Speculation: Ryzen 3000 series

[exquisitechar]

First, their cost won't be double that since it uses the same packaging and IO-die.
Second, they need to entice enthusiast to go for the $429 or $499 part.
Third, I expect the 16C part to be thermally limited (lower all-core speeds) so the MT performance increase wont be that much more compared to 12C. Say 20% instead of 33%.

I expect only on 7nm EUV (does TSMC call that 5nm?) with 12nm IO die will the full potential of 16C be unlocked.

I think that TSMC 7nm+ uses EUV, 5nm comes after.

All good arguments, for sure. I don't doubt that they could price the fastest 16C at $500 and make money, but they could also get away with making it more expensive. I'm open to both possibilities.

 

[krumme]

Can you test the chiplet/io die before it's packaged?

 

[scannall]

Can you test the chiplet/io die before it's packaged?

Yes

 

[DrMrLordX]

Because the 240GE at $75 and 3300 6C/12T at $99 makes it impossible to fit the quad core APUs in the lineup.

AMD has probably fed less than 100k each of 200, 220, and 240GE APUs into OEM channels. They've likely signed contracts with various OEMs to receive a certain number of chips by X date with possible options to receive more in the event of brisk sales or what have you. Several reasons make AMD launching a hypothetical 6c/12t chip at $100 in June (actually more likely September, since the low-end chips will be last to market on AM4) have no bearing on the fate of the GE chips:

1). AMD has, at most, a few million in potential profits on the line when it comes to the GE chips. They're limited-run chips from failed Raven Ridge dice. Barely a blip even on AMD's potential revenue roster. They could bag on their entire remaining stock AND write down all the unsold GE inventory right now without feeling much of a sting. If they wait about five months and do the same thing, it'll hurt even less. Remember that these are chips they could have just thrown away since they probably couldn't sell them as full Raven Ridge APUs otherwise (for whatever reason).
2). AMD can just leave GE APUs to linger in OEM inventories until they slowly sell out or get sent to liquidation and cut off production to prevent further orders. Slightly more profitable than option 1) since there would be no write-down. OEMs won't really be crying either due to the low volumes involved. It's not like AMD needs to be in any rush to get that stuff out of the channel since there's so little of it and since it doesn't compete with Zen2 in any fashion.
3). Most buyers looking at OEM boxes with a GE APU in them will not hear of Ryzen 3 or anything 7nm until next year and won't see Picasso until maybe September. They won't even be aware that there's a 6c/12t chip for $100. It won't be on their radar.

A tiny percentage of AMD's total customer base will know about the $75 240GE AND $100 6c/12t Zen2 chip. A tiny percentage of that group will actually consider a 240GE, briefly (from an OEM tray or whatever).

There is no problem.

A customer buys an AMD computer with gazillion cores and a slow hard drive.

Said customer is frustrated and never buy an AMD computer again.

Great job!

Remember the frustration people felt with poorly-equipped Kaveri and Carrizo laptops in days of yore?

 

[beginner99]

6C 12T Ryzen 3 can easily have MSRP of $99-$120

8C 16T Ryzen 3 at MSRP of $199 is also easily achievable using the same single chiplet SKU.

And? if you can price the 8c at $120 more (=$319) and still be competitive both in absolute performance and pricing, why would you sell it at $199? The 9900k will be invalidated 100% even if you sell at $320 or even $400.

There is a reason NV and Intel have >50% margins and AMD only around 35%. I mean I personally hope that you are right of course and I'm wrong so I can buy a $200 8-core that competes with a $500 9900k. But for AMD that would be bad.They would loose out on future sales.

 

[scannall]

AMD has probably fed less than 100k each of 200, 220, and 240GE APUs into OEM channels. They've likely signed contracts with various OEMs to receive a certain number of chips by X date with possible options to receive more in the event of brisk sales or what have you. Several reasons make AMD launching a hypothetical 6c/12t chip at $100 in June (actually more likely September, since the low-end chips will be last to market on AM4) have no bearing on the fate of the GE chips:

1). AMD has, at most, a few million in potential profits on the line when it comes to the GE chips. They're limited-run chips from failed Raven Ridge dice. Barely a blip even on AMD's potential revenue roster. They could bag on their entire remaining stock AND write down all the unsold GE inventory right now without feeling much of a sting. If they wait about five months and do the same thing, it'll hurt even less. Remember that these are chips they could have just thrown away since they probably couldn't sell them as full Raven Ridge APUs otherwise (for whatever reason).
2). AMD can just leave GE APUs to linger in OEM inventories until they slowly sell out or get sent to liquidation and cut off production to prevent further orders. Slightly more profitable than option 1) since there would be no write-down. OEMs won't really be crying either due to the low volumes involved. It's not like AMD needs to be in any rush to get that stuff out of the channel since there's so little of it and since it doesn't compete with Zen2 in any fashion.
3). Most buyers looking at OEM boxes with a GE APU in them will not hear of Ryzen 3 or anything 7nm until next year and won't see Picasso until maybe September. They won't even be aware that there's a 6c/12t chip for $100. It won't be on their radar.

A tiny percentage of AMD's total customer base will know about the $75 240GE AND $100 6c/12t Zen2 chip. A tiny percentage of that group will actually consider a 240GE, briefly (from an OEM tray or whatever).

There is no problem.



Remember the frustration people felt with poorly-equipped Kaveri and Carrizo laptops in days of yore?

I bought a Trinity laptop. And after tossing an SSD in it, and bumping the ram it performs really well. It was just a shame it came so poorly equipped from the factory.

 

[Timorous]

And? if you can price the 8c at $120 more (=$319) and still be competitive both in absolute performance and pricing, why would you sell it at $199? The 9900k will be invalidated 100% even if you sell at $320 or even $400.

There is a reason NV and Intel have >50% margins and AMD only around 35%. I mean I personally hope that you are right of course and I'm wrong so I can buy a $200 8-core that competes with a $500 9900k. But for AMD that would be bad.They would loose out on future sales.

If AMD are going to be releasing products similar to those leaks, at prices similar to those leaks then I think it is a strategy to get far more OEM take up than they have had so far. Once the OEMs have invested in their infrastructure to support more of your products it makes it easier going forward and then AMD can start raising their margins. I think the idea, assuming the leaks are in the ballpark, is to use the opportunity Intel have given them to forge long lasting OEM relationships to increase the number of OEM offerings. The benefit is that when Intel get their house in order AMD will have greater OEM support than if they go for pure margins right now.

 

[PotatoWithEarsOnSide]

And? if you can price the 8c at $120 more (=$319) and still be competitive both in absolute performance and pricing, why would you sell it at $199? The 9900k will be invalidated 100% even if you sell at $320 or even $400.

There is a reason NV and Intel have >50% margins and AMD only around 35%. I mean I personally hope that you are right of course and I'm wrong so I can buy a $200 8-core that competes with a $500 9900k. But for AMD that would be bad.They would loose out on future sales.

It's easy to push up margins when you dominate the market. AMD are not in that position yet, so they need to manouver themselves as best they can. Once the market share is there, then push margins up a little.
Longer term, it is better to have lower margins on higher volume, than higher margins on low volume.
Whilst I think that cannibalising future sales is a valid argument, I suspect that it impacts Intel far greater than it does AMD; in the first instance they are cannibalising Intel current and future sales, not their own.

 

[Shivansps]

Yet Athlon 220 and 240 seem to be telling the oposite, AMD was too afraid that the Athlons would take too much of the 2200G/2400G sales to the point they made the 220 and 240GE worthless. They were unwilling to increase the IGP freq by 1mhz not to impact 2200G sales, so i find hard to belive they will sudenly change and target Intel no matter the cost.

 

[Mopetar]

The point of going with the chiplet design is to mitigate the price increase associated with 7nm.

Yes, you managed to mitigate the price increase, but it's still a price increase.

I'm not sure if it necessarily is an increase once you consider all factors. I don't know if the slide is completely accurate, but it paints a picture of the price per wafer doubling. AMD would certainly like everything to think that their costs have gone up, because it makes us more willing to pay higher prices, but we can figure out just how much more a wafer would have to cost before it wipes out AMD's ability to gain additional profit without adjusting their own prices.

We know that Zen+ has 213 mm^2 die. I can't find the exact dimensions, but with a die that size, you can expect around 250 chips per wafer assuming perfect yields. With Zen 2 and the much smaller ~75 mm^2 die, you can get over 750 chiplets per wafer. Granted you need to factor in the cost of the IO die as well since the chiplet isn't replacing the previous Zen+ chip by itself, but those IO dies are all being manufactured on a mature process that has much lower demand. From another perspective, AMD pays for those GF wafers regardless so from that point of view the cost doesn't matter.

We can ignore yields for simplicity, but the smaller the die, the better the yields. So even though it's a new node, which typically means more defects, AMDs chiplets are small enough that they'll still get good yields, and since they moved a lot of the IO off of the chiplet, defects are less likely to mean that a chiplet with a defect can't be salvaged in some way. Basically, chiplets have larger percentage of their area devoted to parts that are redundant, so a defect is less likely to hit anything that makes the chip completely useless.

If you just look at the scaling factor alone, AMD can get upwards of 3 times as many chiplets per wafer, so the wafer costs would need to be more than double. There's other things to consider such as using more than one chiplet on a CPU or any extra costs incurred when using the chiplet approach.

AMD will likely charge more, simply because they have a product that can command higher prices, but I suspect that they'd be able to get by even if that weren't the case.

 

[Mockingbird]

I think prices on this list are on the high side, and both 12 and 16 core in the same series seems unlikely to me. I do think they will try to hit Intel hard where it hurts (their sweet spot is the 6C/6T 8600/9600(K)

My guess would be this:

Ryzen 3 3300 - 6C/12T - $149
Ryzen 3 3400 - 6C/12T - $199
Ryzen 3 3400X - 6C/12T - $229

Ryzen 5 3500 - 8C/16T - $249
Ryzen 5 3600 - 8C/16T - $299
Ryzen 5 3600X - 8C/16T - $329

Ryzen 7 3700 - 12C/24T - $349
Ryzen 7 3800 - 12C/24T - $399
Ryzen 7 3800X - 12C/24T - $429

Ryzen 9 3900 - 16C/32T - $449
Ryzen 9 3900X - 16C/32T - $499

Those prices are too low.

I am talking about MSRP, not street price.

For example, Ryzen 5 2600 has an MSRP of $199, but sells for ~$160 right now. Ryzen 7 2700 has an MSRP of $299, but sells for ~$260.

 

[Topweasel]

Those prices are too low.

I am talking about MSRP, not street price.

For example, Ryzen 5 2600 has an MSRP of $199, but sells for ~$160 right now. Ryzen 7 2700 has an MSRP of $299, but sells for ~$260.

Saying these are too low ignores just about everything AMD has done since they launched Ryzen. Their goal has always been disruptive pricing to fuel sales. Not saying you are wrong but there seems to be a lot of trouble to use Intel's pricing model to apply to Ryzen pricing even as far back as Ryzen's launch.

Take the Projected pricing of 3900x. I would actually say it's a little high. But Intel doesn't have a consumer competitor for it and Comet Lake at 10c won't be a match and the last time AMD had double the core count they found a middle ground at half the price of it closest competitor with decent Margin for themselves. But Looking at Threadripper pricing they properly slid down pricing on 2950x while allowing themselves growth in margin on the 2970 and 2990.

So I would put this in the wait and see. I know it won't be priced as high as the 2950. So maybe max $600. But if AMD still plans on being disruptive I think we will see them bring back the $500 price point.

 

[Mockingbird]

If the above graph is true, 80mm2 at 7nm = ~150mm2 at 14nm (14/16nm is above 2.00 and 7nm is bellow 4.00)

Ryzen 14nm APU die is 210mm2 sold at MSRP of $99 (Ryzen 3 2200G), thus cost should be close to 110mm2 at 7nm

Now the IO chip of the new Ryzen 3 is close to 110mm2 manufactured at 14nm GloFo which will make it extremely chip.

6C 12T Ryzen 3 can easily have MSRP of $99-$120

8C 16T Ryzen 3 at MSRP of $199 is also easily achievable using the same single chiplet SKU.

You forgot about increased shipping and packaging cost.

GlobalFundries has to ship the die from New York and TSMC has to ship the die from Taiwan.

The packing cost increase because of increase complexity. The costs are associated with 1.) using a PCB that can support up to 3 dies and 2.) installing two different dies instead of one

Furthermore, "achievable" doesn't it makes any sense.

Read what I said earlier:

AMD already dominate the ~$200 and under (processor) market.

It doesn't make sense for AMD to drop the prices further to compete with itself.

It's the $300+ (processor) market that Intel dominate with products such as the Core, i7-8700K, Core i7-9700K, Core i9-9900K

Let me make it clear why AdoredTV's list is very very likely fake. (If, for whatever reason, you are still unconvinced otherwise.)

Games, by design, are hard to parallel across many cores.

You would see a huge benefit going from 2C to 4C, but a significant (but smaller) benefit going from 4C to 6C.

You hardly see a benefit going from 6C to 8C. (a few games can benefit, but the majority cannot)

That is hence the reason why the the 6C Ryzen 5 is AMD's best selling product.

Now, according to AdoredTV's list, AMD would cut its 6C processor's MSRP in half, from $199 to $99.

Someone would point out that Ryzen 5 2600 is selling for ~$160, which is below MSRP.

This is true, but the hypothetical Ryzen 3 3300 would likely also go on sale after being released (possibly to $80).

So let's think about that for a moment: AMD's new best selling processor is selling for $80. (down from $160)

That would significantly hurt AMD's bottom line.
_______________________________________________________________________________________

More likely than not, AMD is going to keep the number of cores in the Ryzen 3, Ryzen 5, and Ryzen 7 the same.

Ryzen 9 would target new price points.

12C/24T Ryzen 9 3800X at $500 to $600 would target Core i9-9900K

16C/32T Ryzen 9 3900X at $700 to $800 would target Intel's unreleased 10-core Coffee Lake Refresh

There's some overlap with Threadripper, but the people who buy Threadripper probably need the extra PCIe lanes that the Ryzen 9 doesn't have.

 

[amd6502]

Those prices are too low.

That's an understatement. imho they simply have nothing to do with reality and are completely ignoring fundamentals. Namely:

1. The transistors per core has gone up very considerably in Zen2.

2. Transistors are considerably more expensive on 7nm compared to 12nm.

There is nothing in the near future that will beat 12nm for value (except FDSOI which may be irrelevant for AMD).

From both 1. and 2. of the fundamentals it also follows that R&D investement is also going up. It would be foolish and near suicidal to want to cut your margins here. You want the best and most diverse product line (which they have) but you absolutely don't want to give away your best performers, which are the high IPC high frequency zen2 products (anything that uses anything but total reject 7nm silicon or their 12nm and other value nodes).

In addition, MCM packaging adds a smaller amount of cost and MCM does not motivate maximizing volume by having margins come down to razor thin levels. This 7nm MCM consumer generation is likelier to be shorter lived and lower volume compared to 12nm and even compared to the follow up 4000 generation when we likely start seeing monolithic 7nm consumer products on 7nm+. That's another reason why I think they would better off actually increasing the margins. (My guesstimated numbers are with keeping margins same to the 2000 series.)

 

[Abwx]

Yet Athlon 220 and 240 seem to be telling the oposite, AMD was too afraid that the Athlons would take too much of the 2200G/2400G sales to the point they made the 220 and 240GE worthless. They were unwilling to increase the IGP freq by 1mhz not to impact 2200G sales, so i find hard to belive they will sudenly change and target Intel no matter the cost.

In the EU the 200/220/240GE are at 50/63/67€ respectively, at those prices the 240GE make sense comparatively to the 200GE...

https://geizhals.de/?cat=cpuamdam4&xf=12099_Desktop~820_AM4

That being said those prices are very low for what it is, dunno why you are constantly badmouthing AMD while still making your bread and butter with their products, you want those APUs for free, with all features, CUs and anything else unlocked..?.

At some point one should be reasonable, personaly i find that with bottomised pricings they are litteraly subsiding their customers, and you among them, but if ever their offering seems to you not "generous" enough you can always buy the products of the competitor...

 

[Mockingbird]

Third, I expect the 16C part to be thermally limited (lower all-core speeds) so the MT performance increase wont be that much more compared to 12C. Say 20% instead of 33%.

I expect new motherboards that has beefier VRMs and can handle higher TDP.

As an example, assuming that a 14nm wafer costs half as much as a 7nm wafer, but you get 3x as many dies per wafer at 7nm (after taking defects into account):
An R3 would need a 14nm IO plus a 7nm chiplet. Quick math on that puts the manufacturing cost below that for 210mm at 14nm, so it is easy to see why the low end can easily go for $99 if so desired.

process sizes are now just making names that has no basis in reality

TSMC's 7nm is only 2x as dense as GloFo's 14nm, not 4x.

If AMD are going to be releasing products similar to those leaks, at prices similar to those leaks then I think it is a strategy to get far more OEM take up than they have had so far. Once the OEMs have invested in their infrastructure to support more of your products it makes it easier going forward and then AMD can start raising their margins. I think the idea, assuming the leaks are in the ballpark, is to use the opportunity Intel have given them to forge long lasting OEM relationships to increase the number of OEM offerings. The benefit is that when Intel get their house in order AMD will have greater OEM support than if they go for pure margins right now.

The "leaks" are fake.

AMD don't name or priced its products until close to release (usually 1-2 weeks prior).

Longer term, it is better to have lower margins on higher volume, than higher margins on low volume.

It's easy to push up margins when you dominate the market. AMD are not in that position yet, so they need to manouver themselves as best they can. Once the market share is there, then push margins up a little.
Longer term, it is better to have lower margins on higher volume, than higher margins on low volume.
Whilst I think that cannibalising future sales is a valid argument, I suspect that it impacts Intel far greater than it does AMD; in the first instance they are cannibalising Intel current and future sales, not their own.

AMD already dominate the ~$200 and under (processor) market.

It doesn't make sense for AMD to drop the prices further to compete with itself.

It's the $300+ (processor) market that Intel dominate with products such as the Core, i7-8700K, Core i7-9700K, Core i9-9900K

That's the market that AMD would go after.

First, their cost won't be double that since it uses the same packaging and IO-die.
Second, they need to entice enthusiast to go for the $429 or $499 part.
Third, I expect the 16C part to be thermally limited (lower all-core speeds) so the MT performance increase wont be that much more compared to 12C. Say 20% instead of 33%.

I expect only on 7nm EUV (does TSMC call that 5nm?) with 12nm IO die will the full potential of 16C be unlocked.

12C/24T Ryzen 9 3800 for $499

That's 50% more cores than the Core i9-9900K.

That's enticing enough.

Saying these are too low ignores just about everything AMD has done since they launched Ryzen. Their goal has always been disruptive pricing to fuel sales. Not saying you are wrong but there seems to be a lot of trouble to use Intel's pricing model to apply to Ryzen pricing even as far back as Ryzen's launch.

Take the Projected pricing of 3900x. I would actually say it's a little high. But Intel doesn't have a consumer competitor for it and Comet Lake at 10c won't be a match and the last time AMD had double the core count they found a middle ground at half the price of it closest competitor with decent Margin for themselves. But Looking at Threadripper pricing they properly slid down pricing on 2950x while allowing themselves growth in margin on the 2970 and 2990.

So I would put this in the wait and see. I know it won't be priced as high as the 2950. So maybe max $600. But if AMD still plans on being disruptive I think we will see them bring back the $500 price point.

No, AMD goal is to disrupt Intel's pricing.

At the time, AMD doesn't have any competitive products so it doesn't have to worry about reduced profit margin.

Now, AMD dominate the ~$200 and under, it doesn't make sense for AMD to lower the prices and compete with itself.

It's the $300+ (processor) market that Intel continues to dominate with products such as the Core, i7-8700K, Core i7-9700K, Core i9-9900K

AMD can continue to disrupt Intel's pricing by, for example, releasing 12C/24T Ryzen 9 3800 at the same price as the Core i9-9900K's.

 

[somethingclever]

TSMC's 7nm is only 2x as dense as GloFo's 14nm, not 4x.

A halving of length and width is a quarter reduction in area, so you will see 4x more chips in a square wafer. Roughly 3x increase in usable chips is a good estimate.

Now, AMD dominate the ~$200 and under, it doesn't make sense for AMD to lower the prices and compete with itself.

Does AMD really dominate that though? From msrps alone, Intel hex cores are awfully close to AMD hex core list prices. I think AMD staying a step ahead of Intel is a good reason to drop prices again.

 

[Topweasel]

I expect new motherboards that has beefier VRMs and can handle higher TDP.



process sizes are now just making names that has no basis in reality

TSMC's 7nm is only 2x as dense as GloFo's 14nm, not 4x.



The "leaks" are fake.

AMD don't name or priced its products until close to release (usually 1-2 weeks prior).





AMD already dominate the ~$200 and under (processor) market.

It doesn't make sense for AMD to drop the prices further to compete with itself.

It's the $300+ (processor) market that Intel dominate with products such as the Core, i7-8700K, Core i7-9700K, Core i9-9900K

That's the market that AMD would go after.



12C/24T Ryzen 9 3800 for $499

That's 50% more cores than the Core i9-9900K.

That's enticing enough.



No, AMD goal is to disrupt Intel's pricing.

At the time, AMD doesn't have any competitive products so it doesn't have to worry about reduced profit margin.

Now, AMD dominate the ~$200 and under, it doesn't make sense for AMD to lower the prices and compete with itself.

It's the $300+ (processor) market that Intel continues to dominate with products such as the Core, i7-8700K, Core i7-9700K, Core i9-9900K

AMD can continue to disrupt Intel's pricing by, for example, releasing 12C/24T Ryzen 9 3800 at the same price as the Core i9-9900K's.

No they said disruptive pricing. It applied to Intel but they also need people to upgrade. They need people who wouldn't have gotten a Threadripper to see their 3900x knowing it can work on their x370 or B350 board and purchase that.

This can go back and forth. But I find it unlikely that AMD intends to sell a AM4 CPU at ThreadRipper pricing. Why would AMD care if they are competing against TR2 or R2 pricing? They will do what they always do for a CPU launch. Heavily discount previous gen to minimize stock and keep it there till they are all sold.

In the end AMD's job is to maximize profits where supply and demand meet. But this isn't Vega VII where offering it is basically a way to make money on chips they weren't going to sell otherwise. There job isn't going to be "well we have Threadripper 16c at $849, so we can't price our standard desktop processor below that". TR2 isn't a true volume product. It's a median between their volume desktop CPU and volume server CPU (which is a much smaller volume) for a certain type of people. It's a HEDT product not a consumer product. Their job is to make sure that enough people pick the 3900x instead of a 9900k and 10900k. TR will have it's own refresh and they can play with its HEDT pricing on that later. You don't accomplish that by pricing it well into the HEDT stratosphere. A $700+ AM4 CPU goes against everything that AMD wants to accomplish with that platform.

 

[Spartak]

12C/24T Ryzen 9 3800 for $499

That's 50% more cores than the Core i9-9900K.

That's enticing enough.

You expect Intel to stand still this year, and maintain their current price points. There's exactly zero percent chance on that.

 

[Mockingbird]

Does AMD really dominate that though? From msrps alone, Intel hex cores are awfully close to AMD hex core list prices. I think AMD staying a step ahead of Intel is a good reason to drop prices again.

Ryzen 5 2600 is the best selling processor worldwide

The only Intel's processsor priced close to it is the Core i5-8400 and that only has half as many threads and only slightly better single thread performance.

No they said disruptive pricing. It applied to Intel but they also need people to upgrade. They need people who wouldn't have gotten a Threadripper to see their 3900x knowing it can work on their x370 or B350 board and purchase that.

This can go back and forth. But I find it unlikely that AMD intends to sell a AM4 CPU at ThreadRipper pricing. Why would AMD care if they are competing against TR2 or R2 pricing? They will do what they always do for a CPU launch. Heavily discount previous gen to minimize stock and keep it there till they are all sold.

In the end AMD's job is to maximize profits where supply and demand meet. But this isn't Vega VII where offering it is basically a way to make money on chips they weren't going to sell otherwise. There job isn't going to be "well we have Threadripper 16c at $849, so we can't price our standard desktop processor below that". TR2 isn't a true volume product. It's a median between their volume desktop CPU and volume server CPU (which is a much smaller volume) for a certain type of people. It's a HEDT product not a consumer product. Their job is to make sure that enough people pick the 3900x instead of a 9900k and 10900k. TR will have it's own refresh and they can play with its HEDT pricing on that later. You don't accomplish that by pricing it well into the HEDT stratosphere. A $700+ AM4 CPU goes against everything that AMD wants to accomplish with that platform.

Not many people upgrade every year.

The majority of people buying 3rd generations are likely to be on much older processors (i.e. Sandy Bridge, Haswell, etc.)

Also, have you actually seen the MSRP for Ryzen Threadripper 2920X and Ryzen Threadripper 2950X? It's $649 for the former and $899 for the later.

Nowhere did I said that Ryzen 9 2800 and Ryzen 9 2900 should sell for $649 and $899.

Ryzen Threadripper platform is for those who need more PCIe lanes and quad-channel memory.

You expect Intel to stand still this year, and maintain their current price points. There's exactly zero percent chance on that.

There is almost zero chance of desktop Ice Lake launching this year and, even if by some miracle Intel is able to get it out this year, there is no sure indicator that Intel would lower the price point.

 

[PotatoWithEarsOnSide]

The $200 and under market is dominated by OEMs, of which AMD have next to no share at all.
The aim is to be disruptive through the entire product stack, not just enthusiast upwards.

 

[Lovec1990]

AMD is already disruptive just not enough yet

 

[Mockingbird]

The $200 and under market is dominated by OEMs, of which AMD have next to no share at all.
The aim is to be disruptive through the entire product stack, not just enthusiast upwards.

I am talking about retail sales only (box processors).

OEM pricing is an entirely a different thing.

AMD is already disruptive just not enough yet

You don't need to disrupt the market if you already dominate the market.

AMD already dominate the ~$200 and under market.

It's the $300+ market that Intel dominates and AMD needs to disrupt.

e.g.

$329 8C/16T Ryzen 7 3700X with similar performance to $500 Core i9-9900K

$499 12C/24T Ryzen 9 2800 - similar price to Core i9-9900K, but with 50% more cores/threads

 

[jpiniero]

A $700+ AM4 CPU goes against everything that AMD wants to accomplish with that platform.

That's part of the reason why I don't think 16 cores on AM4 is happening. Save that for Threadripper.

 

[Kocicak]

To all who think that the leaked prices are unrealistic:

The 8 core chiplet is a universal part which will be widely used, produced in high quantities and it is small at the same time: it will be cheap.

The 6 core chiplet is a broken part, when counting 8c and 6c chiplets, yield will be approaching 100%.

Margins on the chiplet will vary depending on where the part will be used, high in server processors and low in cheap consumer processors.

AMD can sell those broken chiplets with zero profit, they can even give them away for free. They do not care. It is a waste product. As a waste product, I expect the cheap processors with 6 core chiplets to be available only after the processors with healthy chiplets. Expensive 12c processors may even contain healthy chiplet/s and they may be available immediatelly.

Given the fact that after long years of struggling and having low market share AMD have now chance to get their share back, they really need to sell as many processors as possible - for that they must be cheap.

Leaked pricing is perfectly rational.

 

[jpiniero]

It's the $300+ market that Intel dominates and AMD needs to disrupt.

For that to happen it would have to be comparable and/or faster in games. Moar corez won't get AMD there, faster clocks and/or IPC will.