AMD is shooting for 50 per cent graphics market share says CEO

[HEXUS]

AMD's Dr Su asserts that Vega will be "very very competitive" at the higher end.

Read more.

[EndlessWaves]

"intends to become "very sticky" going forward."

Urgh, vendor specific features? I suppose taken to a modest degree it's a good thing as it encourages AMD and nVidia to diversify into new areas. As long as they stay in the habit of merging to a common technology after a period of time.

The Fury/Fury X ended up very competitive with the 980ti after a couple of driver revisions, the X being faster and the Fury being more power efficient. Given that GP102 is already available I wouldn't be at all surprised if Vega managed to surpass it.

The real test will be at the GTX 1070 level though.

[outwar6010]

Adoredtv did a great video on the issues with amds market share.

[CAT-THE-FIFTH]

Adoredtv did a great video on the issues with amds market share.

I wouldn't put much faith in the position. He said six months Nvidia was doomed due to AMD involvement in consoles,etc(master plan videos),and now he goes the opposite way saying AMD is doomed since he didn't predict Nvidia making so much money(no wonder since Nvidia launched before them,and has their whole range out now) in 1Q,etc.

Then he goes on how Vega is delayed(IIRC) since the concentrating on other areas even though it is quite obvious,that if it uses HBM2,its tied to the production schedule of Hynix who basically said they would START in Q3 2016 in August this year.

Plus even if smaller Vega uses GDDR5/GDDR5X its most likely they are using GF 14NM partly due to WSA and if you look at Polaris 10 and Polaris 11,both chips are on their THIRD stepping already. Not even the RX460 uses a full chip which is why it gets thrashed by the GTX1050TI.

We don't even know how this has impacted Polaris 10 and 11 sales in the OEM space either.

So imagine a larger 300MM2 to 400MM2 chip?? Nvidia used Samsung who licensed the process to GF in 2014 and GF are a year behind Samsung in making 14NM chips,and even Samsung has not produced chips over 200MM2 in size.

Yet,AMD's share price is at a level which has not been seen for a very long time,and even though AMD was considered finished,they have managed to more or less doubled their card sales from one year ago.

Q3 2016 is the biggest quarter for card sales in two years,both companies have benefit and Nvidia more,since they have a full stack out anyway,and the Nvidia cards are all cheaper than their AMD equivalents to make too,hence they have better margins. But compared to one year ago,AMD were in far worse situation on all levels.

Sure Nvidia is making more money,but AMD needs to start somewhere. Even the talk about the GTX200 series taking out the HD4000 series in sales,etc ignores the whole negativity about AMD at the time when they lost over $4 billion between 2007 and 2008,had failures with the HD2000 and HD3000 series and the first Phenom,Nvidia and its Focus Group members like Rollo,were shouting about Nvidia PhysX,etc at every step and they did even sponsor more games too. AMD was considered by many to nearly go bankrupt at that point,so no wonder when Nvidia dropped its prices they got more sales in the end.

Thats why they ended up signing WSA as part of them divesting their fabs. WSA is the main reason AMD has made any losses since.

[khanmein]

adoredtv bias toward AMD but that shills didn't receive groupie luv from red camp & now switched back to NV!

[EvilCycle]

I can't see them hitting 50% this generation as most people I know have already invested in the 10 series Nvidia cards and won't see any point in moving sideways to a new AMD. Gamers on more of a budget have mostly already gone for the RX470/RX480 so GPU sales are going to slow down over the next year or so (unless people still holding on to 9 series cards jump ship to Vega - maybe that is their hope). Still, good luck to them though I mean competition is always a good thing, I just think they aiming a bit too high there.

[CAT-THE-FIFTH]

I can't see them hitting 50% this generation as most people I know have already invested in the 10 series Nvidia cards and won't see any point in moving sideways to a new AMD. Gamers on more of a budget have mostly already gone for the RX470/RX480 so GPU sales are going to slow down over the next year or so (unless people still holding on to 9 series cards jump ship to Vega - maybe that is their hope). Still, good luck to them though I mean competition is always a good thing, I just think they aiming a bit too high there.

But thats the thing Lisa Su said it would take years,so I don't think even AMD expect that - I think they are trying to just grow sales volume,and IIRC they said even ASP have gone up a bit. RTG has only been around for a short period and they are still probably relying on products in development prior to that.

I personally think Zen is more important here,mainly if Zen is a decent product with a decent launch,I expect it has a good chance to uplift the whole AMD product stack very nicely. I think AMD not really updating their performance CPUs for around 4 years has really helped perception of any of their products too. Sadly,they need to fire on all cylinders for CPU and GPU if people want to view them more positively as a company and that is not going to be as easy as having to concentrate on only one set of products IMHO OFC.

[scaryjim]

29/81 is the status quo

They're sharing 110% of the graphics market? No wonder both companies are making good profits this year...

[EvilCycle]

...I personally think Zen is more important here...

Completely agree!

[GrahamC]

I think that there are more than 1 or 2 GTX980ti owners like myself, waiting to see the price/perf ratio of Venga and how it might change the 'Dick Turpin' pricing of Nvidia cards.

[SHSPVR]

I can't see them hitting 50% this generation as most people I know have already invested in the 10 series Nvidia cards and won't see any point in moving sideways to a new AMD. Gamers on more of a budget have mostly already gone for the RX470/RX480 so GPU sales are going to slow down over the next year or so (unless people still holding on to 9 series cards jump ship to Vega - maybe that is their hope). Still, good luck to them though I mean competition is always a good thing, I just think they aiming a bit too high there.

I still have my HD7870

[SineWave]

They can shoot all they want and have abnormally great hardware but if the drivers continue to suck they won't get anywhere. I'm particularly referring to Linux drivers. We need great Vulkan drivers and to be able to play games under Steam and Linux with AMD cards properly.

[kalniel]

They can shoot all they want and have abnormally great hardware but if the drivers continue to suck they won't get anywhere. I'm particularly referring to Linux drivers. We need great Vulkan drivers and to be able to play games under Steam and Linux with AMD cards properly.

That still won't help them get much by way of market share though, since the volume of exclusively linux users is probably very small.

[nobodyspecial]

No chance of hitting 50% (or even 40%) unless they intend to continue to lose money by giving stuff away to keep from losing share. It is not enough to be "competitive". You have to actually WIN something to sell for profit unless the other side simply can't produce enough and that is not the case for NV at this point (maybe if you launched when they did, but not this late as yields of every NV chip hit highs). With NV hitting all markets now basically, they'll lose from that 30%, and I further expect HBM2 to be the same mistake HBM1 was (difficult and expensive going it alone). NV wisely went with the much easier and cheaper to produce memory that was MORE than enough bandwidth (gddr5x) for even the next rev probably. AMD keeps thinking "blue crystals" (see how that worked out for Intel way back when they stopped listening to engineers and went with marketing) will sell stuff instead of WINNING. Nobody cares what is under the hood if it wins everything in sight. If farts somehow was the memory on the top gpu card and won every benchmark I wouldn't touch anything else even if one company claimed HBM49 was 1000 times faster than farts. Unfortunately for AMD it is EXTREMELY difficult to out engineer the other guy (hence competitive, instead of "we'll waste them next Q with Vega") when you laid off most of your engineering talent to stop bleeding cash.

It is true the market is on FIRE, but it's on fire for the best in town, no matter who owns it, or hates it. For every whiner that is on a hardware site about prices, etc, there are 7-10 other people BUYING top end NV cards allowing them to set record profits, sales, revenue and margins. I still can't believe AMD shot for low margin crap instead of top end. That is the difference between profit and loss. I hope ZEN isn't aiming at $300 max, but rather that what we've heard is the LOW end and they'll actually be competing in the $1730 range ALSO for desktop, and competing months later in 16-32 core at the $7000 Intel end on servers. To do anything less than aim for the HIGHEST margin stuff is stupid and should get you FIRED. If NV had 10% market share but sold all of it in the high end while AMD had 90% share and couldn't make a dime in the low end NV and share holders would still laugh. Market share isn't worth squat if the other guy is making all the cash to re-invest etc. If you can't make a dime, market share is pointless and you've failed. AMD should have let go of most of management and ALL of marketing LONG before releasing a SINGLE engineer. You don't have to "MARKET" and winning product with 20+ top hardware sites reviewing every gpu/cpu that is worth talking about. The benchmarks and reviews do it for you. ENGINEERS win races, not marketing. When AMD had the top cpu for 3yrs, they also had the worst marketing dept on earth but still made their highest margins (58.5%) and great profits. Note they sold their chips at EQUAL price to Intel then also. They only thing that kept Intel from taking a huge crap was AMD's inability to produce more than 20% of the total market. I really hope AMD aimed for the sky this time and produced a HUGE ZEN chip (dwarfing Intel's, say 350mm^2+ or roughly xbox1 rather than Intel 246mm^2 10 cores) as that would mean $1700 cpus on desktops and $7000 cpus on server side. It would also mean MARGIN city. Intel makes it's top 4 desktop chips for under $100 and sells them for $434-1730. AMD should follow this model rather than everything they sell being under $150 and losing money hand over fist trying to be small (slow) and cheap.

Marketing leads to "competitive", while engineering leads to "we'll wipe the floor with them by 30-40% and we see Q1 being an major explosion of profits for us". LOL. See NV benchmarks and earnings reports. Sure you can engineer a great small efficient chip (but that takes actual numbers of engineers, see Intel/NV), but if you're lacking personnel, you had better take the large brute force attack first and do it on the highest margin products and sell the junk stuff later. Xbox1/Ps4 dwarf Intel's HEDT chips size, and are made for $90-100 and sold for $100-115 (ps4/xbox1 respectively). For anyone about to claim huge costs more and you can't make any money, I beg to differ. Numbers don't lie here. I suspect Intel's much smaller die and great fabs are creating HEDT chips for less than $70 (I'd guess closer to $60 probably as 1/3 less material and better fabs should be able to hit near that vs. xbox1/ps4). There is a reason why Apple usually kicks everyone to the curb each gen (large dies compared to everyone else in socs) when released. If ZEN is as big as xbox1's die, they will likely kick Intel to the curb and be selling a ~$100 die cost for $430-1700. They'd get rich doing it. I'll buy ZEN if it's over 300mm^2 (which should win many things, and rarely lose as Intel is better at engineering here today) and will drool over my chip if it's 350mm^2+ as that would win everything probably. I will not buy a zen loser. $300-330 vs. a $430 Intel means nothing if Intel wins everything. $100 more on a chip I'll use for the next ~3-5 years is a joke to me. I'll gladly pay $2-3 a month to win everything over it's lifespan.

[nobodyspecial]

https://finance.yahoo.com/news/amd-r...201500443.html
One more point, NV isn't just making more money. AMD had a net loss of $400mil Q3. I expect about the same from Q4. The only reason their shares are up to ~8.75 is because of the HOPE that Zen/Vega can compete. Since neither will be on the report this month, they won't be making money on those two products (even if a winner in either case) until next Q. There is no reason they'll make money this Q. Maybe a slight bump (300mil net loss?) due to more console chip sales going into the refreshes of xbox1/ps4 as they sell well ahead of time before landing on shelves (so much of them are on this Q for all 3 months). But even 10mil x $15 a chip overall is only 150mil more to the bottom line (maybe 250mil net loss?). Note they sold 300mil more crap in Q3 vs. Q2, but doing it got them completely killed. Again, if you're not making profit on gaining market, what is the point in selling stuff? STUPID management. They would have been far better off having a duplicate of Q2 at 300mil less revenue...ROFL. I'll gladly sell a Billion in revenue to make ~70mil vs. selling 1.3B and losing 400mil...Do these people know how to do math?

The WSA is the 2nd worst decision AMD has made in 2 decades. The first was buying ATI for 3x their value when they never made more than 60mil. What retard pays 5.4B for ATI when it takes ~90yrs of their income to get it back even if they peaked for all 90 years. Either management was completely retarded, or ATI had some brilliant marketers. LOL.

A close 3rd would be firing Dirk Meyer, who made all the great money making chips previously, and then taking 5yrs to realize he was RIGHT when he left in 2011 telling them they needed a KING (not chasing mobile/apu/console crap etc). They fired him for ending up CORRECT 5yrs later...LOL. Idiots. 4th worst decision, was probably consoles. They cost them ZEN being seriously late (R&D diverted from CORE cpu/gpu went to consoles) and all recent gpus being behind NV. It cost them the entire cpu race and basically the gpu race also. It also cost them >30% of their engineers in layoffs, etc. Terrible management. Let's hope Lisa has righted the ship to KING MAKERS. But I fear we'll see a <300mm^2 Zen (probably far less) and a Vega facing 1080ti on steroids cranked to max if Vega is reasonably close. Which again in both cases will mean lower end or crap that can't sell at profit.

Meanwhile Intel/NV left them in the dust on watts/perf. NV wisely passed on consoles claiming it would have diverted R&D from core products (Jen said this, again correct, just like Dirk said at AMD). People claimed they were "butt-hurt" over it. No, they're laughing over the success of KING desktop gpus and KING pro gpus. Dirk was the management equivalent of Jen at NV (or Andy Grove at Intel), just working with less resources. Sadly AMD chases the bottom end (low margin crap) vs. NV/Intel who always chase the high-end/high margin. Dirk wanted to ignore low margin/low end junk and go high too. Bummer. NV isn't making money on 1050's, they're making it on 1070/1080's, quadro's and teslas. You need to make high margins before you bother with low margin junk that merely (maybe) adds a few bucks to the bottom line. The high margin stuff adds hundreds of millions (or in Intel's case BILLIONS) to the bottom line. Intel was losing 4B a year chasing the bottom in socs (mobile, which even NV had to give up...LOL, went higher margin AUTO stuff with it) and couldn't over come the problem even with far more engineers. Why does AMD keep thinking they can do what NV and even the mighty Intel failed at?

I do believe NV will go back into mobile once gaming is more relevant there and everyone is using mobile as an extra gaming device hooked to a tv over hdmi etc. As gpu becomes the dominant factor in your purchase they will go back in. They were 5yrs ahead of their time unfortunately and no games to use their "wheel-house". But Vulkan and all the engines using it are putting android etc on steroids for gaming so they'll go back in and just use samsung/qcom/Intel modems as die shrinks allow this with no real power penalty. I suspect 10nm will see them re-enter mobile with someone's modem. At 10nm with GDDR5x or HBM2 with Vulkan (and volta gpus etc) etc you could see your mobile device being as good as an xbox1/ps4. That alone would mean 300mil units sold per year that people would love to play games on or hook to a TV and we're not even talking ~1.2-1.5B phones that get more capable then too. That will dwarf console sales of next gen so far.

I suspect we'll also see an android/linux dual boot PC with a 100w chip soon too. The ARM players are just waiting on the OS and Vulkan to take root so woo gamers (then REAL apps come later). You may even see SteamOS/Android/Linux boxes (all free). Wintel is about to get pinched soon. You won't go Dx12 vs Vulkan unless paid to do so soon. The next gen of kids all grow up with a mobile OS not Windows. They won't care that windows (or even wintel totally) is not on their box as long as that box plays top quality games which Vulkan (and all the engines running it) will allow. I really hope ShieldTV rev2.0 has 2 socs, and amped up to 4ghz with a fan/heatsink on it with GDDR5x or HBM2 and make the console race a REAL race. I'll gladly take an rev2 and android's gaming prices even $5-$30 vs. console at $60 games and you can get dozens of great games for $2-10 already. I have about 80 on my list of wanted games for android. Just waiting on ShieldTV rev2 or an NV based 13in+ tablet. ShieldTV was already impressive, I just want much closer to console perf and 10nm. I'd buy or build an ARM based gaming PC box in a heartbeat also, hopefully complete with an option for NV/AMD discrete gpus, 500w psu, SSD's/HDD's etc. That box is coming folks. Vulkan will bring REAL gaming to a few Billion units yearly, instead of a few hundred million units now. I can't wait for that to take shape

[Strawb77]

"AMD is shooting for 50 per cent graphics market share says CEO"

they should sack him- anybody worth their salt would be aiming at 100%
anyone who expects less than 100% will always get less than 100%
i bet he still gets 100% wages though