AMD really needs to get their footing back to the server/HPC market. They need to follow nvidia in the mobile/ARM space. They need to remake their mobile/smartphone graphics division. And hell, retake their fabs!
Now that's a lot of work and needs a lot of capital. But I seriously can't see AMD growing with their current markets.
Whilst at it, why not ask them to declare bankruptcy as well? Talking is cheap, mate. They haven't made overall positive income in couple of years and you're asking them to bleed more money.
Very true, I don't think he, or many people, realize this margin of growth is stronger than AMD has been in over a decade. They are clearly doing something right...those console design wins certainly help in sales and "free" marketing, and they still have superior integrated graphics, especially in cheap $400 A10-based notebooks.
I doubt this has anything to do with the consoles, which AMD has controlled for 3.5 years now, and everything to do with polaris coming out. AMD finally had a GPU without the 300 stigma attached to it, and while not nvidia numbers, polaris has been selling very well for AMD.
1) New editions of BOTH consoles were released last year at premium prices with higher margin AMD chips. It's likely for every XBOX One S or PS4 Pro sold, AMD is making the same profit they would on 2 or 3 regular model consoles, since the only real difference between these consoles is the APU. 2) Most mainstream console gamers now associate AMD with gaming, and are statistically more likely to use AMD components when building a PC or talking about gaming, and this marketing was essentially "free" because it's advertised by the console manufactures that the consoles have an AMD "Jaguar 8-core graphics engine" or whatever the hell PR wants to spin it as. 3) Because of AMD dominance in console gaming, most games are being optimized for their GPU architecture without incentives directly from AMD, making the gaming performance more consistent on their GPU's. This explains why the RX480 bests the GTX970 across the board in games, when in reality, it shouldn't, as indicated by various synthetic benchmarks. All the more interesting because in my opinion, Nvidia has far superior driver optimization.
They don't have the resources to branch out into everything. They're on the right track, but they've got their hands full getting Ryzen together for DT CPUs, server CPUs, and APUs. On the graphics side Vega and Navi are keeping them busy. Then there's the ongoing semi-custom business, like Scorpio's APU.
They wisely put their ARM plans on the backburner for the moment, but they will likely continue that work later. Mobile graphics, at sub-tablet TDPs? Not very likely. Too much competition anyway, even Nvidia isn't having much luck forcing their way into the smartphone market. You've got PowerVR, Mali, and Adreno to contend with. Not sure if it's worth it, really. Has Nvidia made any serious money selling Tegras recently?
Adreno (A world play on "Radeon") used to be owned by AMD anyway.
Chinese SoC's are quickly ramping up in capability whilst racing to the bottom in terms of price, not an ideal thing AMD wants to enter into right now.
ATI owned it, but back then it was Imageon. That was so long ago they'd have to start over by scaling down GCN. Anyway, speaking of the cheap Chinese SoC problem - if QC is forced to license their stuff out for super cheap by the powers that be, that may even exacerbate the issue as the cutthroat firms will benefit the most.
ARM servers are as dead as they ever were, so I don't expect AMD to have another go. Consortia have tried to put software stacks onto ARM but truth is, everything enterprise is on x86, and it's staying there. Why re-invent the wheel? On any given process node, in terms of performance/Watt, general purpose architectures have more or less converged.
I wasn't really thinking about servers. Ryzen can pack a lot of cores/threads into a single chip, I think most of the server needs will be covered there. I was thinking more about network devices, IoT, etc.
All they need right now is to allow us to buy Ryzen - as in not pricing it out of range.
If they had 3 quad cores SKUs between 99$ and 199$, a 6 cores against the 7600k at 249$, another hexa at 299$, then some octas at 319$, 349$ and a top SKU higher, they would make this platform matter and be back in the game. They could add some more SKUs later on to further exploit certain price points but they better not launch this with 3-5 SKUs at idiotic prices. If they get the pricing right, in 2018 they get to 12-16 billion revenue ,7nm and 12 cores for 2019 and ample resources to keep milking existing markets and go after new ones.
If we can't afford to buy Ryzen, they might as well just give up. Intel made it way easy for AMD and they don't deserve to exist if they don't see the opportunity.
Did I say that I knew that or he knew it? Reread what I and Matt stated. He stated that success and pricing of Zen was predicated on it being successful. I rephrased what he said. I said nothing about Zen actually being successful.
Totally wrong. You need to get a better job so you can afford it. AMD needs to charge what Intel is charging if they are performing like (or better) Intel. No discounts. Charge what you perform as. PERIOD. They just lost 500mil again this year. They can't afford to make chips for poor people, the money is made in the high end, just ask Intel/Nvidia.
They shouldn't sell a ryzen chip that is less than $200 (much like Intel), and should run all the way up to $1730 if they have a chip that performs like one. Save crappy chips for apu. Sure I like cheap cpus too, but I'd rather have AMD charge what they are worth and make MONEY. IF you sell out immediately you're not charging enough. PERIOD. Business 101.
It's good to see AMD finally increasing their margins. They were terrible for the last several years and it makes dedicating money to R & D much more difficult. A strong AMD is good for everyone including Nvidia and Intel.
Mostly looks slightly better because they took all of their one-off-but-happens-every-year GF fine in the 3rd quarter. If you spread that over the year then every quarters margins are rubbish.
So the Gross Margin of 5% in Q3 2016 as a result of having to pay GF all that money didn't happen then?
You can't say margins are growing when you have a 5% margin in one quarter, especially as that expense wasn't for GF purchases for that quarter only so really should have been spread over all of them.
So they are cutting their losses in half and are seeing healthy revenue growth. Give Lisa a year more or two and AMD will finally be back on firmer footing.
me too. but it's not too late. I've been buying some in anticipation of AMD's stock ryzing (heh) soon. Between vega and ryzen, its going to be a good year for AMD.
Unfortunately I think Ryzen being a success has already been baked into the share price. When Ryzen releases, assuming it has something close to parity with Intel, then the stock price will drop by probably 25% as profit-taking occurs. If Vega is awesome, and Ryzen sales are brisk then you will see a recovery in the later part of the year. Buying now, in my opinion, would be foolish.
AMD had another bad year. The third worst in losses, the second worst in revenue. This follows the pattern that AMD has had ever since the disastrous 2012. AMD has had a problem with negative cash flow due to its steady losses and decrease its debt load. I was pleasantly surprised to see that AMD manged to have more cash on hand at the end of 2016 than in 2015: $1.264 billion vs $785 million. I had to dig to find out why. In 2016, AMD sold off, err had a "joint venture" with ATVP JV that netted them $351 million in Q2. During 2016, AMD sold $1.47 billion in new stocks and convertible senior notes. That is close to two billion dollars in INCOME to reduce their debt from $2 billion in 2015 to 1.4 billion in 2016 and keep the cash flow positive. This is at the cost of diluting the stock by almost 30% if those convertible notes are, err, converted. Did that affect the stock price? hardly.
BTW margins are good? ROFL 32%? Good is NV/Intel margins ~60%. Can't believe AMD's stock rise on quarter after quarter of losses. Ryzen had better be good or the stock will crash.
In speaking with actual people who have actually touched the product it is good, it all depends on how well they execute and create mindshare.
I woul think that the server market is probably the area they really need to make an impact as it is the area they most liekly stand to make the most profit.
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.
34 Comments
Back to Article
OEMG - Tuesday, January 31, 2017 - link
AMD really needs to get their footing back to the server/HPC market. They need to follow nvidia in the mobile/ARM space. They need to remake their mobile/smartphone graphics division. And hell, retake their fabs!Now that's a lot of work and needs a lot of capital. But I seriously can't see AMD growing with their current markets.
WorldWithoutMadness - Tuesday, January 31, 2017 - link
Whilst at it, why not ask them to declare bankruptcy as well? Talking is cheap, mate. They haven't made overall positive income in couple of years and you're asking them to bleed more money.Samus - Wednesday, February 1, 2017 - link
Very true, I don't think he, or many people, realize this margin of growth is stronger than AMD has been in over a decade. They are clearly doing something right...those console design wins certainly help in sales and "free" marketing, and they still have superior integrated graphics, especially in cheap $400 A10-based notebooks.TheinsanegamerN - Wednesday, February 1, 2017 - link
I doubt this has anything to do with the consoles, which AMD has controlled for 3.5 years now, and everything to do with polaris coming out. AMD finally had a GPU without the 300 stigma attached to it, and while not nvidia numbers, polaris has been selling very well for AMD.Samus - Wednesday, February 1, 2017 - link
It has everything to do with the consoles.1) New editions of BOTH consoles were released last year at premium prices with higher margin AMD chips. It's likely for every XBOX One S or PS4 Pro sold, AMD is making the same profit they would on 2 or 3 regular model consoles, since the only real difference between these consoles is the APU.
2) Most mainstream console gamers now associate AMD with gaming, and are statistically more likely to use AMD components when building a PC or talking about gaming, and this marketing was essentially "free" because it's advertised by the console manufactures that the consoles have an AMD "Jaguar 8-core graphics engine" or whatever the hell PR wants to spin it as.
3) Because of AMD dominance in console gaming, most games are being optimized for their GPU architecture without incentives directly from AMD, making the gaming performance more consistent on their GPU's. This explains why the RX480 bests the GTX970 across the board in games, when in reality, it shouldn't, as indicated by various synthetic benchmarks. All the more interesting because in my opinion, Nvidia has far superior driver optimization.
Alexvrb - Wednesday, February 1, 2017 - link
They don't have the resources to branch out into everything. They're on the right track, but they've got their hands full getting Ryzen together for DT CPUs, server CPUs, and APUs. On the graphics side Vega and Navi are keeping them busy. Then there's the ongoing semi-custom business, like Scorpio's APU.They wisely put their ARM plans on the backburner for the moment, but they will likely continue that work later. Mobile graphics, at sub-tablet TDPs? Not very likely. Too much competition anyway, even Nvidia isn't having much luck forcing their way into the smartphone market. You've got PowerVR, Mali, and Adreno to contend with. Not sure if it's worth it, really. Has Nvidia made any serious money selling Tegras recently?
StevoLincolnite - Wednesday, February 1, 2017 - link
Adreno (A world play on "Radeon") used to be owned by AMD anyway.Chinese SoC's are quickly ramping up in capability whilst racing to the bottom in terms of price, not an ideal thing AMD wants to enter into right now.
Alexvrb - Friday, February 3, 2017 - link
ATI owned it, but back then it was Imageon. That was so long ago they'd have to start over by scaling down GCN. Anyway, speaking of the cheap Chinese SoC problem - if QC is forced to license their stuff out for super cheap by the powers that be, that may even exacerbate the issue as the cutthroat firms will benefit the most.Meteor2 - Wednesday, February 1, 2017 - link
ARM servers are as dead as they ever were, so I don't expect AMD to have another go. Consortia have tried to put software stacks onto ARM but truth is, everything enterprise is on x86, and it's staying there. Why re-invent the wheel? On any given process node, in terms of performance/Watt, general purpose architectures have more or less converged.Alexvrb - Friday, February 3, 2017 - link
I wasn't really thinking about servers. Ryzen can pack a lot of cores/threads into a single chip, I think most of the server needs will be covered there. I was thinking more about network devices, IoT, etc.jjj - Wednesday, February 1, 2017 - link
All they need right now is to allow us to buy Ryzen - as in not pricing it out of range.If they had 3 quad cores SKUs between 99$ and 199$, a 6 cores against the 7600k at 249$, another hexa at 299$, then some octas at 319$, 349$ and a top SKU higher, they would make this platform matter and be back in the game. They could add some more SKUs later on to further exploit certain price points but they better not launch this with 3-5 SKUs at idiotic prices.
If they get the pricing right, in 2018 they get to 12-16 billion revenue ,7nm and 12 cores for 2019 and ample resources to keep milking existing markets and go after new ones.
If we can't afford to buy Ryzen, they might as well just give up. Intel made it way easy for AMD and they don't deserve to exist if they don't see the opportunity.
Michael Bay - Wednesday, February 1, 2017 - link
It all is predicated on Zen being actually good, which is not happening.baka_toroi - Wednesday, February 1, 2017 - link
What do you mean with "it's not happening"?negusp - Wednesday, February 1, 2017 - link
Success only occurs given that Zen is actually good- pricing, sales, etc.vortmax2 - Wednesday, February 1, 2017 - link
Yes, but there's no way he knows that yet. No one does, yet.negusp - Thursday, February 2, 2017 - link
Did I say that I knew that or he knew it? Reread what I and Matt stated. He stated that success and pricing of Zen was predicated on it being successful. I rephrased what he said. I said nothing about Zen actually being successful.silverblue - Wednesday, February 1, 2017 - link
Come on, you must have a reason for posting that, other than troll bait of course. :)TheJian - Wednesday, February 1, 2017 - link
Totally wrong. You need to get a better job so you can afford it. AMD needs to charge what Intel is charging if they are performing like (or better) Intel. No discounts. Charge what you perform as. PERIOD. They just lost 500mil again this year. They can't afford to make chips for poor people, the money is made in the high end, just ask Intel/Nvidia.They shouldn't sell a ryzen chip that is less than $200 (much like Intel), and should run all the way up to $1730 if they have a chip that performs like one. Save crappy chips for apu. Sure I like cheap cpus too, but I'd rather have AMD charge what they are worth and make MONEY. IF you sell out immediately you're not charging enough. PERIOD. Business 101.
novastar78 - Wednesday, February 1, 2017 - link
You would think that it would be as simple as starting a mobile division, but alas it is not that simple.They sold the business to Qualcomm years ago and, depending on the state of the agreement, they signed a non-compete.
Can you guess what Adreno is an anagram for?
fanofanand - Wednesday, February 1, 2017 - link
It's good to see AMD finally increasing their margins. They were terrible for the last several years and it makes dedicating money to R & D much more difficult. A strong AMD is good for everyone including Nvidia and Intel.Dribble - Wednesday, February 1, 2017 - link
Mostly looks slightly better because they took all of their one-off-but-happens-every-year GF fine in the 3rd quarter. If you spread that over the year then every quarters margins are rubbish.fanofanand - Wednesday, February 1, 2017 - link
Your comment is rubbish. Margins have NOTHING to do with their agreement with GloFo.Dribble - Wednesday, February 1, 2017 - link
So the Gross Margin of 5% in Q3 2016 as a result of having to pay GF all that money didn't happen then?You can't say margins are growing when you have a 5% margin in one quarter, especially as that expense wasn't for GF purchases for that quarter only so really should have been spread over all of them.
beck2050 - Wednesday, February 1, 2017 - link
Still losses for the quarter and the year.TheinsanegamerN - Wednesday, February 1, 2017 - link
but losses are far lower this year then last year. It's still a huge improvement.Mondozai - Wednesday, February 1, 2017 - link
So they are cutting their losses in half and are seeing healthy revenue growth. Give Lisa a year more or two and AMD will finally be back on firmer footing.Meteor2 - Wednesday, February 1, 2017 - link
Just look at that share price -- it's, what tripled in 12 months? I rather wish I'd bought some...TheinsanegamerN - Wednesday, February 1, 2017 - link
me too. but it's not too late. I've been buying some in anticipation of AMD's stock ryzing (heh) soon. Between vega and ryzen, its going to be a good year for AMD.fanofanand - Wednesday, February 1, 2017 - link
Unfortunately I think Ryzen being a success has already been baked into the share price. When Ryzen releases, assuming it has something close to parity with Intel, then the stock price will drop by probably 25% as profit-taking occurs. If Vega is awesome, and Ryzen sales are brisk then you will see a recovery in the later part of the year. Buying now, in my opinion, would be foolish.Meteor2 - Wednesday, February 1, 2017 - link
Yeah indeed. Buy low sell high. Too high a possibility you're quite right and the only way is down if Zen falls short.johnpombrio - Wednesday, February 1, 2017 - link
AMD had another bad year. The third worst in losses, the second worst in revenue. This follows the pattern that AMD has had ever since the disastrous 2012.AMD has had a problem with negative cash flow due to its steady losses and decrease its debt load. I was pleasantly surprised to see that AMD manged to have more cash on hand at the end of 2016 than in 2015: $1.264 billion vs $785 million. I had to dig to find out why. In 2016, AMD sold off, err had a "joint venture" with ATVP JV that netted them $351 million in Q2. During 2016, AMD sold $1.47 billion in new stocks and convertible senior notes. That is close to two billion dollars in INCOME to reduce their debt from $2 billion in 2015 to 1.4 billion in 2016 and keep the cash flow positive. This is at the cost of diluting the stock by almost 30% if those convertible notes are, err, converted. Did that affect the stock price? hardly.
TheJian - Wednesday, February 1, 2017 - link
BTW margins are good? ROFL 32%? Good is NV/Intel margins ~60%. Can't believe AMD's stock rise on quarter after quarter of losses. Ryzen had better be good or the stock will crash.novastar78 - Wednesday, February 1, 2017 - link
In speaking with actual people who have actually touched the product it is good, it all depends on how well they execute and create mindshare.I woul think that the server market is probably the area they really need to make an impact as it is the area they most liekly stand to make the most profit.
Achaios - Thursday, February 2, 2017 - link
Where is that guy who was pushing AMD stock in here and urging everyone and their little sister to buy AMD stock "BECUZ HUGE PROFITS TO CUM MON".On the subject of AMD, they are hopeless. Year after year we get the same results: Loss, loss, loss, loss, minus, minus, minus.
It seems that the folks at AMD are trying to tell us something.