AMD to introduce its Radeon R9 300-series lineup at Computex

[desprado]

Sources with knowledge of Advanced Micro Devices’ roadmap have revealed that the company plans to introduce its new family of Radeon R9 300-series graphics cards closer to Computex than CeBIT, as originally planned. AMD had hoped to push out one or two new cards in March and then re-brand many others to refresh the range. That is no longer the plan.
The challenge for Advanced Micro Devices is that it needs to achieve two rather different (and sometimes opposing) tasks in the first half of this year. It must significantly reduce inventory in the channel and it has to try and boost its dropping market share. If AMD does not ship anything new, its own sales and market share will decline. However, if AMD ships brand-new offerings, older hardware stalls in the channel, impacts prices and makes it difficult to sell new products.
According to sources from the Far East, AMD originally wanted to unleash at least one of its all-new Radeon R9 300-series graphics card at the CeBIT trade-show in early March. These new cards would feature the company’s latest high-end GPU and high-bandwidth memory (HBM). If AMD’s engineers have done a good job, then this Radeon R9 290-series replacement would be the company’s fastest card ever, especially with the latest games running ultra-high-definition resolutions like 4K or 5K. If that happens, then the launch would significantly disrupt sales of all other Radeon R9 series products, so the company decided not to unveil its new top-of-the-range graphics boards this month.
Once AMD has a new Radeon R9-class product ready, it could change the schedule again, but it seems that the company does not want to launch its new product as soon as possible. In the wake of the GeForce GTX 970 memory allocation issue for Nvidia, a new launch might have helped AMD capture market share, but it now seems to be taking other factors into account.
The sources with knowledge of AMD plans revealed to KitGuru that Computex has been targeted as the launch timeframe for AMD’s upcoming graphics adapters and a “full line up” was demanded from senior management and engineers. Instead of launching just one new graphics solution and a load of re-brands, AMD intends to roll-out a whole new range of graphics cards at Computex in early June. At present details about the lineup are not clear, but expect several all-new graphics solutions to be unveiled, whereas some re-branded products will likely show up in the following months.
While AMD does not comment on reports about product plans and launch schedules, the company indirectly confirmed that its new graphics processing units will be out rather later than sooner during the Morgan Stanley Technology, media and telecom conference this week. Apparently, AMD considers second half of this year as a time when it will try to regain lost market share from Nvidia.
“We are confident that as we get into the second half of 2015 with the launch of that [new graphics] product, we will gain back the market share which is low from my standpoint and historically,” said Devinder Kumar, chief financial officer of AMD. “We need to be significantly higher than where we are right now.”
KitGuru Says: The reason why AMD prefers to delay introduction of new graphics processing units seems could be rather simple: sales dynamics of Radeon R9 280- and 290-series is not very good in the first quarter and AMD’s partners need additional time to get rid of them. Still, while AMD’s plan regarding launch timeframe seems to be clear now, the details about the exact lineup will probably remain in doubt until the last moment. KitGuru will continue to investigate AMD’s intentions in the coming weeks

http://www.kitguru.net/components/g...-its-radeon-r9-300-series-lineup-at-computex/

 

[Techhog]

That last part worries me... since it means that this entire thread will be about AMD going out of business next year. Other than that, why can't they just release the 390 and 390X, then firesale the 280(X) and 290(X)? That gets rid of them either way.

 

[desprado]

That last part worries me... since it means that this entire thread will be about AMD going out of business next year. Other than that, why can't they just release the 390 and 390X, then firesale the 280(X) and 290(X)? That gets rid of them either way.

This can be just a rumor and i hope tomorrow AMD shows there line up so that Nvidia price dont go crazy.

 

[SmCaudata]

Too bad if true. I will likely want an upgrade sooner than June. I feel this will hurt marketshare. People who want current Gen likely already bought or are waiting for next gen to drive down prices. Delays won't clear inventory.

 

[S.H.O.D.A.N.]

Other than that, why can't they just release the 390 and 390X, then firesale the 280(X) and 290(X)? That gets rid of them either way.

Presumably because the 390 isn't ready, while the 380 is.

It's also a case of alienating partners with existing stock. You'd force everyone to fire sale. People don't like being told to cut margins into oblivion.

It's really a question of how the 200 series sales look right now. If they keep a steady pace that's bringing in profit while emptying warehouses, then it's all fine and dandy, and the delay makes sense. If the sales are not good, they won't get better just because AMD delays the release of 300.

 

[gradoman]

Looking forward to this! Too bad about the rest of that news though.

 

[S.H.O.D.A.N.]

What doesn't make much sense in this rumor is the idea that, at some point, AMD was perfectly fine with launching a new arch 380 in March and filling the rest of the line with re brands until the 390 shows up in June and then someone said NOPE and told the engineering team to prepare a full new arch stack for launch three months later with no re branding of old designs.

That's a bit of a stretch, ain't it?

 

[Valantar]

It's also a case of alienating partners with existing stock. You'd force everyone to fire sale. People don't like being told to cut margins into oblivion.

It's really a question of how the 200 series sales look right now. If they keep a steady pace that's bringing in profit while emptying warehouses, then it's all fine and dandy, and the delay makes sense. If the sales are not good, they won't get better just because AMD delays the release of 300.

Exactly. As much as I was hoping for a new AMD card launch this month, I'm seeing this in a positive light. We've seen far too many half-assed rebrands from AMD the last couple of years. They need new tech across a broad range of their line-up. And sure, slashing R9 2XX prices would probably clear inventory, but lose money both for AMD and its partners. As of today, the R9 2XX series delivers a good price/performance ratio, it just falls short on power efficiency. That's not really a problem as long as partners ship cards with decent coolers (given that most gamers seem to have massively overpowered PSUs). In other words, they can clear out stocks in the mainstream market all the while not losing money, and afterwards they could launch an all-new, broad range of RX 3XX cards for both mainstream and high end. Seems like a solid plan. Parallell launches for both mainstream and high-end are a good tactic - budget minded buyers still want to feel like they're getting something new without waiting months on end, and like not to get a rebranded model or last year's tech gussied up to look shiny and new.

 

[Head1985]

This is very very very very VERY bad for AMD if TRUE:'(

 

[Flapdrol1337]

I don't get why they wouldn't release the top end card if it's ready to go. Could just price it high so it doesn't compete with the 290X.

 

[RussianSensation]

What doesn't make much sense in this rumor is the idea that, at some point, AMD was perfectly fine with launching a new arch 380 in March and filling the rest of the line with re brands until the 390 shows up in June and then someone said NOPE and told the engineering team to prepare a full new arch stack for launch three months later with no re branding of old designs.

That's a bit of a stretch, ain't it?

No, it makes so much sense! AMD realized they can't just release 390/390X as only 2 new SKUs so all 8000 engineers will sit down and just make an entirely new laptop and desktop lineup of ALL new products in 3-4 months and manufacture hundreds of thousands of them.

/ sarcasm.

The only parts about the article that make sense are AMD waiting to launch the entire lineup and clearing existing inventory. However, anyone who thinks AMD will go back to the drawing board and make some new products in 3-4 months is clearly delusional. The only thing you can do at this point are another respin to try to increase clocks or lower voltage or decide just how much to cut down the higher end parts. From an engineering point of view, the entire R9 300 stack is more or less finalized. You can't just design a 400-550mm2 chip and decide that ya it's not good enough so lets redo it in time for Computex. What an insane theory in the article.

It's possible there are supply constraints with HBM and AMD wants to see where Titan X lands in performance. Last time AMD embarrassed the Titan and 780 with a $399 R9 290 and I am sure AMD will want to repeat this again. Why R9 290 failed miserably were mining prices, lack of after-market AIB cards for launch, and supply shortages. It's better if AMD spends another 3 months finalizing clock speeds, working with AIBs on after-market cards and having good supply at launch. If all 3 of these are met and R9 390X drops at $550-600 with solid performance it will make the Titan X look silly at $1K+, except this time supply constraints and mining inflated prices won't damage the amazing price/performance AMD usually brings to the table with new generations of cards.

 

[dangerman1337]

Presumably because the 390 isn't ready, while the 380 is.

It's also a case of alienating partners with existing stock. You'd force everyone to fire sale. People don't like being told to cut margins into oblivion.

It's really a question of how the 200 series sales look right now. If they keep a steady pace that's bringing in profit while emptying warehouses, then it's all fine and dandy, and the delay makes sense. If the sales are not good, they won't get better just because AMD delays the release of 300.

If the market share is anything I think delaying is just making things worse. They are not going to sell all of those 290s while the 970 (for all of its flaws) and 980 are taking shares in the markets and the minds (why AMD kept producing and sending them off this year is beyond me, should of culled the 290 and replaced it with a GF 28nm SHP 1.2 GCN Hawaii or if that is what Grenada is at least). Those 290s are not going to be cleared at the current prices and they should just say "**** it, do a firesale and bear the costs", its not like they are having profits on any of the 290s.

 

[Stuka87]

This is very very very very VERY bad for AMD if TRUE:'(

Hope this was sarcasm....

What AMD needs to make happen is to have the 300 series on a smaller node. There have been rumors of them going from 28nm to 16FF, although no clue how true those may be. But here is hoping.

 

[Temuka]

Article about nothing... If they really want to clear out 290 and 290x they should price it something like 250 for 290 and 300$ for 290x,but with current prices (and please don't even tell me about rebates and 1 part sale times) they will fail to sale when there is 970 at 320-330$

 

[raghu78]

The only parts about the article that make sense are AMD waiting to launch the entire lineup and clearing existing inventory. However, anyone who thinks AMD will go back to the drawing board and make some new products in 3-4 months is clearly delusional. The only thing you can do at this point are another respin to try to increase clocks or lower voltage or decide just how much to cut down the higher end parts. From an engineering point of view, the entire R9 300 stack is more or less finalized. You can't just design a 400-550mm2 chip and decide that ya it's not good enough so lets redo it in time for Computex. What an insane theory in the article.

It's possible there are supply constraints with HBM and AMD wants to see where Titan X lands in performance. Last time AMD embarrassed the Titan and 780 with a $399 R9 290 and I am sure AMD will want to repeat this again. Why R9 290 failed miserably were mining prices, lack of after-market AIB cards for launch, and supply shortages. It's better if AMD spends another 3 months finalizing clock speeds, working with AIBs on after-market cards and having good supply at launch. If all 3 of these are met and R9 390X drops at $550-600 with solid performance it will make the Titan X look silly at $1K+, except this time supply constraints and mining inflated prices won't damage the amazing price/performance AMD usually brings to the table with new generations of cards.

I agree. :thumbsup: Developing these GPUs takes years of work (atleast 2 -3) so nothing was decided in a quarter or two. The top to bottom product stack launch and clearing out existing inventory are very valid reasons. Also supply is going to be crucial for AMD to cater to the demand as the idea is to gain back the lost market share and do some more. AMD needs sufficient product volume on launch. So they are most probably building up a good volume and setting it all up for late Q2 at Computex.

This was expected as Lisa Su and Devinder Kumar said Q1 will be inventory correction and Q2 will be a return to normal business, which implies the newer products will launch in Q2. The existing R9 2xx product stack is outdated and woefully behind the competition in terms of current market perception (perf/watt).

 

[Subyman]

I'd think they would have known about stock many, many months before now. The only explanation for waiting would be the R9 290 sales have tanked enormously and there are still tons of them, far more than expectations. However, if sales have tanked then that is even more reason to release the new product. Holding back a viable product so an older product can trickle off the shelves for months doesn't make much sense.

 

[Hdgamer]

Yet AMD was raking in the cash during the mining craze. The reason why the sales are low is because it's been to long since they introduced the last cards and now you can find them second hand cheap. They need something new on the market. That's why Nvidia is killing them now. If they just released their new cards they would be fine. Marketshare would go back up.

 

[SimianR]

If there is lots of old R9 290/x inventory then a rebadge actually makes a lot of sense, even if you phase it out eventually. I know some people will disagree, but hey, a new game bundle and a rebadge could do a lot for sales (sadly ).

 

[R0H1T]

Yet AMD was raking in the cash during the mining craze. The reason why the sales are low is because it's been to long since they introduced the last cards and now you can find them second hand cheap. They need something new on the market. That's why Nvidia is killing them now. If they just released their new cards they would be fine. Marketshare would go back up.

No they weren't, sure they sold as many 290(x) as they possibly could but the miners & retailers were making way more than what AMD ended up with.

If I had to guesstimate the miners made the most of the mining boom, then came certain retailers & lastly AMD.

The best they could've done back then was to employ their own mega farm to mine litecoin et al, I bet they would've made way more money with that as compared to selling the GPU's directly

Also the used cards basically killed the sales of new 290(x) after a while especially when the mining craze tanked D:

 

[geoxile]

Interesting claim, a lineup of new cards rather than just rebrands. Hopefully that Hawaii refresh will be a big revision, with some feature updates. The memory compression could definitely help. They should just eat the losses and hurry up and launch the new cards.

 

[RussianSensation]

I'd think they would have known about stock many, many months before now. The only explanation for waiting would be the R9 290 sales have tanked enormously and there are still tons of them, far more than expectations. However, if sales have tanked then that is even more reason to release the new product. Holding back a viable product so an older product can trickle off the shelves for months doesn't make much sense.

No. NV had 120,000 GTX570s to clear. The average going price at the time on Newegg was at least $300 or $36 million dollars. If AMD were to release a better card at $249, no one will want 290/290X.

http://www.kitguru.net/components/g...-gtx570-overstock-delays-gtx660-until-august/

It also makes a lot of sense to build up supply/inventory, get AIB cards ready, get solid drivers out of the box. There a lot of factors to delay 300 series launch. Frankly nothing AMD did since HD4870 worked for them. I think AMD should just focus on 2 main consumers only:

Brand agnostic and price/performance. They need to stop worrying about market share all that nonsense. Just make profitable products that are good for the intended market. Once Mercedes realized to stop catering to non-sense demands of BMW drivers, they gave us the amazing S and C class cars. AMD needs to stop caring about NV's customers and just focus on what they do best. When Mercedes finally realized who gives 2 ****s about BMW drivers is when they embraced who they are and made the best cars they ever made! The end result is BMW drivers keep buying BMW and Mercedes made far superior products for their intended market. Customers of both brands are happy.

 

[Vesku]

Presumably because the 390 isn't ready, while the 380 is.

It's also a case of alienating partners with existing stock. You'd force everyone to fire sale. People don't like being told to cut margins into oblivion.

It's really a question of how the 200 series sales look right now. If they keep a steady pace that's bringing in profit while emptying warehouses, then it's all fine and dandy, and the delay makes sense. If the sales are not good, they won't get better just because AMD delays the release of 300.

They can't be that good given the market share changes in the last couple of quarters. I'm thinking the AIBs basically revolted and said they won't sell the top tier card unless they get decent replacements for the bread and butter categories $150-350. In other words they won't sell the flagship until AMD gives them enough full tonga (960 competitor) and cut down flagship (970 competitor) chips. Such a situation would nicely explain a ~3 month delay as the time needed to test, package and deliver the larger quantity SKUs.

 

[boozzer]

No, it makes so much sense! AMD realized they can't just release 390/390X as only 2 new SKUs so all 8000 engineers will sit down and just make an entirely new laptop and desktop lineup of ALL new products in 3-4 months and manufacture hundreds of thousands of them.

/ sarcasm.

The only parts about the article that make sense are AMD waiting to launch the entire lineup and clearing existing inventory. However, anyone who thinks AMD will go back to the drawing board and make some new products in 3-4 months is clearly delusional. The only thing you can do at this point are another respin to try to increase clocks or lower voltage or decide just how much to cut down the higher end parts. From an engineering point of view, the entire R9 300 stack is more or less finalized. You can't just design a 400-550mm2 chip and decide that ya it's not good enough so lets redo it in time for Computex. What an insane theory in the article.

It's possible there are supply constraints with HBM and AMD wants to see where Titan X lands in performance. Last time AMD embarrassed the Titan and 780 with a $399 R9 290 and I am sure AMD will want to repeat this again. Why R9 290 failed miserably were mining prices, lack of after-market AIB cards for launch, and supply shortages. It's better if AMD spends another 3 months finalizing clock speeds, working with AIBs on after-market cards and having good supply at launch. If all 3 of these are met and R9 390X drops at $550-600 with solid performance it will make the Titan X look silly at $1K+, except this time supply constraints and mining inflated prices won't damage the amazing price/performance AMD usually brings to the table with new generations of cards.

Right on. 3 to 4 months to make brand new line up ahahahahhaahhaa so freaking hilarious.

 

[Vesku]

Right on. 3 to 4 months to make brand new line up ahahahahhaahhaa so freaking hilarious.

To deliver the chips already planned rather than solely the flagship. Flagship initial quantity needed to launch is quite a bit smaller than for cheaper SKUs. Sounds like AMD wanted their partners to launch the flagship while keeping the rest of the old lineup for another quarter or two, maybe even 3 this is AMD afterall, with at best some clock bumps before eventually delivering new lower SKU chips. One way for the AIB partners to ensure they get the higher volume chips in a timely manner would be to refuse to sell the flagship until they are delivered.

 

[RussianSensation]

Right on. 3 to 4 months to make brand new line up ahahahahhaahhaa so freaking hilarious.

The article is clueless. NV themselves admitted Titan X took THOUSANDS of Human engineers-hours to make. All of a sudden we are supposed to believe AMD can re-engineer their entire 300 series line-up on in 3-4 months just because senior management said so? I can't believe at the stupidity of the journalist integrity for actually writing such garbage. Clearly an article aimed at getting ad revenue clicks because it creates controversy. BS written for ad revenue detected.