mostly untrue) perception that their cards and drivers are more finicky and difficult to live with for less tech-savvy people.
On a side note, I assume that the 19%-81% ratio mentioned above simply ignores all the Intel GPU equipped PCs out there, while the Steam survey of course doesn't. This means, that if we convert the numbers in a similar way, the ratio between AMD and nVidia users among the Steam userbase is actually closer to 35%-65%.
"You have to remember, that the Steam hardware survey does not represent a random group of samples from the overall market, but rather those who are biased toward gaming (by virtue of having Steam installed)."
If you aren't gaming and can't be bothered to install Steam, then built-in graphics are good-enough. (High end CAD machines are a *tiny* minority). If you are looking at real revenue (at least for the Radeon division), I'd look to Steam results.
The worst part, and it counts for Nvidia was well, is that there is little reason to upgrade your .28nm parts. You might get more transistors, you might get a little better design, you might pull less watts (with a Nvidia system), but you don't get the big gains you expected on old generation changes. .18nm might go faster and cooler, but don't expect cheaper: Moore's law is more like a guideline now.
Market share of installed GPUs is different to Market share of Sales this quarter.
If historically it's always been 50/50, and 10% of people buy a new GPU each year, even if you have 100% of new sales, it will take a long time for the installed base to be 100% yours.
This is a subset of all computers out there. Not every computer has Steam installed. Plus there are quite a few Nvidia mobile GPUs represented in that list and they for sure have a dominant position in laptops. Though the numbers are interesting, but looking at the details it is much more muddy than any of the arguments presented. There are a lot of crap cards in the numbers as opposed to simply talking about what's selling from both Red and Green's big guns $200+.
81% of the DISCRETE market is what he meant. NV doesn't make apus for PC which steam counts for both Intel and AMD. Thought that may change soon with ARM PC's (fully loaded like and Intel/amd PC, big drives, discrete gpu, 16-32GB mem, 500-1000psu etc etc) at some point I hope.
Unless AMD made their ZEN cpu the size of Intel's current CPU+GPU sides, there isn't much point in hyping it up. It needs to be big enough so Intel's only response is to make a brand new design that is PURE CPU. for desktops. If they went anything under 50% bigger than Intel's cpu side they'll fail and be priced down while Intel simply plops down 2 quads and sells for the same price (probably at first just turning off gpu's to get watts in the 100-125w range). My devils canyon hits 68w or so with gpu off, so if AMD makes a huge chip (gpus on top skylake chips are larger than the cpu side now), they can theoretically smoke Intel, and anything intel can slap together for a few years like ~2000. The difference being, they can now produce FAR more than a meager 20% they were stuck at before. This time you have GF/Samsung/TSMC that can all put out your chips if desired. If they went huge they could possibly build 50% of the market or more, if they went anything less than 125% I think they won't make much because Intel will just plop a cherry picked core out, and drop prices on all others $50-100 to make sure AMD collects nothing while they come up with a REAL answer. But you can't price down something that is beating your top end by 25-75% across the board in games. That takes a new design. We've seen the 8 core vs. 4 core (see AMD and how that worked out, loses everything but corner cases and runs hot, massive watts etc), and Intel would end up in that boat for a while. But again, ONLY if AMD went HUGE for a TRUE KING, like dirk said they needed when he left in 2011. Get a king or you have ZERO pricing power. If they beat Intel 20% or more in everything I'd gladly pay $50 on top of Intel's top chip at the time on desktop (say $400, where skylake running $350). But if they tie, and die is not larger, I'll go Intel again. Management has had 5 years to follow what Dirk said. Being 2nd and "OK, but cheaper" just loses money, and never has pricing power to MAKE wads of cash. It never gains market share either. Perform so you can CHARGE accordingly or go bankrupt in this business.
I hope they went big. We still don't know, only have IPC info, but not how big it is, speeds, cache, watts etc. That one stat doesn't say much and could even be shooting low on many things. You definitely don't want another bullsnozer...LOL. 300mm^2 sounds good :) But my guess is management is still dumb and went with about the same as Intel's cpu side, so they can keep putting gpus with it reasonably. I was really hoping two tech designs (because going my way for cpu, HUGE, would make a APU kind of out). Maybe a high clocked dual with 4 threads would be ok. It, can't be much worse than a current quad apu from AMD if it's higher clocked and 40% IPC improvement anyway). Go big or go home AMD. 2nd place just=30% margins v.s NV/Intel 56-64% margins. Learn/adapt or die.
Wow....if AMD followed this FUD, they'd run aground for sure. Yes, they should make a beast of a CPU with high IPC to beat the i7-6xxxK CPUs so they're talk about them. But, an extra 25-75% in GAMES??? WTH do you think CPUs do? Render polygons? Yes, the weak IPC of AMD chips makes them shortfall Intel by a few FPS by default, but your CPU won't double your FPS. Sorry. The market is driving toward ultra-lowpower CPUs, not behemoth power monsters. While there is a significant margin to be had by that top 1% of CPU purchases, it's the mainstream stuff that gives you revenue. AMD's current problem is that even with a lower cost than Intel, they still suffer an significant IPC loss which they have to make up by OCing the chip and STILL have less FPS for the same gfx card. This costs wattage, making the system overall less efficient (power-wise, noise-wise, size-wise). Being able to use the new process node similar to Intel's, they'll once again have the opportunity to compete in performance per watt. A flagship that can trounce Intel by simply throwing raw cores at a problem may help, but they need to win in single-thread as well.
Because Nvidia has a marketing team, AMD really doesn't. On the GPU side, AMD offers a better , or at least a comparable,card at every price point. But people still buy Nvidia because they get the marketing slammed in their face.
The numbers you are referring to are for one quarter (3 months). That typically is slanted based on product releases. Earlier in the same page it indicated Intel at 71%, Nvidia at 15%, and AMD at 11% in GPU percentages. I think these numbers are more representative. than you 80% Nvidia number. Nvidia having 4 points of market share more sounds true.
No, they are not. They compete by coupling a brute force approach to low prices. That only works as a long term strategy if it leads to an increase in market share. And it doesn't. Therefore the company is loosing money.
As for competing for market share by appealing to consumers, NVDIA cards tend to be more practical. The brute force approach of AMD causes inconsistencies in the frame rate and higher power usage.
I think the new focus on smaller cards is really smart though, and hopefully AMD has the technology in 2016 to really make that concept work. AMD needs to be different from NVDIA in order to compete in a meaningful way.
Just as a heads up, the only current solution that really has serious frame rate inconsistencies is SLI, while CF and single cards from both companies are quite smooth by now.
I'm referring to the fact that AMD cards have a tendency to have lower minimum frame rates during parts of a game. I'm not referring to micro-stuttering.
NVDIA cards are better optimized for the settings that people actually use.
"Who has interest in funding AMD... except Intel?" All those that have already invested millions in it... at some point you prefer to keep putting money in hoping someone will buy the company and give you some benefit or at least that the company will eventually be profitable rather than let the company sink and end up with a huge loss.
Well they are a 10k employee company so all the numbers get big, per employee per quarter it's about $100k revenue, $30k gross profit, $10k loss. Not that absurd numbers for a company in trouble. And AMD used to have a ton of assets in factories and facilities, it's like selling your house/car and rent/lease it back which would give you cash to burn for a while even though you lose money every month. The $371m deal they made with Nantong for the chip assembly and test facilities should keep them floating a few quarters more. But they had better turn their core business around very soon, because that was the last major physical asset they own.
They have about 5-7 years (not sure exactly) until bankruptcy, because that's when their loans are due. They are in an overall net liability position (meaning if they were to declare bankruptcy, anyone they owe money to wouldn't get all of their money back), but no one cares for now until the $2b in loans are due, which starts in about 7 years. And even then they will probably try to extend or refinance the loans, as there's almost no way they will be able to repay them.
Cash wise they have enough cash to keep going for a while. It's the big loan balance that will hit them.
You can't say they have 5-7 years then say "Cash wise they heave enough to keep going for a while". You know that "a while" is a lot less, perhaps a year. If you take away the cash needed for cash flow to just keep running the company then that doesn't leave much, and at their current rate of losses that is going fast.
AMD has the cash to continue to stay in operation until around mid-2018, when about $600 Million in debt comes due. An addition $400 Million comes due in 2019. For a company currently only valuated at $1.5 Billion, those debt payment will equal 66% of the companies total valuation.
Point being, AMD needs to bring in an operating profit of $1 Billion over the next 3-4 years just to cover those payments.
The danger for AMD, is if Zen doesn't turn the financials around, the investors could start dumping the stock as those payments approach, making it impossible for AMD to raise enough cash to make those payments. The investors are already VERY skittish, and in my opinion, Zen is AMDs last shot.
AMD would sell itself or the GPU division long before running out of cash. Zen is fundamental and does need to be ok but it doesn't have to be great . Intel has been offering less and less for more , they slowed the process shrink pace, they need to keep financials up to keep the stock price up so they are vulnerable. Intel can't afford too many price cuts or their stock price could drop 50-75%. And they have just 1 core going from 5W SoC to server. So AMD doesn't have to be good in all markets and Intel has given them a lot of extra room compared to 5-10 years ago. As for your math and concerns , they are overblown. AMD has 785 million in cash, they are getting 320 mil in Q2 from a sale and they have 2.26B debt but what matter more is that if they are doing okish, it won't be difficult to get more debt. They are projecting " to return to non-GAAP operating profitability in 2H 2016"and " to generate positive free cash flow from operations in 2016." Long term, the PC is dead and AMD needs to address glasses and robots in consumer or just server if they leave consumer. Intel seems to see robots as their savior long term.
Intel won't let AMD die. It would mean getting broken up into several smaller companies. Intel will probably do something to give AMD money if things get really bad. Maybe just giving AMD an "anonymous" donation...
Either way, Zen is probably going to be make it or break it for AMD. The only way they can become profitable is to have a strong (relatively speaking) CPU market that can actually churn out a profit.
Because its the EBIT that defines the real success of your company towards your owners and investors, the "profit" or "loss" merely defines how much taxes you plan to pay.
Your table is presenting this quarter's GM as 23% instead of 30%.
I think Devander needs to learn that accrual accounting expects estimations of things like taxes to be made throughout the year instead of waiting until the end of the year to book all the tax in Q4. He's been there fifteen years so he should have at least a little bit of insight into what drives the tax expense throughout the year. Or maybe he doesn't.
You guys missed a lot. First they posted a new roadmap, kinda. About Zen they say high core count and multi-threading. On the server side they add about Zen - Disruptive Memory Bandwidth and High Native I/O Capacity. They also mention in 2017 a server APU - Multi-Teraflops for HPC and Workstation with Transformational Memory Architecture and Scale-Up Graphics Performance - w/e that means, maybe HBM and CF with discrete GPUs...
In the call they said Zen server is for 2017 and they claimed to have secured several design wins with Zen in server. Polaris was always mentioned as shipping mid-year so nothing new there. At some point they said more than 40% IPC gain for Zen. The new semi-custom win ramps in the second half , likely Nintendo , could be a big gain if Nintendo goes with a 4K capable GPU on finFET and maybe includes some VR glasses with a 500$ console.
As for the numbers , it's the second quarter the computing and graphics segment grows after tanking hard for quite a while and that's good. They also said they expect the PC market to be down low to mid single digits this year and they expect their units to be up.
This year Polaris, Zen in desktop and Nintendo are opportunities, next year Zen in server and APU and maybe a little ARM server. Not too optimistic about Zen in single threaded, doubt it will beat Intel but they can offer decent perf at decent power at much better price per core(no GPU). If they beat Intel and offer 2-3 times more cores, even better but they can survive with less.
All they need to match Intel's single threaded performance is to make a chip with comparable throughput. AMD have historically had process disadvantage, forcing them to skimp on the transistors. In the pentium 4 days they had arch advantage which mitigated that, but today intel has good arch.
When talking about HPC APUs, "scale-up graphics performance" could also mean crossfire/shared compute opportunities with multiple APUs on the same board.
It is doable though, the easiest way to go and would give them a big advantage over the others. VR glasses are just a screen and not much else so cost wise it is very doable- small screens with no touch layer and cover glass are cheap no matter what lies some might be telling the public. Also do remember hat in consoles folks don't aim to make money with the hardware and that Nintendo was the first one do push gesture based controller. ofc not saying the glasses would be 4k since you need to run those at high FPS but they don't need 4k glasses now.
"AMD attributes the seasonally lower sales of semi-custom SoCs as the reason for the drop from last quarter, and the year-over-year drop is due to lower game console royalties, and lower server and embedded sales."
I'm mystified - how did AMD take in less console royalties even as console sales boomed in 2015?
Sony and MS would buy those CPUs the quarter before to have systems built Christmas and sell hard. We can likely expect a small boost in the semi-custom business in 1Q 2016. Manufacturing time for those consoles would dictate you want them built in the end of the summer to be ready and sitting on shelves in November.
Seasonal buying patterns may account the drop in revenue in Q4 2015 versus Q3 2015 but it doesn't explain why Q4 2015 revenue ($488 million) was also lower than Q4 2014 ($577 million). The year-ago comparison largely takes the seasonal fluctuations out of the picture. I am still scratching my head how semi-custom and console revenue in the fourth quarter of 2015 ended up 15% lower than the fourth quarter of 2014.
That's on console royalties (older gen consoles that have declining sales) ,server and embedded. Server must have been some 100 million a year ago and likely dropped a lot since. Plus maybe a bit of a shift in semi-custom timing for orders. It does seem like too much and the funnier part is that revenue outlook for Q1 is down 14% , with PC dropping less than that. 2015 vs 2014 in semi-custom was up. Units way up, ASP plenty down but revenue was up. The entire segment was down on year on everything else being down.
The segment that semi custom is in is mixed with other sales channels such as embedded, server chips, and probably their memory sales. Also the amount that AMD was paid per console chipset was lower in 2015 than it was in 2014 as part of the pricing agreement with Sony and Microsoft. It will be slightly lower in 2016 as well, however, it is expected that both Sony and Microsoft will be lowering their retail prices this year to drive volume. AMD anticipates the higher volume to overcome the lower selling price they will be receiving to drive higher overall revenues from console sales.
In short, the drop in overall sales cannot be blamed solely on console sales. I'm sure a big chunk of it is due to lower Opteron sales year over year.
That may have been true if it had happened 10 years ago, but now, both have plenty of competition from mobile. Intel may be the only game in town in desktop, but ARM (through Qualcomm, Samsung, Apple, Mediatek...) is selling more designs per year than intel. Similarly, in mobile, Nvidia is competing (and lost, really) with Qualcomm (Adreno, former ATI mobile group), ARM (in Samsung and Mediatek chips), PowerVR (Apple chips) and others (vivando...).
So, there is a case to make for both, but it's unlikely they would be split. Moreover, the US administration has shown great reluctance in applying antitrust lately...
I think Zen is too little too late. The only thing it will do is force Intel to lower its price, Xeon-D, and Xeon E5. I just do not see AMD's way out, and their Radeon Graphics are doing shit.
May be they should be acquired by Intel. On a Financial Level, no one is interested to pick up a downtrend going company with lots of debt and little to no upside.
Intel could pick it up for less then $4B. Just Why hasn't Intel made the move yet?
( Monopoly doesn't really count as argument anymore, Desktop, Laptop market is shrinking. )
In tel has literally nothing to gain from buying AMD. Few companies would, as their biggest asset, their x86 license, is officially non-transferable. The rest of their assets (designs for both their CPU and GPU) have little value to anyone else, as those are very competitive businesses and a slew of underpowered competitors exist from mobile (which is the growing market).
They lost 84 cents per share in 2015. That's crazy, that's almost half their stock price. The remaining ATI engineers in Toronto can probably pool their money together and buy the company back and rename it ATI.
I'm glad anandtech kept posting those ridiculous articles about superstar ABC or XYZ joining AMD, helped get the stock price back above $2 long enough for me to dump it. The company has some cool technology but the crappy management and brain drain of talent has sunk the company.
Its bad for all engineers in the industry when a company fails but the market place has no mercy.
AMD keeps losing money but I guess its not that much each quarter so they can just drag on for another year hoping for the next CPU or GPU to save them. I think at some point they would have to sell out and get what they can but hey somehow RIM is still around these days.
Everyone saw this coming honestly. AMD has stopped innovating years ago. They rely on market hype and recycle old products. They completely abandoned desktop CPU market in favor of just focusing on APU's for the mobile market, GPU's and Server CPU's. Since then Intel has made leaps and bounds in SoC R&D, allowing their mobile chips to not only outperform AMD APU's in both CPU power and integrated on-die GPU power as well. Nvidia is still updating GPU architecture (Pascal will be introduced this year) thus broadening the gap between the two. AMD's R300 series was supposed to be the next best thing, going toe to toe with Nvidias GTX 900 series and then some, but once the smoke cleared, they were just rebrands with almost 0 performance or innovation gains.
AMD is in financial distress due to lack of innovation and no real desire or means to fully compete anymore. They are milking the contracts they have with MS and Sony for the APU’s put into the new consoles, but other than that they are in real trouble. Eventually Fed will step in and funnel them money (if they aren’t already) to just keep them afloat on paper, as not to create an Intel/Nvidia monopoly, which by default is already happening. I would gladly buy AMD products as I have in the past, when they can go toe to toe with current offerings for a lesser price.
I wish Nvidia would stop innovating like AMD, they'd probably come up with something decent. The biggest innovation we've had from them is pulling out the FP64 and compute flexibility to free up power for gaming and selling FP16 capabilities. It's not AMD who's going to be putting its competitors memory onto its next generation chips designed for an API pioneered by its competitor, now is it?
Incidentally, if anything it looks like Polaris, which had working samples at CES is going to beat Pascal, which had 980M modules used in its place, and that (surprise surprise) has an updated GPU architecture.
I have a feeling that last sentence is their biggest problem. If they're even they need to have a better price as well.
HBM is an industry standard, same as HBM2. AMD was the first one to implement it with their partner, Hynix. Now, what did that give them?
The truth is that both AMD and Nvidia were hurt, badly, by the stagnation of foundry process. Nvidia has been on 28nm since the GF680, in march 2012 (similar to AM with the 7970 in january 2012). Seeing no end to it, Nvidia sacrificed 64 bit capabilities for better gaming performance. AMD decided to skip a refresh outside of Fury.
This year will be the first new process introduced in GPU since 2012. It should give us great improvements (from FinFET and smaller dies).
Your post assumes AMD doesn't have a single product on their roadmap. In addition, as regards your last sentence, AMD do actually have products that undercut equally performing NVIDIA offerings. The 360 isn't worth it over the 950 (which is why I bought the latter), but despite the power consumption difference, AMD are certainly competitive as regards their graphics products more often than not.
It's no accident that their server market share has plummeted, more so than their desktop CPU market share, due to Bulldozer. Doesn't it stand to reason that a good product with Zen would drag them back up from the depths in both markets, even if just a little? It's not as if AMD can't develop a good architecture, they just had to make the most of their failed experiment in order to actually have some income whilst they developed a replacement. It takes four or five years to bring out a completely new architecture.
It is interesting to note that the last time AMD was really competitive in the CPU market was, um, 12 years ago (Feb 2004) when the newly released Prescott Pentium 4 CPU's were beaten in terms of gaming performance by the then "Athlon 64" CPU"s.
In the space of these ensuing 12 years, AMD have failed to produce a CPU that like the "Athlon 64", would give INtel a run for their money.
It seems unlikely they will ever succeed, and that is bad for the end user.
I've made a ton of money by buying AMD stock when it's down and selling when it's up. I buy their CPUs and GPUs because they are less expensive and do what I want. And remember Nexgen? I sold their motherboards before they were absorbed by AMD. Still have a P100 or whatever. Thank you that it's America and we're not stuck with Intelisma!
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
69 Comments
Back to Article
Wreckage - Tuesday, January 19, 2016 - link
I'm always amazed that they can lose $100,000,000+ every few months and stay in business.ddriver - Tuesday, January 19, 2016 - link
The only quarters where AMD is "profitable" are the product of exaggerating loses slightly in other quarters. AMD has lost billions the last decade and there is no outlook it will ever get any better.So the big question is - why bother running it? The only thing AMDs existence achieves is to theoretically undermine Intel's de-facto monopoly. It only makes Intel look less of a monopolist, but it doesn't push Intel to innovate or improve the value of its products.
The last time AMD stood a chance to turn into a profitable company, it was destroyed from within, overpaying for ATi, just to have AMD+ATi worth a small fraction of what it paid for ATi, which led to losing its fabs, further decreasing profit margins. And now AMD is stuck forever right where Intel wants it to be - an impotent pseudo competitor.
Who has interest in funding AMD... except Intel?
DigitalFreak - Tuesday, January 19, 2016 - link
"The only thing AMDs existence achieves is to theoretically undermine Intel's de-facto monopoly. It only makes Intel look less of a monopolist, but it doesn't push Intel to innovate or improve the value of its products."To a lesser extent, the same could be said about their relationship with Nvidia.
ddriver - Tuesday, January 19, 2016 - link
AMD is fairly competitive to nvidia.HighTech4US - Tuesday, January 19, 2016 - link
Yet 8 out of 10 buy Nvidia.Nvidia 81.1% market share vs 18.8% for AMD.
https://forum.beyond3d.com/posts/1881582
ddriver - Tuesday, January 19, 2016 - link
I find this strange, there is nothing in the product lines of both companies to justify those figures. I myself mostly run radeons, but I don't really care about gaming, radeons offer better FP64 performance for the $.JoeMonco - Tuesday, January 19, 2016 - link
And next to nothing uses FP64 so that better performance matters to next to no average consumer.The_Assimilator - Wednesday, January 20, 2016 - link
Pretty simple, nVIDIA released Maxwell 2 before AMD released Fiji. Maxwell 2, in particular GTX 970, was a killer price/performance point for people who had been waiting to upgrade. By the time Fiji arrived, not only was it too expensive for what it brought to the table, its market had pretty much dried up. Oh, and there's the fact that nVIDIA has a competent marketing department.AMD's consistent recycling of old products instead of introducing new ones, and hyping their new products while lying about their features (see: "Fury X is a great overclocker") doesn't endear consumers to them.
xthetenth - Wednesday, January 20, 2016 - link
It's interesting that the GTX 970 is so popular when the 290 has been a superior price/performance pick for the majority of the time since the 970 dropped. And to think the 960 actually sold when there was usually a $250 290. People are bad at doing their research and all too willing to do companies' marketing for them.Notmyusualid - Wednesday, January 20, 2016 - link
Sorry to nit-pick, but on the Steam hardware survey:http://store.steampowered.com/hwsurvey?platform=pc
First graph you see, last 6 months data, 26.23% used ATI (AMD). 54.61% for Nvidia.
szimm - Wednesday, January 20, 2016 - link
You have to remember, that the Steam hardware survey does not represent a random group of samples from the overall market, but rather those who are biased toward gaming (by virtue of having Steam installed). Also, IIRC, participation in the Steam survey is voluntary - this further narrows the group by essentially filtering out those who do not care so much about the PC hardware market, to want to be part of the statistics and let Valve scan their PCs.So the only conclusion you can really draw from that, is that among people who are gamers and knowledgeable about tech, AMD has a slightly higher approval rating than in the overall market. This would fit well with AMDs lack of effective marketing, and with the (mostly untrue) perception that their cards and drivers are more finicky and difficult to live with for less tech-savvy people.
On a side note, I assume that the 19%-81% ratio mentioned above simply ignores all the Intel GPU equipped PCs out there, while the Steam survey of course doesn't. This means, that if we convert the numbers in a similar way, the ratio between AMD and nVidia users among the Steam userbase is actually closer to 35%-65%.
beck2050 - Thursday, January 21, 2016 - link
Sales in the last quarter of DISCRETE gpus were 81 to 19. Steam is not a sales report.wumpus - Tuesday, January 26, 2016 - link
"You have to remember, that the Steam hardware survey does not represent a random group of samples from the overall market, but rather those who are biased toward gaming (by virtue of having Steam installed)."If you aren't gaming and can't be bothered to install Steam, then built-in graphics are good-enough. (High end CAD machines are a *tiny* minority). If you are looking at real revenue (at least for the Radeon division), I'd look to Steam results.
The worst part, and it counts for Nvidia was well, is that there is little reason to upgrade your .28nm parts. You might get more transistors, you might get a little better design, you might pull less watts (with a Nvidia system), but you don't get the big gains you expected on old generation changes. .18nm might go faster and cooler, but don't expect cheaper: Moore's law is more like a guideline now.
rtho782 - Wednesday, January 20, 2016 - link
Market share of installed GPUs is different to Market share of Sales this quarter.If historically it's always been 50/50, and 10% of people buy a new GPU each year, even if you have 100% of new sales, it will take a long time for the installed base to be 100% yours.
eanazag - Wednesday, January 20, 2016 - link
This is a subset of all computers out there. Not every computer has Steam installed. Plus there are quite a few Nvidia mobile GPUs represented in that list and they for sure have a dominant position in laptops. Though the numbers are interesting, but looking at the details it is much more muddy than any of the arguments presented. There are a lot of crap cards in the numbers as opposed to simply talking about what's selling from both Red and Green's big guns $200+.TheJian - Thursday, January 21, 2016 - link
81% of the DISCRETE market is what he meant. NV doesn't make apus for PC which steam counts for both Intel and AMD. Thought that may change soon with ARM PC's (fully loaded like and Intel/amd PC, big drives, discrete gpu, 16-32GB mem, 500-1000psu etc etc) at some point I hope.Unless AMD made their ZEN cpu the size of Intel's current CPU+GPU sides, there isn't much point in hyping it up. It needs to be big enough so Intel's only response is to make a brand new design that is PURE CPU. for desktops. If they went anything under 50% bigger than Intel's cpu side they'll fail and be priced down while Intel simply plops down 2 quads and sells for the same price (probably at first just turning off gpu's to get watts in the 100-125w range). My devils canyon hits 68w or so with gpu off, so if AMD makes a huge chip (gpus on top skylake chips are larger than the cpu side now), they can theoretically smoke Intel, and anything intel can slap together for a few years like ~2000. The difference being, they can now produce FAR more than a meager 20% they were stuck at before. This time you have GF/Samsung/TSMC that can all put out your chips if desired. If they went huge they could possibly build 50% of the market or more, if they went anything less than 125% I think they won't make much because Intel will just plop a cherry picked core out, and drop prices on all others $50-100 to make sure AMD collects nothing while they come up with a REAL answer. But you can't price down something that is beating your top end by 25-75% across the board in games. That takes a new design. We've seen the 8 core vs. 4 core (see AMD and how that worked out, loses everything but corner cases and runs hot, massive watts etc), and Intel would end up in that boat for a while. But again, ONLY if AMD went HUGE for a TRUE KING, like dirk said they needed when he left in 2011. Get a king or you have ZERO pricing power. If they beat Intel 20% or more in everything I'd gladly pay $50 on top of Intel's top chip at the time on desktop (say $400, where skylake running $350). But if they tie, and die is not larger, I'll go Intel again. Management has had 5 years to follow what Dirk said. Being 2nd and "OK, but cheaper" just loses money, and never has pricing power to MAKE wads of cash. It never gains market share either. Perform so you can CHARGE accordingly or go bankrupt in this business.
I hope they went big. We still don't know, only have IPC info, but not how big it is, speeds, cache, watts etc. That one stat doesn't say much and could even be shooting low on many things. You definitely don't want another bullsnozer...LOL. 300mm^2 sounds good :) But my guess is management is still dumb and went with about the same as Intel's cpu side, so they can keep putting gpus with it reasonably. I was really hoping two tech designs (because going my way for cpu, HUGE, would make a APU kind of out). Maybe a high clocked dual with 4 threads would be ok. It, can't be much worse than a current quad apu from AMD if it's higher clocked and 40% IPC improvement anyway). Go big or go home AMD. 2nd place just=30% margins v.s NV/Intel 56-64% margins. Learn/adapt or die.
Ammaross - Friday, January 22, 2016 - link
Wow....if AMD followed this FUD, they'd run aground for sure.Yes, they should make a beast of a CPU with high IPC to beat the i7-6xxxK CPUs so they're talk about them. But, an extra 25-75% in GAMES??? WTH do you think CPUs do? Render polygons? Yes, the weak IPC of AMD chips makes them shortfall Intel by a few FPS by default, but your CPU won't double your FPS. Sorry.
The market is driving toward ultra-lowpower CPUs, not behemoth power monsters. While there is a significant margin to be had by that top 1% of CPU purchases, it's the mainstream stuff that gives you revenue. AMD's current problem is that even with a lower cost than Intel, they still suffer an significant IPC loss which they have to make up by OCing the chip and STILL have less FPS for the same gfx card. This costs wattage, making the system overall less efficient (power-wise, noise-wise, size-wise). Being able to use the new process node similar to Intel's, they'll once again have the opportunity to compete in performance per watt. A flagship that can trounce Intel by simply throwing raw cores at a problem may help, but they need to win in single-thread as well.
xdrol - Monday, January 25, 2016 - link
Steam survey is about current owners, not new owners (aka sales) of different chips..Bateluer - Wednesday, January 20, 2016 - link
Because Nvidia has a marketing team, AMD really doesn't. On the GPU side, AMD offers a better , or at least a comparable,card at every price point. But people still buy Nvidia because they get the marketing slammed in their face.eanazag - Wednesday, January 20, 2016 - link
The numbers you are referring to are for one quarter (3 months). That typically is slanted based on product releases. Earlier in the same page it indicated Intel at 71%, Nvidia at 15%, and AMD at 11% in GPU percentages. I think these numbers are more representative. than you 80% Nvidia number. Nvidia having 4 points of market share more sounds true.HighTech4US - Thursday, January 21, 2016 - link
AMD has been losing large amounts of market share to Nvidia for well over a year.Look at the chart here:
https://forum.beyond3d.com/posts/1873668
AMD has lost over 50% of it's market share and has dropped from 37.9% to 18% whereas Nvidia has gone from 62% to 81.9%
So no this is not just a "one quarter" or a slant.
RafaelHerschel - Wednesday, January 20, 2016 - link
No, they are not. They compete by coupling a brute force approach to low prices. That only works as a long term strategy if it leads to an increase in market share. And it doesn't. Therefore the company is loosing money.As for competing for market share by appealing to consumers, NVDIA cards tend to be more practical. The brute force approach of AMD causes inconsistencies in the frame rate and higher power usage.
I think the new focus on smaller cards is really smart though, and hopefully AMD has the technology in 2016 to really make that concept work. AMD needs to be different from NVDIA in order to compete in a meaningful way.
xthetenth - Wednesday, January 20, 2016 - link
Just as a heads up, the only current solution that really has serious frame rate inconsistencies is SLI, while CF and single cards from both companies are quite smooth by now.RafaelHerschel - Wednesday, January 20, 2016 - link
I'm referring to the fact that AMD cards have a tendency to have lower minimum frame rates during parts of a game. I'm not referring to micro-stuttering.NVDIA cards are better optimized for the settings that people actually use.
Strunf - Wednesday, January 20, 2016 - link
"Who has interest in funding AMD... except Intel?"All those that have already invested millions in it... at some point you prefer to keep putting money in hoping someone will buy the company and give you some benefit or at least that the company will eventually be profitable rather than let the company sink and end up with a huge loss.
Kjella - Wednesday, January 20, 2016 - link
Well they are a 10k employee company so all the numbers get big, per employee per quarter it's about $100k revenue, $30k gross profit, $10k loss. Not that absurd numbers for a company in trouble. And AMD used to have a ton of assets in factories and facilities, it's like selling your house/car and rent/lease it back which would give you cash to burn for a while even though you lose money every month. The $371m deal they made with Nantong for the chip assembly and test facilities should keep them floating a few quarters more. But they had better turn their core business around very soon, because that was the last major physical asset they own.MadAd - Wednesday, January 20, 2016 - link
So after that, in about a year will that make AMD about ready to announce bankruptcy?Lonyo - Wednesday, January 20, 2016 - link
They have about 5-7 years (not sure exactly) until bankruptcy, because that's when their loans are due. They are in an overall net liability position (meaning if they were to declare bankruptcy, anyone they owe money to wouldn't get all of their money back), but no one cares for now until the $2b in loans are due, which starts in about 7 years. And even then they will probably try to extend or refinance the loans, as there's almost no way they will be able to repay them.Cash wise they have enough cash to keep going for a while. It's the big loan balance that will hit them.
Dribble - Wednesday, January 20, 2016 - link
You can't say they have 5-7 years then say "Cash wise they heave enough to keep going for a while". You know that "a while" is a lot less, perhaps a year. If you take away the cash needed for cash flow to just keep running the company then that doesn't leave much, and at their current rate of losses that is going fast.gamerk2 - Wednesday, January 20, 2016 - link
AMD has the cash to continue to stay in operation until around mid-2018, when about $600 Million in debt comes due. An addition $400 Million comes due in 2019. For a company currently only valuated at $1.5 Billion, those debt payment will equal 66% of the companies total valuation.Point being, AMD needs to bring in an operating profit of $1 Billion over the next 3-4 years just to cover those payments.
The danger for AMD, is if Zen doesn't turn the financials around, the investors could start dumping the stock as those payments approach, making it impossible for AMD to raise enough cash to make those payments. The investors are already VERY skittish, and in my opinion, Zen is AMDs last shot.
jjj - Wednesday, January 20, 2016 - link
AMD would sell itself or the GPU division long before running out of cash.Zen is fundamental and does need to be ok but it doesn't have to be great . Intel has been offering less and less for more , they slowed the process shrink pace, they need to keep financials up to keep the stock price up so they are vulnerable. Intel can't afford too many price cuts or their stock price could drop 50-75%. And they have just 1 core going from 5W SoC to server. So AMD doesn't have to be good in all markets and Intel has given them a lot of extra room compared to 5-10 years ago.
As for your math and concerns , they are overblown. AMD has 785 million in cash, they are getting 320 mil in Q2 from a sale and they have 2.26B debt but what matter more is that if they are doing okish, it won't be difficult to get more debt. They are projecting " to return to non-GAAP operating profitability in 2H 2016"and " to generate positive free cash flow from operations in 2016."
Long term, the PC is dead and AMD needs to address glasses and robots in consumer or just server if they leave consumer. Intel seems to see robots as their savior long term.
SlyNine - Wednesday, January 20, 2016 - link
Pc is dead... believe it when I see it.Ammaross - Friday, January 22, 2016 - link
"Pc is dead... believe it when I see it."Because an iPad is a viable computing platform for anyone that does more than Facebook... FFS
Aspiring Techie - Wednesday, January 20, 2016 - link
Intel won't let AMD die. It would mean getting broken up into several smaller companies. Intel will probably do something to give AMD money if things get really bad. Maybe just giving AMD an "anonymous" donation...Either way, Zen is probably going to be make it or break it for AMD. The only way they can become profitable is to have a strong (relatively speaking) CPU market that can actually churn out a profit.
ShieTar - Thursday, January 21, 2016 - link
Because its the EBIT that defines the real success of your company towards your owners and investors, the "profit" or "loss" merely defines how much taxes you plan to pay.MartinT - Tuesday, January 19, 2016 - link
Makes you wonder how their stock close to doubled since late last summer, only to now be back to almost where it started.All in the fundamentals, right?
HighTech4US - Tuesday, January 19, 2016 - link
The run-up was pure speculation.The run-down is because reality set in.
Intel999 - Tuesday, January 19, 2016 - link
@authorYour table is presenting this quarter's GM as 23% instead of 30%.
I think Devander needs to learn that accrual accounting expects estimations of things like taxes to be made throughout the year instead of waiting until the end of the year to book all the tax in Q4. He's been there fifteen years so he should have at least a little bit of insight into what drives the tax expense throughout the year. Or maybe he doesn't.
Ryan Smith - Tuesday, January 19, 2016 - link
"Your table is presenting this quarter's GM as 23% instead of 30%."We realized that right when you did. Fixed.=)
jjj - Tuesday, January 19, 2016 - link
You guys missed a lot.First they posted a new roadmap, kinda.
About Zen they say high core count and multi-threading. On the server side they add about Zen - Disruptive Memory Bandwidth and High Native I/O Capacity.
They also mention in 2017 a server APU - Multi-Teraflops for HPC and Workstation with Transformational Memory Architecture and Scale-Up Graphics Performance - w/e that means, maybe HBM and CF with discrete GPUs...
In the call they said Zen server is for 2017 and they claimed to have secured several design wins with Zen in server. Polaris was always mentioned as shipping mid-year so nothing new there. At some point they said more than 40% IPC gain for Zen.
The new semi-custom win ramps in the second half , likely Nintendo , could be a big gain if Nintendo goes with a 4K capable GPU on finFET and maybe includes some VR glasses with a 500$ console.
As for the numbers , it's the second quarter the computing and graphics segment grows after tanking hard for quite a while and that's good. They also said they expect the PC market to be down low to mid single digits this year and they expect their units to be up.
This year Polaris, Zen in desktop and Nintendo are opportunities, next year Zen in server and APU and maybe a little ARM server.
Not too optimistic about Zen in single threaded, doubt it will beat Intel but they can offer decent perf at decent power at much better price per core(no GPU). If they beat Intel and offer 2-3 times more cores, even better but they can survive with less.
ddriver - Tuesday, January 19, 2016 - link
All they need to match Intel's single threaded performance is to make a chip with comparable throughput. AMD have historically had process disadvantage, forcing them to skimp on the transistors. In the pentium 4 days they had arch advantage which mitigated that, but today intel has good arch.Alexvrb - Tuesday, January 19, 2016 - link
When talking about HPC APUs, "scale-up graphics performance" could also mean crossfire/shared compute opportunities with multiple APUs on the same board.vladx - Wednesday, January 20, 2016 - link
"Nintendo goes with a 4K capable GPU on finFET and maybe includes some VR glasses with a 500$ console"Ha, that will never happen.
jjj - Wednesday, January 20, 2016 - link
It is doable though, the easiest way to go and would give them a big advantage over the others. VR glasses are just a screen and not much else so cost wise it is very doable- small screens with no touch layer and cover glass are cheap no matter what lies some might be telling the public. Also do remember hat in consoles folks don't aim to make money with the hardware and that Nintendo was the first one do push gesture based controller.ofc not saying the glasses would be 4k since you need to run those at high FPS but they don't need 4k glasses now.
frenchy_2001 - Wednesday, January 20, 2016 - link
1) Nintendo makes money with their HW. Sega and Sony were the first to push for "Lose money on HW, make it up on SW".2) Nintendo has never been about overpowered specs and super impressive graphics. I doubt this will change now.
3) VR requires both hi-res and hi refresh. This is part of the problems faced by Oculus/HTC. There is no cheap VR screen today.
Conclusion: I really doubt Nintendo will even aim for your dream.
Kjella - Tuesday, January 19, 2016 - link
While we're nitpicking all the table headings say Q3 2015, not Q4 2015.ddarko - Tuesday, January 19, 2016 - link
"AMD attributes the seasonally lower sales of semi-custom SoCs as the reason for the drop from last quarter, and the year-over-year drop is due to lower game console royalties, and lower server and embedded sales."I'm mystified - how did AMD take in less console royalties even as console sales boomed in 2015?
eanazag - Tuesday, January 19, 2016 - link
Sony and MS would buy those CPUs the quarter before to have systems built Christmas and sell hard. We can likely expect a small boost in the semi-custom business in 1Q 2016. Manufacturing time for those consoles would dictate you want them built in the end of the summer to be ready and sitting on shelves in November.ddarko - Tuesday, January 19, 2016 - link
Seasonal buying patterns may account the drop in revenue in Q4 2015 versus Q3 2015 but it doesn't explain why Q4 2015 revenue ($488 million) was also lower than Q4 2014 ($577 million). The year-ago comparison largely takes the seasonal fluctuations out of the picture. I am still scratching my head how semi-custom and console revenue in the fourth quarter of 2015 ended up 15% lower than the fourth quarter of 2014.jjj - Tuesday, January 19, 2016 - link
That's on console royalties (older gen consoles that have declining sales) ,server and embedded.Server must have been some 100 million a year ago and likely dropped a lot since. Plus maybe a bit of a shift in semi-custom timing for orders. It does seem like too much and the funnier part is that revenue outlook for Q1 is down 14% , with PC dropping less than that.
2015 vs 2014 in semi-custom was up. Units way up, ASP plenty down but revenue was up. The entire segment was down on year on everything else being down.
Intel999 - Thursday, January 21, 2016 - link
The segment that semi custom is in is mixed with other sales channels such as embedded, server chips, and probably their memory sales. Also the amount that AMD was paid per console chipset was lower in 2015 than it was in 2014 as part of the pricing agreement with Sony and Microsoft. It will be slightly lower in 2016 as well, however, it is expected that both Sony and Microsoft will be lowering their retail prices this year to drive volume. AMD anticipates the higher volume to overcome the lower selling price they will be receiving to drive higher overall revenues from console sales.In short, the drop in overall sales cannot be blamed solely on console sales. I'm sure a big chunk of it is due to lower Opteron sales year over year.
jjj - Tuesday, January 19, 2016 - link
Console royalties means Nintendo and other older gen consoles. The PS4 and XB! are categorized as semi-custom.testbug00 - Tuesday, January 19, 2016 - link
price per chip bought by Microsoft and Sony lessens. I think AMD also makes a little bit of money from the Xbox 360 and Wii/Wii U consoles.Sony and Microsoft's purchases for the holiday season also fall under AMD's 3Q2015 results as time required to build and ship to stores.
Duckeenie - Tuesday, January 19, 2016 - link
Nvidia has >80% discreet graphics market share according to the first couple of Google sources I managed to find.HighTech4US - Tuesday, January 19, 2016 - link
https://forum.beyond3d.com/posts/188158281.1%
tuxRoller - Tuesday, January 19, 2016 - link
I look forward to the day when amd goes under and Intel, along with nvidia, are split up under antitrust.frenchy_2001 - Wednesday, January 20, 2016 - link
That may have been true if it had happened 10 years ago, but now, both have plenty of competition from mobile.Intel may be the only game in town in desktop, but ARM (through Qualcomm, Samsung, Apple, Mediatek...) is selling more designs per year than intel.
Similarly, in mobile, Nvidia is competing (and lost, really) with Qualcomm (Adreno, former ATI mobile group), ARM (in Samsung and Mediatek chips), PowerVR (Apple chips) and others (vivando...).
So, there is a case to make for both, but it's unlikely they would be split. Moreover, the US administration has shown great reluctance in applying antitrust lately...
iwod - Tuesday, January 19, 2016 - link
I think Zen is too little too late. The only thing it will do is force Intel to lower its price, Xeon-D, and Xeon E5. I just do not see AMD's way out, and their Radeon Graphics are doing shit.May be they should be acquired by Intel. On a Financial Level, no one is interested to pick up a downtrend going company with lots of debt and little to no upside.
Intel could pick it up for less then $4B. Just Why hasn't Intel made the move yet?
( Monopoly doesn't really count as argument anymore, Desktop, Laptop market is shrinking. )
frenchy_2001 - Wednesday, January 20, 2016 - link
In tel has literally nothing to gain from buying AMD.Few companies would, as their biggest asset, their x86 license, is officially non-transferable.
The rest of their assets (designs for both their CPU and GPU) have little value to anyone else, as those are very competitive businesses and a slew of underpowered competitors exist from mobile (which is the growing market).
webdoctors - Tuesday, January 19, 2016 - link
They lost 84 cents per share in 2015. That's crazy, that's almost half their stock price. The remaining ATI engineers in Toronto can probably pool their money together and buy the company back and rename it ATI.I'm glad anandtech kept posting those ridiculous articles about superstar ABC or XYZ joining AMD, helped get the stock price back above $2 long enough for me to dump it. The company has some cool technology but the crappy management and brain drain of talent has sunk the company.
Its bad for all engineers in the industry when a company fails but the market place has no mercy.
Jumangi - Wednesday, January 20, 2016 - link
AMD keeps losing money but I guess its not that much each quarter so they can just drag on for another year hoping for the next CPU or GPU to save them. I think at some point they would have to sell out and get what they can but hey somehow RIM is still around these days.DnaAngel - Wednesday, January 20, 2016 - link
Everyone saw this coming honestly. AMD has stopped innovating years ago. They rely on market hype and recycle old products. They completely abandoned desktop CPU market in favor of just focusing on APU's for the mobile market, GPU's and Server CPU's. Since then Intel has made leaps and bounds in SoC R&D, allowing their mobile chips to not only outperform AMD APU's in both CPU power and integrated on-die GPU power as well. Nvidia is still updating GPU architecture (Pascal will be introduced this year) thus broadening the gap between the two. AMD's R300 series was supposed to be the next best thing, going toe to toe with Nvidias GTX 900 series and then some, but once the smoke cleared, they were just rebrands with almost 0 performance or innovation gains.AMD is in financial distress due to lack of innovation and no real desire or means to fully compete anymore. They are milking the contracts they have with MS and Sony for the APU’s put into the new consoles, but other than that they are in real trouble. Eventually Fed will step in and funnel them money (if they aren’t already) to just keep them afloat on paper, as not to create an Intel/Nvidia monopoly, which by default is already happening.
I would gladly buy AMD products as I have in the past, when they can go toe to toe with current offerings for a lesser price.
xthetenth - Wednesday, January 20, 2016 - link
I wish Nvidia would stop innovating like AMD, they'd probably come up with something decent. The biggest innovation we've had from them is pulling out the FP64 and compute flexibility to free up power for gaming and selling FP16 capabilities. It's not AMD who's going to be putting its competitors memory onto its next generation chips designed for an API pioneered by its competitor, now is it?Incidentally, if anything it looks like Polaris, which had working samples at CES is going to beat Pascal, which had 980M modules used in its place, and that (surprise surprise) has an updated GPU architecture.
I have a feeling that last sentence is their biggest problem. If they're even they need to have a better price as well.
frenchy_2001 - Wednesday, January 20, 2016 - link
HBM is an industry standard, same as HBM2. AMD was the first one to implement it with their partner, Hynix. Now, what did that give them?The truth is that both AMD and Nvidia were hurt, badly, by the stagnation of foundry process. Nvidia has been on 28nm since the GF680, in march 2012 (similar to AM with the 7970 in january 2012). Seeing no end to it, Nvidia sacrificed 64 bit capabilities for better gaming performance. AMD decided to skip a refresh outside of Fury.
This year will be the first new process introduced in GPU since 2012. It should give us great improvements (from FinFET and smaller dies).
silverblue - Wednesday, January 20, 2016 - link
Your post assumes AMD doesn't have a single product on their roadmap. In addition, as regards your last sentence, AMD do actually have products that undercut equally performing NVIDIA offerings. The 360 isn't worth it over the 950 (which is why I bought the latter), but despite the power consumption difference, AMD are certainly competitive as regards their graphics products more often than not.It's no accident that their server market share has plummeted, more so than their desktop CPU market share, due to Bulldozer. Doesn't it stand to reason that a good product with Zen would drag them back up from the depths in both markets, even if just a little? It's not as if AMD can't develop a good architecture, they just had to make the most of their failed experiment in order to actually have some income whilst they developed a replacement. It takes four or five years to bring out a completely new architecture.
Achaios - Saturday, January 23, 2016 - link
It is interesting to note that the last time AMD was really competitive in the CPU market was, um, 12 years ago (Feb 2004) when the newly released Prescott Pentium 4 CPU's were beaten in terms of gaming performance by the then "Athlon 64" CPU"s.In the space of these ensuing 12 years, AMD have failed to produce a CPU that like the "Athlon 64", would give INtel a run for their money.
It seems unlikely they will ever succeed, and that is bad for the end user.
beck2050 - Monday, January 25, 2016 - link
2015, the operating loss of $253 million. It's amazing year after year they can stay in business, let alone design new products.Sam Snead - Tuesday, January 26, 2016 - link
I've made a ton of money by buying AMD stock when it's down and selling when it's up. I buy their CPUs and GPUs because they are less expensive and do what I want. And remember Nexgen? I sold their motherboards before they were absorbed by AMD. Still have a P100 or whatever. Thank you that it's America and we're not stuck with Intelisma!HighTech4US - Monday, April 11, 2016 - link
Wow buy low sell high. Who would ever have thought up such a system.Funny how only you know when exactly to buy and sell AMD.
You seem like gamblers I know who N E V E R lose.