$35M? It's weird, AMD is a global company making highly complicated cutting edge tech that's used everywhere, kicking Intel's butt in desktop segment and they only made $35M the last quarter? Sure its better than a loss but still.
It's crazy a movie or music star made more money than this global company of 10K employees. Ppl can't accuse them of rolling in dough haha.
No one has accused them of rolling in the dough. Not sure why you think they are beating intel butt in desktop market..you do know just because they have a new chip out doesn't automatically mean instant win? Lots of people have contracts with intel, prefer intel chips. Intel still rules in the SFF market and high end gaming.
Erm, no. Ryzen 3000 essentially made Intel's entire lineup, except for MAYBE 9900K, completely pointless at their current prices. You can put a 3900X in mini ITX X470, and later an 3950X. SFF is covered lol. 3900X gives 9900K a run for its money in "high end gaming" So much so you would be hard pressed to notice the difference, does it vastly more efficiently, and oblierates in other tasks: while costing the same and including a cooler capable of sustaining impressive boost clocks. You also get a platform that isn't dead, and if you go with X570, new features that Intel doesn't even support.
People who "prefer Intel chips" as you put it, are tools with Ryzen 3000 on the market. If you can't see the butt getting kicked, you are blind, sorry. Even my 2700X makes half the Intel lineup pointless.
Probably will have to ask for a new server on one of the projects I manage. Purchasing Dept will surely get me Intel. I'm happy that you know that AMD has the better price/performance out there, but I'm guessing that you are on your teens if you think that that will instantly make everyone buy AMD. You also seem to forget that AMD still sucks in laptop CPUs and that most people nowadays buy laptops. Wish things were as easy as you see them.
>You also seem to forget that AMD still sucks in laptop CPUs and that most people nowadays buy laptops.
You either haven't seen the reviews of Ryzen 2xxx U/H & 3xxx U/H series or basically made this up just to make your point! AMD lags in the single core dept for sure however they more than make up for it in the graphics department, the only (major) problem for them is the core count & probably not producing enough chips. If however they do have the capacity for more production, then I can only assume Intel is back to their old ways of <you know what> like we saw recently with that user "benchmark" tweak 😑
Not even Charlie of Semiaccurate is unfair like you. After an analysis of 3000 serie (leaving away expensive low volume 3900 and 3950 cpus), he said that at the end it is only a matter of selling price, and on per core basis the two cpus are comparable. Bel intel will lower the prices at an AMD level soon.
Yes, Ryzen is great and if I were building today I would definitely buy a Ryzen 3000 series, but he is referring to sales, not performance. In sales, Intel is still winning. It takes more than a year or two of good performance to gain marketshare, especially with the high volume enterprise market... It takes consistency. Dont forget the really great Ryzen (3xxx seies) just came out this summer. I am sure in a few more years AMD will gain more share.
No they are not. the DIY PC and the HEDT platform is now AMD.
Intel sales are OEM, prebuilts, laptops, chromebooks and server.
Next year will be detrimental for AMD. They needs to disrupt Intel markets if they want to consolidate their emphasis on the semiconductors level. I do believe AMD will double their actual size as a company, but they need to disrupt Intel other markets if they want to grow even more. They need more than just taking market share, they need to take and create markets.
They need to make money, and to do that they have to offer better product which can sustain high margins. Up to now CPUs margins are good but GPUs are a hole that swallows all the profits. They have to stop that if they want to really grow and not being left at surviving in a corner.
Their new product line was not even incorporated in these results. They did really well with over Ryzen 2000, Vega and Polaris for having those numbers. Google is rumored to ditch Intel. If that happens, a lot of other companies are going to follow.
Agreed, but DIY and HEDT is a tiny peice. That is what I was saying. AMD does have plans in the other markets and OEM wins as well... Next year we will see 7nm laptop chips that will help. With that said AMD will not likely ever dominate. They can however be very healthy and competitive if they keep executing.
DIY PC's is a dyeing market - with very limited appeal. I better attitude for AMD would be support it fan based instead trying to disrupt the completion. Such a negative attitude makes customers like myself don't desire or trust them. Honest Truth.
maybe where you are.. but i have my doubts its dying as much as you claim., if it was, then there wouldnt be so many boards, or cpu's released as there is.
heh.. more like your blind fanaticism towards intel is why you dont desire or trust them, HStewart.
it made intels old lineup irrelevant. but intels old lineup is...well old. if they release cometlake with an 8 core 16 thread i7 instead of i9 and an 8 core 8 thread i5 (the old i7) they basically become very relevant again. things change year to year and sometimes month to month. when amd gets a win everybody declares the death of intel. and then a couple months later when intel comes out with something better nobody has anything say then. just like how an 8700k clocked to 5ghz even beat amds 8 cores from the first gen, even in multithreaded apps. itll go back and forth between them. intel really need to get on to 10nm for desktop but they can compete on 14nm for a bit longer
Folks on PC enthusiasts just see things from their perspective.
Form the perspective of the CTO of Ford/GM/Pfizer/Google/Bank of America...they will be buying those Dell or HP desktops with Intel inside in the millions.
Your valiant purchase of a 3000 series was cute but it wont change a thing.
Ford and them don't give 2 flying fucks what CPU is in the machine. They care if Dell can supply them with a machine, it just happens to be powered by Intel.
The problem isn't that AMD's CPUs aren't what people are after, it's that big builders like Dell aren't selling them.
Problem is that this kind of advantages it is enjoying now, that is the issues Intel has had with its 10nm, is probably not going to repeat in the future. Next year with this product and Intel in its full power is not going to provide more money than today.
The thing is Intel was always having the process node advantage, now it is gone for good. It will never happen again. TSMC is already rolling out EUV 7nm, 5nm and 3nm is on track. Intel will never get this back.
I'm not denying that Intel more advance PP has given it a big help in the past. But actual PP are really expensive, both for production and for development, and AMD, even though using TSMC advanced PP, is not guaranteed to always have this advantage ove Intel. You are one of those forgetting that actual Intel offering is a restyling of something 4 years old, while 10nm is still not fully deployed and 7nm is coming. With those PP there will be changes bigger than simple shrinks. Intel has a new architecture to roll out (and it would already be here against Zen if 10nm were ok) and new interconnections allows for things that AMD cannot do even with a newer PP. So enjoy AMD small advantage when comparing its latest products with Intel 4 years old ones, but be prepared to a fast change of the market when Intel will strike back at full power.
Most of Intel's advantages on their "old" architecture come from tweaking that 14nm so much that it will actually be difficult for their 7/10nm to compete against it for at least a generation on higher power parts.
AMD will probably have the performance advantage till 2021 or so, there I expect it to even out again if AMD does not improve.
We have yet to see all of AMD's architecture improvements with this generation, the 3950X will be quite interesting, and AMD will not release their best chips now as they don't need to. Most of the 3000 series gains seem to be from the improved process node, rather than the architecture, so I'm expecting good gains again from 4000 series.
Verneer is rumored to bring 4 thread SMT, but I'd only expect them to implement this in the server environment if they do, doesn't really make sense in the consumer space. I bet on them splitting server and desktop with 4000 series.
Assuming Intel sleeping for few more years. Obviously this will not happen :). At some point Intel will do a check, can i manufacture?? yes? ok, no? ok i'll will begin to manufacture at TSMC and Samsung, leaving AMD in the dust. It takes only less than one year to convert strategical production at foundries thanks to actual advanced software tools developed by TSMC and others.
You need to work on your uarch to be manufactured at other foundries. It is not as simple as this. Also, Intel is investing massively in their fabs and these results will not bring process node forward, only production. If the product is not good, it will not matter anyway.
I doubt Intel will stay behind AMD for long, and even though AMD is ahead of them now in terms of architecture design and manufacturing node, it won't stay that way for long. AMD has not made Intel obsolete even with it's newest chips, and when you factor in drivers/stability and years of conditioning that Intel is just better it's hard for me to call this a win. Intel products are not bad, now there's actual competition. It's good to see Intel on the receiving end of a price war, but AMD isn't banking yet and it's going to take a while for the money to really start flowing in, probably a lot longer for Intel to put out their next generation of CPUs and turn this around again.
I've loved and used AMD chips since the 90s, but I'm shocked that the company has remained in business somehow. Now they're being squeezed from the other direction by Arm chips and the future is just as uncertain as it always has been.
At this pace, it will take 3 years for AMD to pay off their debt. I really hope they do so. If they crash again, I believe it will be the end for goods, however I don't think it will happens. They have promising uarchs and they are really getting momentum in a lot of markets. I am still really curious to see what Samsung will come out with RDNA.
It is lack of laptop sales. PC/desktop sales are in decline, but laptops are still going strong, of which intel has a massive majority in market share.
Unfortunately for AMD it has very weak laptop offer. They just have those APUs which are good for medium sized laptop and nothing for ultra slim ones. More over, they do not have a single GPU that is power efficient enough to replace Nvidia discrete offer in that market.
Well, if they price it well, then AMD GPU can compete. Remember, the desktop GPUs are way overlcocked beyond the efficiency curve. They can sell for cost price and make a profit on the CPU, would help branding a lot if they can sell all AMD machines.
This quarter at $35M on a total revenue of more than $1500M just shows that AMD is alreading selling at cost. Do you want them to go in red again? If they ever do at this point with these products I doubt it can ever become profitable for the rest of its (would be short) life.
The lack of 6 and 8 core options for laptops is glaring.
I'm not in for an AMD laptop APU until they have 8 cores and AV1 decoding on the GPU. Navi's media decoder on the desktop GPUs does not support AV1, so I could be skipping until at least 2021.
It's not only a problem of n# of cores. AMD has not SKU good for ultra thin laptop or 2in1. And the integrated APU makes it a bad choice for high end ones that need a (big) discrete GPU. Intel and Nvidia have the entire market of such devices, and they will until AMD will make APU with smaller iGPU and more core for high end laptop or office oriented ones or smaller packaged version for ultra thin client.
I'll concede that's it's not just about # of cores. But that's a big deficiency given that Intel has 8 cores on mobile and is rumored to be going for 10 cores. If I want a top of the line AMD mobile Picasso APU, the Ryzen 7 3750H, that's just 4 cores. The Renoir successor may not even have Navi graphics, and even if it does, Navi won't do AV1. It's not what I want to be using for 5+ years.
On the very low end, they did recently release two low-powered (6 W) SKUs for devices such as fanless Chromebooks... but they are 28nm Excavator parts rather than Zen based. Noooooo!
MSI estimated that gaming laptop market is 5% out of 160mln. laptops sold per year. Majority of the market will be happy with 4 cores or less, as long as the laptop is snappy, looks good, and has good battery life. Afterall, Intel is also releasing many dual core and quadcore parts. So 7nm Renoir will be competitive in that main, non-gaming laptop market. Now, regarding the VCE- it would not be surprising for Renoir to support the latest codecs, even if it is Vega. I would not be discounting Renoir simply based on an early guess about VCE.
I can currently find Ryzen 7 3750H, the top AMD mobile chip, in ~$700 systems. They seem to have discrete GPUs, but I wouldn't call them gaming laptops per se. $700 is the price point I would want to pay for laptop to use for the next several years (and I like to use them right through their destruction). You can find decent ~$300-$400 laptops, but the best chip out of the lineup at $700 works and gives some future proofing.
I would be pleasantly surprised if the VCE supports AV1.
News was amd was getting a push into the laptop market and this appears to be the start of it. Good graphics, keyboard, panel.. Sure it might be a 1660TI but it's still a Ryzen onboard.
AMD's chips in mobile are currently not very power efficient, and they will still be using the older architecture and 14nm node until 4000 series at the very earliest.
That's not how much AMD makes, that's how much is left after expenses. It's pretty reasonable for a company that's heavily into R&D to spend a lot on that, and this sum probably also includes paying off debts, etc. I'd say that in general, a company that invests most of its money in its future is a healthy company. I think that it's better to look at the revenue and gross margin. That tells you how well AMD is doing, not the net income. Though AMD certainly has its work cut for it there. It's way under Intel and still under NVIDIA, though it's close enough to NVIDIA to hopefully surpass it at some point.
"It's pretty reasonable for a company that's heavily into R&D to spend a lot on that, and this sum probably also includes paying off debts, etc" AMD's rivals have the same issues but they solve them and stil they have a lot of money left at the end of the sums. AMD is winning on the market (share) just because they sell at a discount price with respect to its rivals. Should they apply the same margins others do they will be just have much less "success" despite their good momentum.
As a share of revenue, AMD spends 16% more on R&D than Intel does.
This is the reason for the discrepancy. If AMD spent as little as Intel, they would have fully ~$196M more in profit than Intel.
Remember also that Intel is engaged in more 'exploratory" markets, and a much higher share of their spending goes to 'waste' - things like paying for compiler cheats, etc. These benefit *relative* performance v. the competition, but they aren't actual gains in equity - just burning cash for market share.
This argument should been seen reversed. Seen the margin applied, AMD just has a discount price of 16% with respect to Intel. This low margin is the responsible to the fact that AMD has few money to invest in R&D to create better products. If they would place their product with the same margins they would loose all the advantages they are enjoying now. Seen that Intel is not going to stay at this performance level (4 years old) forever, AMD has to improve its product a lot to keep the pace Intel will set when completing its 7nm PP with all the kind of connections they can use (and AMD not).
If we want AMD to stick around, we need to encourage AMD to raise prices. Business 101. Competing on cost is not a long term winning strategy. Never has been, never will be.
Building trust into consumer by offering better product at competitive prices DOES work. Best example is KIA. Look at what they were offering in 2010... in 2014... and today. It took 10 years, but now the quality of Korean cars are on par with the Japanese. In 5 years, we might see second hand market price rise close to Japanese second hand cars.
Kia isn't a great example. But, we can go with it. Since Dec 31, 2010, Kia Motor's stock price (KRX) has dropped 13.5%. Kia's 2010 profit was 2.25 trillion won, Kia's 2018 profit was 1.2 trillion won, down 46.6%.
AMD as a company has two big competitive risks. (1) Intel could get 10 nm and then 7 nm working. AMD has used up its lower process size ammo and doesn't have much of any additional counters for that. If that happens, AMD won't have a compelling way to win on the high end. (2) Zhaoxin, ARM, or another company could make a usable CPU on the low end. AMD's low-price strategy could be undercut by another even lower priced company.
Yes, both risk are unlikely, but they are both quite feasible, especially in the long term.
Thus, it is quite possible for AMD to be bested on both sides, by performance and by price. That sandwich position is not a good long-term position to be in.
If they are at the same price, they wouldn't sell as the average consumer will just go with Intel as they have had Intel in their machines for years.
Same reason as why they go for Nvidia (though in Nvidia's case, AMD's GPU have not been competitive since the R9 200 series).
Competing on price/performance is what AMD should do and is what they're doing and its going well for them.
I am really interested how Q3 2019 goes, most people I know were waiting on that generation of Ryzen as single threaded performance for gaming was near on-par and the 3600 is insanely well priced (I and three other friends I know of waited to replace our 2600K and 4770K with them, I'm getting max 65W usage versus my ~150W on the 4770K with similar/better performance and no fans taking off).
You just reinforced what I said. If cost is the sole reason to buy AMD, then AMD has a losing strategy -- customers will go elsewhere the second they have the opportunity. AMD needs to have stronger reasons than cost for customers to buy them. AMD needs to work harder on differentiating their products, so that they are the default go to company, not Intel. And naming your chips with more cores mainstream instead of HEDT is just marketing fluff, not a real differentiating factor. Give us something Intel can't give us.
Low cost is only a successful strategy short term.
Traditionally, in the weeks ahead of a new cpu architecture launch, sales on current chips will dip--true for everyone in the business. The Q2 numbers don't include Zen 2 sales to date, which Su said a little while ago are already 3.5X their initial Zen + sales numbers. Q3 & 4 coming up should show markedly different and much more robust numbers, and on the server side, too. Personally, I'm all for AMD plowing every cent it can back into R&D--it isn't enough to get ahead--they have to stay ahead moving forward.
well they were actually losing money for years until the first zen came out. amd are coming from way behind. itll get better for them if they keep things going well. thats the story of amd. boom and bust
well they were actually losing money for years until the first zen came out. amd are coming from way behind. itll get better for them if they keep things going well. thats the story of amd. boom and bust.
plus something that has been and still is holding back is global foundries. they used to be a part of amd but amd sold them so they wouldnt be liable for the risk of running a foundry. but they have wafer supply aggreements with them so its basically like they still are taking the risk.
amd has to pay for so many wafers every year from them whether they buy them or not. that means that even if they have to sell something for nothing thats what they have to do. its really holding them back to an extent. the reason for the io die being on 14nm (or 12 or whatever) and the reason that 580 gpus are selling so cheap along with vega is because they made to many for the crypto boom but also because they have to keep buying wafers from glofo. most of the production has been moved over to tsmc which is a good move but they still have to buy wafers from glofo. what do you do with all those?
They have chosen the path of selling at low margins to get market share (that seems to be what everyone loos at when deciding the win or fail of a strategy). However the conditions that allow them to offer better CPU than Intel will not last forever. Seen that Intel has the laptop market in its hands (AMD's APUs are not a threat there) and is heavily concentrated in loosing as small as possible room in datacenter market, the only real win AMD is getting is on the desktop market which is shrinking. Yet, few % market share points are going to gave them a bit of money (while Intel is loosing almost nothing as seen in their own quarter report) but this is the time for AMD to make money. really. This is their most favorable period in the last 20 years seen the fails Intel has accumulated with their 10nm PP. When their 7nm will arrive the party will abruptly end for AMD.
Okay, Intel 7nm will arrive in 2022, or something. That should give AMD multiple years of partying. Intel has snatched the mobile market, though I was suprised to see Ryzen mobile 3000 in a local store. AMD has their main strength in desktop. The competition is about servers. 'Rome' servers from AMD should come soon and are very competive against 'Cascade lake.'
7nm will simply bring AMD where it was 2 years ago, that is in nowhere land. 10nm, when fully deployed, are enough to stop AMD "expansion". Unless before Intel actions AMD shareholders will ask the company to stop trading market share for revenues.
Intel's 7 nm is more like TSMC's 5 nm. Intel's 10 nm is more like TSMC's 7 nm. Don't expect Intel to continue to falter with their fabrication. They bit off more than they could chew and couldn't economically solve the yield issues they were having with 10 nm with the way they were going about it. 7 nm is likely to be a different ball game and it won't be delayed by their 10 nm struggles.
It's AMD which is in trouble going forward. They have another year+ of good times before things start to get rough for them. They are going to have to execute flawlessly to match the expectations the market has for them according to their share price.
Consider that Intel will most likely recover from their fabrication mishaps, that Intel is coming out with a series of improved cores, and that Intel is coming out with a discrete GPU that if successful could put pressure on AMD's GPU sales and margins. Intel can provide many things that AMD cannot, such as networking and storage, so I think there is a reasonable cap on what share AMD is likely to win in the data center market without having markedly superior processor options. And I don't expect AMD to maintain superiority in data center processing through 2020.
“Don’t expect that thing that has been promised thisYearTM for the past 4 years to be delayed forever”
The current expected state is that intel will never reach 10nm. Obviously they will but the chance they will get there at any given year is less than half.
Intel is already sending 10nm CPU for the laptop we are going to see in holidays period. These new CPUs will simply oversell AMD offer 20:1 which is enough for Intel to protect its laptop market against any discount AMD will do with its APUs for the rest of 2020, that is until the new architecture will be released.
4 cores laptop parts with poor yield maxing out at 3.9GHz with a boost around 4GHz. 10nm will never be high end. I doubt we are ever going to see server chips using it. It will cost a fortune for intel to produce 32 cores + CPUs. As long as Intel is monolithic, they will never match AMD on cost.
As if 10nm today yields will be the same for the rest of its life... AMD offers 4 core CPU that can reach... 3.7GHz! They have a huge GPU so will have these new Ice lake. Intel use lower power (at same perf) and are going to be placed everywhere, from low end to high end 2in1, covering the entire market bu the workstation/gaming one where they have other products that AMD has not. Can't see how the introduction of the 10nm CPU are going to worsen Intel position that already is of almost a complete monopoly in laptop market.
4GHz boost on a laptop part is pretty good, my 4720HQ doesn't go past about 3.6GHz and usually drops to like 2.8GHz all-core if I render. The 7700HQ that superseded it in current gen only boosts to 3.8GHz.
Only the latest 8750H supposedly hits 4.1GHz, but I bet that's only for a short window on one core, all core it probably throttles quite a bit.
Intel GPU do not need to be successful to simply destroy AMD APU segment and limiting Nvidia low level market monopoly. Those GPU will become the de-facto standard of the market and being 10% slower than the rivals (if ever slower) won't automatically make them a fail. Surpassing their performance/resource ratio will be very difficult and expensive for the rivals which will start selling at very discount price and probably abandoning the market (I'm thinking about Nvidia and its now obsolete proposals like GT1030 or even older boards that seems to still sell a lot). AMD APUs will need to find a niche to survive as none will be choosing them for performance (lower or similar to the competitions offer), consumption or features (they are quite behind Nvidia in all these aspects, and if Nvidia will struggle to compete against Intel Xe GPUs, AMD has no chance at all).
Ya, the last time they were in a favorable position was during the Athlon 64 and Athlon 64 X2 procs. During that time though, they went with high margins, instead of gathering market share.
That's why I didn't bother getting an X2. The cheapest one was double the price of the cheapest dual core from Intel. Regular Athlon 64 wasn't even an option for me, since it sucked with multitasking. Have one high demand task and your system was useless doing anything else, unlike the old single core HT enabled Intel procs.
They seem to be going the opposite route this time. Hopefully it works for them.
That is not true. Zen 1 was kicking Intel HEDT platform to the dust while being mainstream, and they were on parity process node. Intel process node advantage will never be a reality again. We are at the end of the road and they are now behind while they were having a 4 years lead. Samsung and TSMC are now the process node leaders, not Intel. You are making the assumption that AMD will not advance through that period of time. The rumors are that they will bring 4 threads per core, during the meantime Intel is trap with security vulnerabilities killing their performances and losing the trust of their customers. Google and Amazon are now considering EPYC way more seriously, and we know that Microsoft will probably use Zen for their Azure gaming servers.
Quick note for anybody reading this thread: This post is pretending reality is red. Not so. I suggest for anybody curious to revisit reviews to see how it really was...
A big part of the publishing of AMD's Q2 results was their weak guidance for Q3. Previously, AMD had maintained an optimistic view of the 2nd half of 2019 while Intel and NVIDIA and other semiconductor manufacturers had paired back their expectations for the back half of the year at their Q1 conference calls. Now AMD has cut their guidance for the rest of the year. That is most likely why their stock went down in after hours trading following the release of the earnings report press release.
Stockholders needs different numbers than a mere $52M of net revenue to give a meaning to the price the stock has reached. AMD cannot continue trading money for market share. Market share that will most probably been lost on next Intel return with a working PP. Someone just forgot that AMD actual products had to battle with Intel 10nm new architecture. Intel failed at providing it making this AMD product a little better than what Intel is offering on its now ancient 14nm PP. When Intel will full deploy 10nm and better with the 7nm advantages where they are probably exploiting all the interconnection they have created up to know, AMD advantage with small chiplets will simply evaporate and won't be able to sell its product but with a further price cut. Those $52 may become $100M in next quarters but they could stay that for next decade leaving AMD where it was 2 years ago.. in surviving mode (that's where Intel want and needs it to stay).
AMD can trade potential profit for market share, as its value would be branding. AMD's chiplet advantage will not go away for another gen or two, and AMD will not stay idle. Both Intel and AMD said they are working on 3D chiplet architecture, which will be interesting.
I bet within the next few years we will finally move away from the current motherboard design, it was done many decades ago and doesn't reflect current requirements as well anymore, especially as latency becomes ever more important due to increased power. We have already moved everything to pretty much PCIe.
"AMD can trade potential profit for market share, as its value would be branding." No if at the end you have miserable net profits with respect to the total revenue. Shareholders now want money, not the promises which have been given for almost a decade. It is time for AMD to get what they have been developing for all this time. If they are not able to that now with all the advantages they have, the chance they will can in future will shrink.
AMD are gaining market share in laptops fast already. It is just that laptop market is hard to break into, and AMD are quite passive in it- they are trying to just sell chips. Instead, they should be investing into building their own speed/quality standards and a good name in it, that could later carry and advertise them. Still, as long as AMD sell APUs cheaper than Intel, and offer some driver support- market share will continue to grow. Picasso APUs are great (but official tuning utilities, and stock settings on most laptops are not; unofficial Ryzen Controller fixes most of that). Of course, Renoir will be even better. But I'm not sure users will be benefiting from it, if it is pricier, and is placed in same poor laptop designs.
I have an honest question, does anybody actually own AMD notebook, but is out just to sway customer my impression of AMD ( maybe mostly fans ) is that all gaming and nothing but hate everywhere else against Intel. To me this attitude does not help AMD but actually hurts them.
I would not blame this one the vendor - but blamed it on demand. Plus stuff like lake of Thunderbolt 3 does not help.
yes.. i own 2 of them. and your impression is probably more bias, then wrong. TB3 ?? ever consider that this may not be a feature that most need, or want ?
I bought my dad an AMD powered laptop for christmas, because it was a great deal, performed amazingly, and had an NVMe SSD. The same laptop in Intel would have been too much for me to afford (He also only browses the internet with it). It's an incredible device with great battery life.
IF AMD got $10 every time someone said they're turning the corner and next quarter will be better, they could double their revenue, hahaha.
They're on the THIRD generation of Ryzen CPUs, they should be pumping up margins and raking in the dough by now...I don't see anytime in the future they'll have a better lineup than right now.
Your hope for AMD's ryzen sales may prove short lived. AMD has traditionally favored the value and value/enthusiast communities. The introduction of TR allowed them to get into the high market segment. With the price of the Radeon VII at $700 [2] and the price of 3950X at $750 [1] I, and I'm certain many other customers, are feeling betrayed. A GPU contains the physical chip, VRM, connectors, PCB, memory, and more. A CPU is just the CPU and a small PCB with a few small components. I bought my 6 core Phenom II 1090T (which I'm still running on), for < $200. "The last thing we would dream of doing is complain about pricing on these [AMD Thuban core] parts." -- Anand Lal Shimpi [3] I'm uncertain what AMD is trying to pull on us, be it that 5nm and 3nm will also double in price and therefore they have to get us used to it, or simply that AMD/Lisa Sue is feeling greedy these days, but I will not buy it. I wanted to, but the pricing is unconscionable, I might as well get a TR and a TR has more PCIe connections, and thus is currently more useful in the way of expansion and Ethernet.
None of us are feeling betrayed, the 3000 series Ryzen is amazing, the 3600 has got to be one of the best value for money in a very long time, it feels like recommending the 2600K again when people ask.
Yes, it is amazing, but I was commenting on the lack of amazing pricing, not amazing performance. I saved up the cash for the top end only to find out that the top end has risen from $329[1] to $749[2]! 227.657% of 2700X, yikes!!! Since when is a doubling in price good for the consumer?
and intels pricing of its top end cpus is any better ??? come on. you can ask Intel the SAME question each year it releases a new batch of cpu's. ryzen 3900X, $699 cdn, 9900K, $699 cdn as well, and just for contrast to that, the X series of intels cpus, $1210 to $2800 cdn. even the Threadripper 2990WX is less then the top X series cpu at $2500, and that gets you 14 more cores to boot.
I don't disagree with your latest post at all. Indeed, I'd love it if Intel sold it's processors at a better value for it's customers. My point of contention is with AMD choosing *not* to sell their products at a similarly good value to the customer as they did in the past. Chiplets were supposed to save money through lessening rejects and having more chips in the higher speed/silicon quality brackets. Why am I expected to pay so much more then? Even the 5700XT is a 251mm2 square chunk of 7nm silicon and is priced at $400[1], not to mention the components and GDDR6 RAM (the latest and greatest). Yet we pay $749 for 2 chunks each 80.7651mm2 [2] (see my comments below for the total silicon, it's not worth posting in 2 places in the same thread).
" I'd love it if Intel sold it's processors at a better value for it's customers. " and who wouldnt ? but intel wont, and they will charge what they can because ( at the time ) there was no alternative. but now, there is an alternative. and intel will have to either find a way to make their products worth the price they charge, or lower its pricing.
so what you are saying is, because what AMD charged for their products in the past, and now that they have products that, mostly with the cpu side, compete with intel very well, they dont have the right to charge the same amount as the intel equivalent is priced at ? i would say the value with amd, is the part that for the same price as the 9900k, which is 8core/12 threads, you get 4 more cores/8 more threads. comparing cores/threads, the 8 core version. 3800x, is 150 bucks less, and for the most part, are on par with the intel cpus in performance, correct ? as for the gpu side, i dont know what to say, as for me at least, paying the extra cash for a feature i wont use now, or for the time being, isnt worth it to me, aside from the part that the RTX series, is priced out of my reach. BTW, i WISH the 5700XT was 400 bucks, but here, its 530 for the entry level card. even the 5700 starts at 460
The TR crowd though, yeah, might be annoyed, but that's not really an issue, instead you should be happy that the next time you buy you're getting way more performance/dollar, and it has been rumored for months that the 3000 series was going to be quite a step up.
In terms of TR pricing, who knows yet, AMD can adjust to what they want. Interesting rumors of the top-end being a 64 core part, I wonder what the Epyc line-up will be.
I'm not at all upset on their account, they obviously have money and it's good for AMD to break into their market. I was waiting patiently as anyone who did not have the funds for TR could do. My comment on them was to point out how AMD is pricing the 3000 series as though it were TR and with less PCIe lanes.
" My comment on them was to point out how AMD is pricing the 3000 series as though it were TR and with less PCIe lanes. " thats funny.. cause it isnt priced that way, maybe it is just in the USA ?
I can't say what's it's like in your country, but here's what it's like here: The TR 1920X 12 core had an MSRP of $799 with the 1950X 16 core being $999[1]. It has since dropped to $266.49[2] as has the 16 core part, but to $469.99 [3]. The TR 2920X 12 core had an MSRP of $649 with the 2950X 16 core being $899[4]. It has since dropped to $599.99[5] as has the 16 core part, but to $665.00[6]. AMD's 3900X 12 core has an MSRP of $499 and the 3950X 16 core has an MSRP of $749[7]. Therefore, the pricing of the higher end parts in the US is simmilar to the pricing of TR, even at the original MSRP.
well.. for here.. add a good 200 bucks to each of those prices, and as i mentioned in the other post, TR starts at 860 for the 2920x, and the 3900x is the same price as the 9900k, and IF i was going to upgrade now, i would probably get then 3800x, or if i was able to swing it, the 3900x. as for the 19xx TRs, the 1950x is 699, 1900x, 499, but with TR2 out, why by TR1?? is MSRP even followed when new comp hardware comes out ??
Q1: I was offering options, not saying I would buy TR1. Q2: No, but I'm comparing across time for the purpose of assessing value of the current non-TR components.
Read the rest of my post. $700 for GPU that has a lot more 7nm silicon (and other parts) (331mm2)[1] then the CPU and $649 for the CPU[2]. Even the total silicon is smaller.
161.5302mm2 7nm silicon 122.6512mm2 14nm silicon +___________ 284.1814mm2 Total silicon
Also, AMD has been a great value buy for years as I proved in my first post. Hence, betrayal.
its not about living is a rosy garden. my reply to you, could be copied and pasted here. amd was the value cpu option, mainly because they were outperformed by even intels lower end parts, but now that they are on par, and even faster then intel, they should be allowed to charge appropriately for their products, as the performance is there, correct ?
Correct. BUT, it's not about "allowed" (or permitted, etc.), it's about the customers. Me, you, everyone else. "Mankind was my business. The common good was my business," said Marley to Scrooge IIRC. By the same token, if AMD is still in business for the customers, which I, perhaps foolishly, assumed to be the case, then they aught to remain the value option.
i think they are still the value option, but just not in the same ( pricing ) way. the value now, is more for the same ( or less ) they intel charges, IE 8 cores for 150 bucks less, or 4 more cores for the same price as intels 9900k :-)
maybe you are just used to AMD charging the prices they did, cause they had no choice because the performance wasnt there, but now that can charge more and its a little shocking to see AMD being able to charge more now ? :-)
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.
113 Comments
Back to Article
webdoctors - Wednesday, July 31, 2019 - link
$35M? It's weird, AMD is a global company making highly complicated cutting edge tech that's used everywhere, kicking Intel's butt in desktop segment and they only made $35M the last quarter? Sure its better than a loss but still.It's crazy a movie or music star made more money than this global company of 10K employees. Ppl can't accuse them of rolling in dough haha.
imaheadcase - Wednesday, July 31, 2019 - link
No one has accused them of rolling in the dough. Not sure why you think they are beating intel butt in desktop market..you do know just because they have a new chip out doesn't automatically mean instant win? Lots of people have contracts with intel, prefer intel chips. Intel still rules in the SFF market and high end gaming.AshlayW - Wednesday, July 31, 2019 - link
Erm, no. Ryzen 3000 essentially made Intel's entire lineup, except for MAYBE 9900K, completely pointless at their current prices. You can put a 3900X in mini ITX X470, and later an 3950X. SFF is covered lol. 3900X gives 9900K a run for its money in "high end gaming" So much so you would be hard pressed to notice the difference, does it vastly more efficiently, and oblierates in other tasks: while costing the same and including a cooler capable of sustaining impressive boost clocks. You also get a platform that isn't dead, and if you go with X570, new features that Intel doesn't even support.People who "prefer Intel chips" as you put it, are tools with Ryzen 3000 on the market. If you can't see the butt getting kicked, you are blind, sorry. Even my 2700X makes half the Intel lineup pointless.
heffeque - Wednesday, July 31, 2019 - link
Probably will have to ask for a new server on one of the projects I manage. Purchasing Dept will surely get me Intel.I'm happy that you know that AMD has the better price/performance out there, but I'm guessing that you are on your teens if you think that that will instantly make everyone buy AMD.
You also seem to forget that AMD still sucks in laptop CPUs and that most people nowadays buy laptops.
Wish things were as easy as you see them.
R0H1T - Thursday, August 1, 2019 - link
>You also seem to forget that AMD still sucks in laptop CPUs and that most people nowadays buy laptops.You either haven't seen the reviews of Ryzen 2xxx U/H & 3xxx U/H series or basically made this up just to make your point! AMD lags in the single core dept for sure however they more than make up for it in the graphics department, the only (major) problem for them is the core count & probably not producing enough chips. If however they do have the capacity for more production, then I can only assume Intel is back to their old ways of <you know what> like we saw recently with that user "benchmark" tweak 😑
Gondalf - Wednesday, July 31, 2019 - link
Not even Charlie of Semiaccurate is unfair like you.After an analysis of 3000 serie (leaving away expensive low volume 3900 and 3950 cpus), he said that at the end it is only a matter of selling price, and on per core basis the two cpus are comparable.
Bel intel will lower the prices at an AMD level soon.
goatfajitas - Wednesday, July 31, 2019 - link
Yes, Ryzen is great and if I were building today I would definitely buy a Ryzen 3000 series, but he is referring to sales, not performance. In sales, Intel is still winning. It takes more than a year or two of good performance to gain marketshare, especially with the high volume enterprise market... It takes consistency. Dont forget the really great Ryzen (3xxx seies) just came out this summer. I am sure in a few more years AMD will gain more share.eva02langley - Wednesday, July 31, 2019 - link
No they are not. the DIY PC and the HEDT platform is now AMD.Intel sales are OEM, prebuilts, laptops, chromebooks and server.
Next year will be detrimental for AMD. They needs to disrupt Intel markets if they want to consolidate their emphasis on the semiconductors level. I do believe AMD will double their actual size as a company, but they need to disrupt Intel other markets if they want to grow even more. They need more than just taking market share, they need to take and create markets.
CiccioB - Wednesday, July 31, 2019 - link
They need to make money, and to do that they have to offer better product which can sustain high margins.Up to now CPUs margins are good but GPUs are a hole that swallows all the profits.
They have to stop that if they want to really grow and not being left at surviving in a corner.
eva02langley - Wednesday, July 31, 2019 - link
Their new product line was not even incorporated in these results. They did really well with over Ryzen 2000, Vega and Polaris for having those numbers. Google is rumored to ditch Intel. If that happens, a lot of other companies are going to follow.trivik12 - Friday, August 2, 2019 - link
That is bullshit. Intel makes custom chips for Google. That relationship is very strong. They were also early access for Optane memory.Google/Facebook/Amazon all will also make their own SOC/ASIC etc. Still majority of servers/cloud will run on Intel for sure
goatfajitas - Wednesday, July 31, 2019 - link
Agreed, but DIY and HEDT is a tiny peice. That is what I was saying. AMD does have plans in the other markets and OEM wins as well... Next year we will see 7nm laptop chips that will help. With that said AMD will not likely ever dominate. They can however be very healthy and competitive if they keep executing.HStewart - Saturday, August 3, 2019 - link
DIY PC's is a dyeing market - with very limited appeal. I better attitude for AMD would be support it fan based instead trying to disrupt the completion. Such a negative attitude makes customers like myself don't desire or trust them. Honest Truth.Korguz - Tuesday, August 6, 2019 - link
maybe where you are.. but i have my doubts its dying as much as you claim., if it was, then there wouldnt be so many boards, or cpu's released as there is.heh.. more like your blind fanaticism towards intel is why you dont desire or trust them, HStewart.
bobhumplick - Wednesday, July 31, 2019 - link
it made intels old lineup irrelevant. but intels old lineup is...well old. if they release cometlake with an 8 core 16 thread i7 instead of i9 and an 8 core 8 thread i5 (the old i7) they basically become very relevant again. things change year to year and sometimes month to month. when amd gets a win everybody declares the death of intel. and then a couple months later when intel comes out with something better nobody has anything say then. just like how an 8700k clocked to 5ghz even beat amds 8 cores from the first gen, even in multithreaded apps. itll go back and forth between them. intel really need to get on to 10nm for desktop but they can compete on 14nm for a bit longerjabber - Thursday, August 1, 2019 - link
Folks on PC enthusiasts just see things from their perspective.Form the perspective of the CTO of Ford/GM/Pfizer/Google/Bank of America...they will be buying those Dell or HP desktops with Intel inside in the millions.
Your valiant purchase of a 3000 series was cute but it wont change a thing.
Xyler94 - Tuesday, August 6, 2019 - link
Ford and them don't give 2 flying fucks what CPU is in the machine. They care if Dell can supply them with a machine, it just happens to be powered by Intel.The problem isn't that AMD's CPUs aren't what people are after, it's that big builders like Dell aren't selling them.
yeeeeman - Wednesday, July 31, 2019 - link
AMD needs to keep executing like this for a few more years before they can gain momentum and increase their revenues.CiccioB - Wednesday, July 31, 2019 - link
Problem is that this kind of advantages it is enjoying now, that is the issues Intel has had with its 10nm, is probably not going to repeat in the future.Next year with this product and Intel in its full power is not going to provide more money than today.
eva02langley - Wednesday, July 31, 2019 - link
The thing is Intel was always having the process node advantage, now it is gone for good. It will never happen again. TSMC is already rolling out EUV 7nm, 5nm and 3nm is on track. Intel will never get this back.CiccioB - Wednesday, July 31, 2019 - link
I'm not denying that Intel more advance PP has given it a big help in the past.But actual PP are really expensive, both for production and for development, and AMD, even though using TSMC advanced PP, is not guaranteed to always have this advantage ove Intel.
You are one of those forgetting that actual Intel offering is a restyling of something 4 years old, while 10nm is still not fully deployed and 7nm is coming.
With those PP there will be changes bigger than simple shrinks.
Intel has a new architecture to roll out (and it would already be here against Zen if 10nm were ok) and new interconnections allows for things that AMD cannot do even with a newer PP.
So enjoy AMD small advantage when comparing its latest products with Intel 4 years old ones, but be prepared to a fast change of the market when Intel will strike back at full power.
RSAUser - Thursday, August 1, 2019 - link
Most of Intel's advantages on their "old" architecture come from tweaking that 14nm so much that it will actually be difficult for their 7/10nm to compete against it for at least a generation on higher power parts.AMD will probably have the performance advantage till 2021 or so, there I expect it to even out again if AMD does not improve.
We have yet to see all of AMD's architecture improvements with this generation, the 3950X will be quite interesting, and AMD will not release their best chips now as they don't need to. Most of the 3000 series gains seem to be from the improved process node, rather than the architecture, so I'm expecting good gains again from 4000 series.
Verneer is rumored to bring 4 thread SMT, but I'd only expect them to implement this in the server environment if they do, doesn't really make sense in the consumer space. I bet on them splitting server and desktop with 4000 series.
MASSAMKULABOX - Thursday, August 1, 2019 - link
Never say never..willis936 - Wednesday, July 31, 2019 - link
AMD is already on par with intel. They would need to keep their already built momentum for a few more years to make intel a totally irrelevant option.Gondalf - Wednesday, July 31, 2019 - link
Assuming Intel sleeping for few more years. Obviously this will not happen :).At some point Intel will do a check, can i manufacture?? yes? ok, no? ok i'll will begin to manufacture at TSMC and Samsung, leaving AMD in the dust.
It takes only less than one year to convert strategical production at foundries thanks to actual advanced software tools developed by TSMC and others.
eva02langley - Wednesday, July 31, 2019 - link
You need to work on your uarch to be manufactured at other foundries. It is not as simple as this. Also, Intel is investing massively in their fabs and these results will not bring process node forward, only production. If the product is not good, it will not matter anyway.niva - Wednesday, July 31, 2019 - link
I doubt Intel will stay behind AMD for long, and even though AMD is ahead of them now in terms of architecture design and manufacturing node, it won't stay that way for long. AMD has not made Intel obsolete even with it's newest chips, and when you factor in drivers/stability and years of conditioning that Intel is just better it's hard for me to call this a win. Intel products are not bad, now there's actual competition. It's good to see Intel on the receiving end of a price war, but AMD isn't banking yet and it's going to take a while for the money to really start flowing in, probably a lot longer for Intel to put out their next generation of CPUs and turn this around again.I've loved and used AMD chips since the 90s, but I'm shocked that the company has remained in business somehow. Now they're being squeezed from the other direction by Arm chips and the future is just as uncertain as it always has been.
eva02langley - Wednesday, July 31, 2019 - link
At this pace, it will take 3 years for AMD to pay off their debt. I really hope they do so. If they crash again, I believe it will be the end for goods, however I don't think it will happens. They have promising uarchs and they are really getting momentum in a lot of markets. I am still really curious to see what Samsung will come out with RDNA.CiccioB - Wednesday, July 31, 2019 - link
Can you describe which math you used to calculate a 3 year debt repayment?eva02langley - Wednesday, July 31, 2019 - link
https://wccftech.com/amd-meets-earnings-estimates-...y/y 392M$ debt repaid --> 1300M$ in debt.
RSAUser - Thursday, August 1, 2019 - link
I doubt AMD will ever be gone for good, the US would bail them out as Army rules dictate that there must be at least 2 suppliers.IlllI - Wednesday, July 31, 2019 - link
It is lack of laptop sales. PC/desktop sales are in decline, but laptops are still going strong, of which intel has a massive majority in market share.CiccioB - Wednesday, July 31, 2019 - link
Unfortunately for AMD it has very weak laptop offer.They just have those APUs which are good for medium sized laptop and nothing for ultra slim ones.
More over, they do not have a single GPU that is power efficient enough to replace Nvidia discrete offer in that market.
RSAUser - Thursday, August 1, 2019 - link
Well, if they price it well, then AMD GPU can compete. Remember, the desktop GPUs are way overlcocked beyond the efficiency curve. They can sell for cost price and make a profit on the CPU, would help branding a lot if they can sell all AMD machines.MASSAMKULABOX - Thursday, August 1, 2019 - link
SO, sell at cost , and make it up in Volume ..I see /sCiccioB - Friday, August 2, 2019 - link
This quarter at $35M on a total revenue of more than $1500M just shows that AMD is alreading selling at cost.Do you want them to go in red again? If they ever do at this point with these products I doubt it can ever become profitable for the rest of its (would be short) life.
nandnandnand - Wednesday, July 31, 2019 - link
The lack of 6 and 8 core options for laptops is glaring.I'm not in for an AMD laptop APU until they have 8 cores and AV1 decoding on the GPU. Navi's media decoder on the desktop GPUs does not support AV1, so I could be skipping until at least 2021.
CiccioB - Wednesday, July 31, 2019 - link
It's not only a problem of n# of cores.AMD has not SKU good for ultra thin laptop or 2in1. And the integrated APU makes it a bad choice for high end ones that need a (big) discrete GPU.
Intel and Nvidia have the entire market of such devices, and they will until AMD will make APU with smaller iGPU and more core for high end laptop or office oriented ones or smaller packaged version for ultra thin client.
nandnandnand - Wednesday, July 31, 2019 - link
I'll concede that's it's not just about # of cores. But that's a big deficiency given that Intel has 8 cores on mobile and is rumored to be going for 10 cores. If I want a top of the line AMD mobile Picasso APU, the Ryzen 7 3750H, that's just 4 cores. The Renoir successor may not even have Navi graphics, and even if it does, Navi won't do AV1. It's not what I want to be using for 5+ years.On the very low end, they did recently release two low-powered (6 W) SKUs for devices such as fanless Chromebooks... but they are 28nm Excavator parts rather than Zen based. Noooooo!
neblogai - Wednesday, July 31, 2019 - link
MSI estimated that gaming laptop market is 5% out of 160mln. laptops sold per year. Majority of the market will be happy with 4 cores or less, as long as the laptop is snappy, looks good, and has good battery life. Afterall, Intel is also releasing many dual core and quadcore parts. So 7nm Renoir will be competitive in that main, non-gaming laptop market.Now, regarding the VCE- it would not be surprising for Renoir to support the latest codecs, even if it is Vega. I would not be discounting Renoir simply based on an early guess about VCE.
nandnandnand - Wednesday, July 31, 2019 - link
I can currently find Ryzen 7 3750H, the top AMD mobile chip, in ~$700 systems. They seem to have discrete GPUs, but I wouldn't call them gaming laptops per se. $700 is the price point I would want to pay for laptop to use for the next several years (and I like to use them right through their destruction). You can find decent ~$300-$400 laptops, but the best chip out of the lineup at $700 works and gives some future proofing.I would be pleasantly surprised if the VCE supports AV1.
https://slickdeals.net/newsearch.php?src=SearchBar...
just4U - Saturday, August 3, 2019 - link
I dunno.. you ask me this is a pretty damn good gaming laptop with Ryzen 3750..https://www.newegg.com/gold-steel-asus-tuf-gaming-...
News was amd was getting a push into the laptop market and this appears to be the start of it. Good graphics, keyboard, panel.. Sure it might be a 1660TI but it's still a Ryzen onboard.
RSAUser - Thursday, August 1, 2019 - link
AMD's chips in mobile are currently not very power efficient, and they will still be using the older architecture and 14nm node until 4000 series at the very earliest.They are not competitive in mobile.
ET - Wednesday, July 31, 2019 - link
That's not how much AMD makes, that's how much is left after expenses. It's pretty reasonable for a company that's heavily into R&D to spend a lot on that, and this sum probably also includes paying off debts, etc. I'd say that in general, a company that invests most of its money in its future is a healthy company. I think that it's better to look at the revenue and gross margin. That tells you how well AMD is doing, not the net income. Though AMD certainly has its work cut for it there. It's way under Intel and still under NVIDIA, though it's close enough to NVIDIA to hopefully surpass it at some point.CiccioB - Wednesday, July 31, 2019 - link
"It's pretty reasonable for a company that's heavily into R&D to spend a lot on that, and this sum probably also includes paying off debts, etc"AMD's rivals have the same issues but they solve them and stil they have a lot of money left at the end of the sums. AMD is winning on the market (share) just because they sell at a discount price with respect to its rivals.
Should they apply the same margins others do they will be just have much less "success" despite their good momentum.
willis936 - Wednesday, July 31, 2019 - link
You think AMD rivals have done anything other than slash R&D expenses over the last decade?CiccioB - Wednesday, July 31, 2019 - link
Huh?AMD rivals have made lots of money up to now.
What do yo mean with "slash R&D"?
What do you think stock companies intents are?
willis936 - Wednesday, July 31, 2019 - link
Read my comment again.Sahrin - Wednesday, July 31, 2019 - link
As a share of revenue, AMD spends 16% more on R&D than Intel does.This is the reason for the discrepancy. If AMD spent as little as Intel, they would have fully ~$196M more in profit than Intel.
Remember also that Intel is engaged in more 'exploratory" markets, and a much higher share of their spending goes to 'waste' - things like paying for compiler cheats, etc. These benefit *relative* performance v. the competition, but they aren't actual gains in equity - just burning cash for market share.
CiccioB - Wednesday, July 31, 2019 - link
This argument should been seen reversed.Seen the margin applied, AMD just has a discount price of 16% with respect to Intel.
This low margin is the responsible to the fact that AMD has few money to invest in R&D to create better products.
If they would place their product with the same margins they would loose all the advantages they are enjoying now. Seen that Intel is not going to stay at this performance level (4 years old) forever, AMD has to improve its product a lot to keep the pace Intel will set when completing its 7nm PP with all the kind of connections they can use (and AMD not).
RSAUser - Thursday, August 1, 2019 - link
Intel has those margins as they are still using a tech that's been paid off for years.dullard - Wednesday, July 31, 2019 - link
If we want AMD to stick around, we need to encourage AMD to raise prices. Business 101. Competing on cost is not a long term winning strategy. Never has been, never will be.eva02langley - Wednesday, July 31, 2019 - link
Building trust into consumer by offering better product at competitive prices DOES work. Best example is KIA. Look at what they were offering in 2010... in 2014... and today. It took 10 years, but now the quality of Korean cars are on par with the Japanese. In 5 years, we might see second hand market price rise close to Japanese second hand cars.It does work and it is working right now.
dullard - Wednesday, July 31, 2019 - link
Kia isn't a great example. But, we can go with it. Since Dec 31, 2010, Kia Motor's stock price (KRX) has dropped 13.5%. Kia's 2010 profit was 2.25 trillion won, Kia's 2018 profit was 1.2 trillion won, down 46.6%.dullard - Wednesday, July 31, 2019 - link
AMD as a company has two big competitive risks. (1) Intel could get 10 nm and then 7 nm working. AMD has used up its lower process size ammo and doesn't have much of any additional counters for that. If that happens, AMD won't have a compelling way to win on the high end. (2) Zhaoxin, ARM, or another company could make a usable CPU on the low end. AMD's low-price strategy could be undercut by another even lower priced company.Yes, both risk are unlikely, but they are both quite feasible, especially in the long term.
Thus, it is quite possible for AMD to be bested on both sides, by performance and by price. That sandwich position is not a good long-term position to be in.
dullard - Wednesday, July 31, 2019 - link
The sooner AMD gets the "we have a premium product and will charge a premium price" idea, the better their long term prospects will be.RSAUser - Thursday, August 1, 2019 - link
If they are at the same price, they wouldn't sell as the average consumer will just go with Intel as they have had Intel in their machines for years.Same reason as why they go for Nvidia (though in Nvidia's case, AMD's GPU have not been competitive since the R9 200 series).
Competing on price/performance is what AMD should do and is what they're doing and its going well for them.
I am really interested how Q3 2019 goes, most people I know were waiting on that generation of Ryzen as single threaded performance for gaming was near on-par and the 3600 is insanely well priced (I and three other friends I know of waited to replace our 2600K and 4770K with them, I'm getting max 65W usage versus my ~150W on the 4770K with similar/better performance and no fans taking off).
dullard - Friday, August 2, 2019 - link
You just reinforced what I said. If cost is the sole reason to buy AMD, then AMD has a losing strategy -- customers will go elsewhere the second they have the opportunity. AMD needs to have stronger reasons than cost for customers to buy them. AMD needs to work harder on differentiating their products, so that they are the default go to company, not Intel. And naming your chips with more cores mainstream instead of HEDT is just marketing fluff, not a real differentiating factor. Give us something Intel can't give us.Low cost is only a successful strategy short term.
WaltC - Wednesday, July 31, 2019 - link
Traditionally, in the weeks ahead of a new cpu architecture launch, sales on current chips will dip--true for everyone in the business. The Q2 numbers don't include Zen 2 sales to date, which Su said a little while ago are already 3.5X their initial Zen + sales numbers. Q3 & 4 coming up should show markedly different and much more robust numbers, and on the server side, too. Personally, I'm all for AMD plowing every cent it can back into R&D--it isn't enough to get ahead--they have to stay ahead moving forward.bobhumplick - Wednesday, July 31, 2019 - link
well they were actually losing money for years until the first zen came out. amd are coming from way behind. itll get better for them if they keep things going well. thats the story of amd. boom and bustbobhumplick - Wednesday, July 31, 2019 - link
well they were actually losing money for years until the first zen came out. amd are coming from way behind. itll get better for them if they keep things going well. thats the story of amd. boom and bust.plus something that has been and still is holding back is global foundries. they used to be a part of amd but amd sold them so they wouldnt be liable for the risk of running a foundry. but they have wafer supply aggreements with them so its basically like they still are taking the risk.
amd has to pay for so many wafers every year from them whether they buy them or not. that means that even if they have to sell something for nothing thats what they have to do. its really holding them back to an extent. the reason for the io die being on 14nm (or 12 or whatever) and the reason that 580 gpus are selling so cheap along with vega is because they made to many for the crypto boom but also because they have to keep buying wafers from glofo. most of the production has been moved over to tsmc which is a good move but they still have to buy wafers from glofo. what do you do with all those?
CiccioB - Wednesday, July 31, 2019 - link
They have chosen the path of selling at low margins to get market share (that seems to be what everyone loos at when deciding the win or fail of a strategy).However the conditions that allow them to offer better CPU than Intel will not last forever.
Seen that Intel has the laptop market in its hands (AMD's APUs are not a threat there) and is heavily concentrated in loosing as small as possible room in datacenter market, the only real win AMD is getting is on the desktop market which is shrinking.
Yet, few % market share points are going to gave them a bit of money (while Intel is loosing almost nothing as seen in their own quarter report) but this is the time for AMD to make money. really. This is their most favorable period in the last 20 years seen the fails Intel has accumulated with their 10nm PP. When their 7nm will arrive the party will abruptly end for AMD.
Rudde - Wednesday, July 31, 2019 - link
Okay, Intel 7nm will arrive in 2022, or something. That should give AMD multiple years of partying.Intel has snatched the mobile market, though I was suprised to see Ryzen mobile 3000 in a local store. AMD has their main strength in desktop. The competition is about servers. 'Rome' servers from AMD should come soon and are very competive against 'Cascade lake.'
CiccioB - Wednesday, July 31, 2019 - link
7nm will simply bring AMD where it was 2 years ago, that is in nowhere land.10nm, when fully deployed, are enough to stop AMD "expansion". Unless before Intel actions AMD shareholders will ask the company to stop trading market share for revenues.
Yojimbo - Wednesday, July 31, 2019 - link
Intel's 7 nm is more like TSMC's 5 nm. Intel's 10 nm is more like TSMC's 7 nm. Don't expect Intel to continue to falter with their fabrication. They bit off more than they could chew and couldn't economically solve the yield issues they were having with 10 nm with the way they were going about it. 7 nm is likely to be a different ball game and it won't be delayed by their 10 nm struggles.It's AMD which is in trouble going forward. They have another year+ of good times before things start to get rough for them. They are going to have to execute flawlessly to match the expectations the market has for them according to their share price.
Consider that Intel will most likely recover from their fabrication mishaps, that Intel is coming out with a series of improved cores, and that Intel is coming out with a discrete GPU that if successful could put pressure on AMD's GPU sales and margins. Intel can provide many things that AMD cannot, such as networking and storage, so I think there is a reasonable cap on what share AMD is likely to win in the data center market without having markedly superior processor options. And I don't expect AMD to maintain superiority in data center processing through 2020.
willis936 - Wednesday, July 31, 2019 - link
“Don’t expect that thing that has been promised thisYearTM for the past 4 years to be delayed forever”The current expected state is that intel will never reach 10nm. Obviously they will but the chance they will get there at any given year is less than half.
Yojimbo - Wednesday, July 31, 2019 - link
They are already there.CiccioB - Wednesday, July 31, 2019 - link
Intel is already sending 10nm CPU for the laptop we are going to see in holidays period.These new CPUs will simply oversell AMD offer 20:1 which is enough for Intel to protect its laptop market against any discount AMD will do with its APUs for the rest of 2020, that is until the new architecture will be released.
eva02langley - Wednesday, July 31, 2019 - link
4 cores laptop parts with poor yield maxing out at 3.9GHz with a boost around 4GHz. 10nm will never be high end. I doubt we are ever going to see server chips using it. It will cost a fortune for intel to produce 32 cores + CPUs. As long as Intel is monolithic, they will never match AMD on cost.CiccioB - Wednesday, July 31, 2019 - link
As if 10nm today yields will be the same for the rest of its life...AMD offers 4 core CPU that can reach... 3.7GHz! They have a huge GPU so will have these new Ice lake.
Intel use lower power (at same perf) and are going to be placed everywhere, from low end to high end 2in1, covering the entire market bu the workstation/gaming one where they have other products that AMD has not.
Can't see how the introduction of the 10nm CPU are going to worsen Intel position that already is of almost a complete monopoly in laptop market.
RSAUser - Thursday, August 1, 2019 - link
4GHz boost on a laptop part is pretty good, my 4720HQ doesn't go past about 3.6GHz and usually drops to like 2.8GHz all-core if I render. The 7700HQ that superseded it in current gen only boosts to 3.8GHz.Only the latest 8750H supposedly hits 4.1GHz, but I bet that's only for a short window on one core, all core it probably throttles quite a bit.
CiccioB - Wednesday, July 31, 2019 - link
Intel GPU do not need to be successful to simply destroy AMD APU segment and limiting Nvidia low level market monopoly.Those GPU will become the de-facto standard of the market and being 10% slower than the rivals (if ever slower) won't automatically make them a fail. Surpassing their performance/resource ratio will be very difficult and expensive for the rivals which will start selling at very discount price and probably abandoning the market (I'm thinking about Nvidia and its now obsolete proposals like GT1030 or even older boards that seems to still sell a lot).
AMD APUs will need to find a niche to survive as none will be choosing them for performance (lower or similar to the competitions offer), consumption or features (they are quite behind Nvidia in all these aspects, and if Nvidia will struggle to compete against Intel Xe GPUs, AMD has no chance at all).
khanikun - Wednesday, July 31, 2019 - link
Ya, the last time they were in a favorable position was during the Athlon 64 and Athlon 64 X2 procs. During that time though, they went with high margins, instead of gathering market share.That's why I didn't bother getting an X2. The cheapest one was double the price of the cheapest dual core from Intel. Regular Athlon 64 wasn't even an option for me, since it sucked with multitasking. Have one high demand task and your system was useless doing anything else, unlike the old single core HT enabled Intel procs.
They seem to be going the opposite route this time. Hopefully it works for them.
eva02langley - Wednesday, July 31, 2019 - link
That is not true. Zen 1 was kicking Intel HEDT platform to the dust while being mainstream, and they were on parity process node. Intel process node advantage will never be a reality again. We are at the end of the road and they are now behind while they were having a 4 years lead. Samsung and TSMC are now the process node leaders, not Intel. You are making the assumption that AMD will not advance through that period of time. The rumors are that they will bring 4 threads per core, during the meantime Intel is trap with security vulnerabilities killing their performances and losing the trust of their customers. Google and Amazon are now considering EPYC way more seriously, and we know that Microsoft will probably use Zen for their Azure gaming servers.I am not seeing the same way at all.
Klimax - Friday, August 2, 2019 - link
Quick note for anybody reading this thread: This post is pretending reality is red. Not so. I suggest for anybody curious to revisit reviews to see how it really was...Yojimbo - Wednesday, July 31, 2019 - link
A big part of the publishing of AMD's Q2 results was their weak guidance for Q3. Previously, AMD had maintained an optimistic view of the 2nd half of 2019 while Intel and NVIDIA and other semiconductor manufacturers had paired back their expectations for the back half of the year at their Q1 conference calls. Now AMD has cut their guidance for the rest of the year. That is most likely why their stock went down in after hours trading following the release of the earnings report press release.CiccioB - Wednesday, July 31, 2019 - link
Stockholders needs different numbers than a mere $52M of net revenue to give a meaning to the price the stock has reached.AMD cannot continue trading money for market share.
Market share that will most probably been lost on next Intel return with a working PP.
Someone just forgot that AMD actual products had to battle with Intel 10nm new architecture. Intel failed at providing it making this AMD product a little better than what Intel is offering on its now ancient 14nm PP.
When Intel will full deploy 10nm and better with the 7nm advantages where they are probably exploiting all the interconnection they have created up to know, AMD advantage with small chiplets will simply evaporate and won't be able to sell its product but with a further price cut.
Those $52 may become $100M in next quarters but they could stay that for next decade leaving AMD where it was 2 years ago.. in surviving mode (that's where Intel want and needs it to stay).
CiccioB - Wednesday, July 31, 2019 - link
I correct, $35M of net revenue, not 52.RSAUser - Thursday, August 1, 2019 - link
AMD can trade potential profit for market share, as its value would be branding.AMD's chiplet advantage will not go away for another gen or two, and AMD will not stay idle. Both Intel and AMD said they are working on 3D chiplet architecture, which will be interesting.
I bet within the next few years we will finally move away from the current motherboard design, it was done many decades ago and doesn't reflect current requirements as well anymore, especially as latency becomes ever more important due to increased power. We have already moved everything to pretty much PCIe.
CiccioB - Friday, August 2, 2019 - link
"AMD can trade potential profit for market share, as its value would be branding."No if at the end you have miserable net profits with respect to the total revenue.
Shareholders now want money, not the promises which have been given for almost a decade.
It is time for AMD to get what they have been developing for all this time. If they are not able to that now with all the advantages they have, the chance they will can in future will shrink.
5080 - Wednesday, July 31, 2019 - link
A new Ryzen on 7nm for notebooks can't come fast enough for AMD. The desktop market is just too small now to make any significant market gains.neblogai - Wednesday, July 31, 2019 - link
AMD are gaining market share in laptops fast already. It is just that laptop market is hard to break into, and AMD are quite passive in it- they are trying to just sell chips. Instead, they should be investing into building their own speed/quality standards and a good name in it, that could later carry and advertise them.Still, as long as AMD sell APUs cheaper than Intel, and offer some driver support- market share will continue to grow. Picasso APUs are great (but official tuning utilities, and stock settings on most laptops are not; unofficial Ryzen Controller fixes most of that). Of course, Renoir will be even better. But I'm not sure users will be benefiting from it, if it is pricier, and is placed in same poor laptop designs.
HStewart - Saturday, August 3, 2019 - link
I have an honest question, does anybody actually own AMD notebook, but is out just to sway customer my impression of AMD ( maybe mostly fans ) is that all gaming and nothing but hate everywhere else against Intel. To me this attitude does not help AMD but actually hurts them.I would not blame this one the vendor - but blamed it on demand. Plus stuff like lake of Thunderbolt 3 does not help.
Korguz - Tuesday, August 6, 2019 - link
yes.. i own 2 of them. and your impression is probably more bias, then wrong.TB3 ?? ever consider that this may not be a feature that most need, or want ?
Xyler94 - Tuesday, August 6, 2019 - link
How often do you actually use Thunderbolt?I bought my dad an AMD powered laptop for christmas, because it was a great deal, performed amazingly, and had an NVMe SSD. The same laptop in Intel would have been too much for me to afford (He also only browses the internet with it). It's an incredible device with great battery life.
webdoctors - Wednesday, July 31, 2019 - link
IF AMD got $10 every time someone said they're turning the corner and next quarter will be better, they could double their revenue, hahaha.They're on the THIRD generation of Ryzen CPUs, they should be pumping up margins and raking in the dough by now...I don't see anytime in the future they'll have a better lineup than right now.
RSAUser - Thursday, August 1, 2019 - link
They won't up the margins as they would rather sell more CPUs.Where margins are good is laptops, but AMD is not competitive there.
ballsystemlord - Wednesday, July 31, 2019 - link
Your hope for AMD's ryzen sales may prove short lived. AMD has traditionally favored the value and value/enthusiast communities. The introduction of TR allowed them to get into the high market segment. With the price of the Radeon VII at $700 [2] and the price of 3950X at $750 [1] I, and I'm certain many other customers, are feeling betrayed.A GPU contains the physical chip, VRM, connectors, PCB, memory, and more. A CPU is just the CPU and a small PCB with a few small components.
I bought my 6 core Phenom II 1090T (which I'm still running on), for < $200. "The last thing we would dream of doing is complain about pricing on these [AMD Thuban core] parts." -- Anand Lal Shimpi [3]
I'm uncertain what AMD is trying to pull on us, be it that 5nm and 3nm will also double in price and therefore they have to get us used to it, or simply that AMD/Lisa Sue is feeling greedy these days, but I will not buy it.
I wanted to, but the pricing is unconscionable, I might as well get a TR and a TR has more PCIe connections, and thus is currently more useful in the way of expansion and Ethernet.
[1]: https://www.anandtech.com/show/14605/the-and-ryzen...
[2]: https://www.anandtech.com/show/13923/the-amd-radeo...
[3]: https://www.anandtech.com/print/3675/tigerdirect-o...
RSAUser - Thursday, August 1, 2019 - link
None of us are feeling betrayed, the 3000 series Ryzen is amazing, the 3600 has got to be one of the best value for money in a very long time, it feels like recommending the 2600K again when people ask.ballsystemlord - Saturday, August 3, 2019 - link
Yes, it is amazing, but I was commenting on the lack of amazing pricing, not amazing performance. I saved up the cash for the top end only to find out that the top end has risen from $329[1] to $749[2]! 227.657% of 2700X, yikes!!!Since when is a doubling in price good for the consumer?
[1]: https://www.anandtech.com/show/12625/amd-second-ge...
[2]: https://www.anandtech.com/show/14605/the-and-ryzen...
Korguz - Tuesday, August 6, 2019 - link
and intels pricing of its top end cpus is any better ??? come on. you can ask Intel the SAME question each year it releases a new batch of cpu's. ryzen 3900X, $699 cdn, 9900K, $699 cdn as well, and just for contrast to that, the X series of intels cpus, $1210 to $2800 cdn. even the Threadripper 2990WX is less then the top X series cpu at $2500, and that gets you 14 more cores to boot.ballsystemlord - Tuesday, August 6, 2019 - link
I don't disagree with your latest post at all. Indeed, I'd love it if Intel sold it's processors at a better value for it's customers.My point of contention is with AMD choosing *not* to sell their products at a similarly good value to the customer as they did in the past. Chiplets were supposed to save money through lessening rejects and having more chips in the higher speed/silicon quality brackets. Why am I expected to pay so much more then?
Even the 5700XT is a 251mm2 square chunk of 7nm silicon and is priced at $400[1], not to mention the components and GDDR6 RAM (the latest and greatest). Yet we pay $749 for 2 chunks each 80.7651mm2 [2] (see my comments below for the total silicon, it's not worth posting in 2 places in the same thread).
[1]: https://www.anandtech.com/show/14618/the-amd-radeo...
[2]: https://www.anandtech.com/show/13852/amd-no-chiple...
Korguz - Wednesday, August 7, 2019 - link
" I'd love it if Intel sold it's processors at a better value for it's customers. " and who wouldnt ? but intel wont, and they will charge what they can because ( at the time ) there was no alternative. but now, there is an alternative. and intel will have to either find a way to make their products worth the price they charge, or lower its pricing.so what you are saying is, because what AMD charged for their products in the past, and now that they have products that, mostly with the cpu side, compete with intel very well, they dont have the right to charge the same amount as the intel equivalent is priced at ? i would say the value with amd, is the part that for the same price as the 9900k, which is 8core/12 threads, you get 4 more cores/8 more threads. comparing cores/threads, the 8 core version. 3800x, is 150 bucks less, and for the most part, are on par with the intel cpus in performance, correct ? as for the gpu side, i dont know what to say, as for me at least, paying the extra cash for a feature i wont use now, or for the time being, isnt worth it to me, aside from the part that the RTX series, is priced out of my reach. BTW, i WISH the 5700XT was 400 bucks, but here, its 530 for the entry level card. even the 5700 starts at 460
ballsystemlord - Wednesday, August 7, 2019 - link
See my reply where you said a similar thing below.RSAUser - Thursday, August 1, 2019 - link
The TR crowd though, yeah, might be annoyed, but that's not really an issue, instead you should be happy that the next time you buy you're getting way more performance/dollar, and it has been rumored for months that the 3000 series was going to be quite a step up.In terms of TR pricing, who knows yet, AMD can adjust to what they want. Interesting rumors of the top-end being a 64 core part, I wonder what the Epyc line-up will be.
ballsystemlord - Saturday, August 3, 2019 - link
I'm not at all upset on their account, they obviously have money and it's good for AMD to break into their market. I was waiting patiently as anyone who did not have the funds for TR could do. My comment on them was to point out how AMD is pricing the 3000 series as though it were TR and with less PCIe lanes.Korguz - Tuesday, August 6, 2019 - link
" My comment on them was to point out how AMD is pricing the 3000 series as though it were TR and with less PCIe lanes. " thats funny.. cause it isnt priced that way, maybe it is just in the USA ?ballsystemlord - Tuesday, August 6, 2019 - link
I can't say what's it's like in your country, but here's what it's like here:The TR 1920X 12 core had an MSRP of $799 with the 1950X 16 core being $999[1]. It has since dropped to $266.49[2] as has the 16 core part, but to $469.99 [3].
The TR 2920X 12 core had an MSRP of $649 with the 2950X 16 core being $899[4]. It has since dropped to $599.99[5] as has the 16 core part, but to $665.00[6].
AMD's 3900X 12 core has an MSRP of $499 and the 3950X 16 core has an MSRP of $749[7].
Therefore, the pricing of the higher end parts in the US is simmilar to the pricing of TR, even at the original MSRP.
* A few of these processors are on sale and the price will rise back up sooner or later.
[1]: https://www.anandtech.com/show/11697/the-amd-ryzen...
[2]: https://www.newegg.com/p/274-000K-001E8?Descriptio...
[3]: https://www.newegg.com/amd-ryzen-threadripper-1950...
[4]: https://www.anandtech.com/show/13124/the-amd-threa...
[5]: https://www.newegg.com/amd-ryzen-threadripper-2920...
[6]: https://www.newegg.com/p/274-000K-001X0?Descriptio...
[7]: https://www.anandtech.com/show/14605/the-and-ryzen...
Korguz - Wednesday, August 7, 2019 - link
well.. for here.. add a good 200 bucks to each of those prices, and as i mentioned in the other post, TR starts at 860 for the 2920x, and the 3900x is the same price as the 9900k, and IF i was going to upgrade now, i would probably get then 3800x, or if i was able to swing it, the 3900x. as for the 19xx TRs, the 1950x is 699, 1900x, 499, but with TR2 out, why by TR1?? is MSRP even followed when new comp hardware comes out ??ballsystemlord - Wednesday, August 7, 2019 - link
Q1: I was offering options, not saying I would buy TR1.Q2: No, but I'm comparing across time for the purpose of assessing value of the current non-TR components.
Qasar - Thursday, August 1, 2019 - link
" and I'm certain many other customers, are feeling betrayed. " how so ?? please explain.ballsystemlord - Saturday, August 3, 2019 - link
Read the rest of my post. $700 for GPU that has a lot more 7nm silicon (and other parts) (331mm2)[1] then the CPU and $649 for the CPU[2]. Even the total silicon is smaller.161.5302mm2 7nm silicon
122.6512mm2 14nm silicon
+___________
284.1814mm2 Total silicon
Also, AMD has been a great value buy for years as I proved in my first post. Hence, betrayal.
[1]: https://www.anandtech.com/show/13923/the-amd-radeo...
[2]: https://www.anandtech.com/show/13852/amd-no-chiple...
Qasar - Tuesday, August 6, 2019 - link
then it just must be you, as i dont feel betrayed.ballsystemlord - Tuesday, August 6, 2019 - link
You can live in your rosy garden as you please, I'll stop disturbing you.Korguz - Wednesday, August 7, 2019 - link
its not about living is a rosy garden. my reply to you, could be copied and pasted here. amd was the value cpu option, mainly because they were outperformed by even intels lower end parts, but now that they are on par, and even faster then intel, they should be allowed to charge appropriately for their products, as the performance is there, correct ?ballsystemlord - Wednesday, August 7, 2019 - link
Correct.BUT, it's not about "allowed" (or permitted, etc.), it's about the customers. Me, you, everyone else. "Mankind was my business. The common good was my business," said Marley to Scrooge IIRC. By the same token, if AMD is still in business for the customers, which I, perhaps foolishly, assumed to be the case, then they aught to remain the value option.
Korguz - Thursday, August 8, 2019 - link
i think they are still the value option, but just not in the same ( pricing ) way. the value now, is more for the same ( or less ) they intel charges, IE 8 cores for 150 bucks less, or 4 more cores for the same price as intels 9900k :-)ballsystemlord - Thursday, August 8, 2019 - link
I understand. That's a reasonable position. I still disagree. But that's my opinion, and your's is equally arguable.Korguz - Friday, August 9, 2019 - link
maybe you are just used to AMD charging the prices they did, cause they had no choice because the performance wasnt there, but now that can charge more and its a little shocking to see AMD being able to charge more now ? :-)ballsystemlord - Friday, August 9, 2019 - link
I agree.ballsystemlord - Friday, August 9, 2019 - link
Addendum: A 16 core EPYC (7282), is cheaper then the 16 core ryzen! If only EPIC were overclockable....Korguz - Saturday, August 10, 2019 - link
:-) i wouldnt mind playing with one of the 64 core epyc's my self :-)