Early AMD Ryzen 9000 Zen 5 listings suggest lower prices than Zen 4

MarcusNumb said:

Wow, if those prices are true, this will be a huge kick in the nutts to Intel. Imagine at least 10% gain in IPC and 10% cheaper!!!!

Click to expand...

The exact opposite is true: this smacks of desperation. Given the inflation since the Zen4 launch and today, AMD is selling its latest products at a 20-25% discount, meaning it doesn't believe the performance gain warrants even parity pricing. AMD already leads in the DIY market, so pricing isn't likely to gain it market share there, and it'll take more than a 10% IPC boost to give it a better foothold among VARs.

(and please: before the fanboys attack, I've been 100% AMD in my home machines for several years now).

Yhe Endymio, it the true. and it's still unknown if this is the same cores on the laptop models which reduced lower-bit SIMD performance for this 10%.....

Endymio said:

The exact opposite is true: this smacks of desperation. Given the inflation since the Zen4 launch and today, AMD is selling its latest products at a 20-25% discount, meaning it doesn't believe the performance gain warrants even parity pricing. AMD already leads in the DIY market, so pricing isn't likely to gain it market share there, and it'll take more than a 10% IPC boost to give it a better foothold among VARs.

(and please: before the fanboys attack, I've been 100% AMD in my home machines for several years now).

Click to expand...

Parity pricing against what? Zen4 launch prices?

Zen4: new motherboard chipsets, new CPU socket, expensive manufacturing tech for chiplets, semi expensive manufacturing tech for IO die, top notch IO capabilities...

Zen5: no really new chipsets, old CPU socket, chiplets use (probably) cheaper manufacturing tech, IO die uses same (=cheaper because older manufacturing tech), IO capabilities stay same, no need to design new IO die...

Zen5 is much cheaper to manufacture than Zen4. To remind, Zen4 was considered too expensive on launch. 1+1=...

HardReset said:

Parity pricing against what? Zen4 launch prices?

Click to expand...

Yes.

HardReset said:

Zen4: new motherboard chipsets, new CPU socket, expensive manufacturing tech for chiplets, semi expensive manufacturing tech for IO die, top notch IO capabilities...

Click to expand...

Eh? You seem to be agreeing that Zen5 brings far less value to the table than its predecessor. And none of this changes the fact that when you reduce pricing 25% on equal volume, total revenue drops 25%. This isn't a positive for AMD, no matter how you slice it.

HardReset said:

Zen5 is much cheaper to manufacture than Zen4.

Click to expand...

Err, what? Zen5 moved from TSMC's N4P node to N4X (N3 for the "c" variant), and from N5 to N4 for CCDs. Only the I/O chiplet is unchanged at 6nm. Both these new nodes are more expensive than the preceeding one (though N4P to N4X is admittedly a minor bump.)

Zen 5 uses the same socket, so not sure why you even bring that up. And while AMD likely did save on "design costs" for Zen 5, you realize that doesn't affect manufacturing costs, right?

Endymio said:

Zen 5 uses the same socket, so not sure why you even bring that up. And while AMD likely did save on "design costs" for Zen 5, you realize that doesn't affect manufacturing costs, right?

Click to expand...

Much of the R&D costs of the new socket design needed to be absorbed in the first generation. While not huge, it's not zero. The real reason I think these chips will end up cheaper is that AMD is trying to push AI on their chips, the best way to get people to buy them is *drum roll* a price cut. I have a feeling that AMD has plans to monetize their local AI features.

I do think that we are reaching the tipping point of whether or not AI is going to sink or swim.

yRaz said:

Much of the R&D costs of the new socket design needed to be absorbed in the first generation.

Click to expand...

Sure. But saying "our R&D costs were low" is just another way of saying "we really didn't improve much". AMD's margins might improve slightly, but their total revenues are going to take a significant dip.

Endymio said:

Eh? You seem to be agreeing that Zen5 brings far less value to the table than its predecessor. And none of this changes the fact that when you reduce pricing 25% on equal volume, total revenue drops 25%. This isn't a positive for AMD, no matter how you slice it.

Click to expand...

If something is cheaper, then value is smaller? No wonder GPU prices are going up all the time. And if volume is greater, then what happens on revenue? It gets up of course.

Endymio said:

Err, what? Zen5 moved from TSMC's N4P node to N4X (N3 for the "c" variant), and from N5 to N4 for CCDs. Only the I/O chiplet is unchanged at 6nm. Both these new nodes are more expensive than the preceeding one (though N4P to N4X is admittedly a minor bump.)

Zen 5 uses the same socket, so not sure why you even bring that up. And while AMD likely did save on "design costs" for Zen 5, you realize that doesn't affect manufacturing costs, right?

Click to expand...

You sure they are actually more expensive? 5nm node was best one available on 2022. Now best available is expensive 3nm. Also 6nm is much older process now so it also should be cheaper.

In other words, 2022 Zen4 had quite new 6nm node and best available 5nm node. Now Zen5 has same 6nm but 2 years older and 4nm is not best available. Total cost should be lower but of course it's impossible to say how much AMD really pays.

Re-using same IO die of course lower manufacturing costs because they probably have some excess chips that were meant for Zen4 CPUs but will now be used on Zen5.

Endymio said:

Sure. But saying "our R&D costs were low" is just another way of saying "we really didn't improve much". AMD's margins might improve slightly, but their total revenues are going to take a significant dip.

Click to expand...

I didn't say they were low. Gonna from PGA to LGA is not cheap and LGA has a lot of downsides that don't get talked about. They went to LGA with Threadripper and it a lot of problems that the solution to was basically "repaste and reseat the CPU"

I'm not saying it was cheap, but they only predicted so many sales and needed to absorb that in the first generation of the socket. The cost of development ended up being about $50 a system. This was split between the a premium on the chipset and a premium on the cpus.

Puiu said:

"European prices are typically listed without the VAT" - ah... what? no they aren't. prices usually include VAT.

From the website's price and VAT policy: All prices include 22% VAT.

This means that the listed prices have a 22% included VAT, although I think the prices are too good to be true. We can see that the listings have a "20% discount if you buy online" already.

Click to expand...

TSMC's N4P is cheaper than th N5 at the time of Zen 4's release. The dies are a bit smaller and AMD looks like it wants to compete rather than gouge. Prices have been reported by a few leakers to be cheaper than Zen 4 and AMD won't be losing money unlike Intel does with it's much higher fabrication costs for Raptor Lake and Arrow Lake will be even more epensive. AMD knows this and Intel will be under enormous pressure to not raise prices.

Endymio said:

Yes.


Eh? You seem to be agreeing that Zen5 brings far less value to the table than its predecessor. And none of this changes the fact that when you reduce pricing 25% on equal volume, total revenue drops 25%. This isn't a positive for AMD, no matter how you slice it.


Err, what? Zen5 moved from TSMC's N4P node to N4X (N3 for the "c" variant), and from N5 to N4 for CCDs. Only the I/O chiplet is unchanged at 6nm. Both these new nodes are more expensive than the preceeding one (though N4P to N4X is admittedly a minor bump.)

Zen 5 uses the same socket, so not sure why you even bring that up. And while AMD likely did save on "design costs" for Zen 5, you realize that doesn't affect manufacturing costs, right?

Click to expand...

Let's be serious, Zen 4 launched with a completely new platform. It's normal for Zen 5 to feel "less important", but even so, it's part of the long term support for AM5 which makes it very important.

The real upgrades in this generation will be for laptops and it seems AMD is putting a lot of effort there. It's also what I want them to do since I plan on upgrading my laptop next year.

Puiu said:

The real upgrades in this generation will be for laptops and it seems AMD is putting a lot of effort there. It's also what I want them to do since I plan on upgrading my laptop next year.

Click to expand...

Agreed, I think my only issue is that AMD don’t get enough design wins, Qualcomm got loads of Design Wins for the completely unknown Snapdragon CPU, but AMD seem to struggle to get many design wins for a genuinely better product than what Intel produces.

Burty117 said:

Agreed, I think my only issue is that AMD don’t get enough design wins, Qualcomm got loads of Design Wins for the completely unknown Snapdragon CPU, but AMD seem to struggle to get many design wins for a genuinely better product than what Intel produces.

Click to expand...

Looking at the rumours, it seems OEMs are banking on AMD doing well and it seems they will also have good early product stocks. Intel is launching late and will probably be expensive and OEMs need to cash in on that AI hype.

Puiu said:

Let's be serious, Zen 4 launched with a completely new platform. It's normal for Zen 5 to feel "less important", but even so, it's part of the long term support for AM5 which makes it very important.

The real upgrades in this generation will be for laptops...

Click to expand...

So Zen 5 desktop isn't a 'real upgrade' over Zen 4? And I'm not debating how Zen 5 "feels", but rather disputing the OP's claim that this (rumored) price cut somehow bodes trouble for Intel.

Puiu said:

Looking at the rumours, it seems OEMs are banking on AMD doing well and it seems they will also have good early product stocks

Click to expand...

Except for the fact that we've heard the same rumors for every AMD CPU launch of the last two decades: "AMD's about the take the OEM markets by storm".

Endymio said:

So Zen 5 desktop isn't a 'real upgrade' over Zen 4? And I'm not debating how Zen 5 "feels", but rather disputing the OP's claim that this (rumored) price cut somehow bodes trouble for Intel.


Except for the fact that we've heard the same rumors for every AMD CPU launch of the last two decades: "AMD's about the take the OEM markets by storm".

Click to expand...

Another strawman argument...

Techspot is NOT saying it doesn't bode well for iNTEL the Corporation.... they are saying it doesn't look good for iNTEL fans and Gamers looking to buy iNTEL. Nobody here is discussing Stocks or even Business/Enterprise chips, because AMD serves those markets with things other than AM5 9000 series.

OEM's will eat up AMD's APU/NPU...

Endymio said:

So Zen 5 desktop isn't a 'real upgrade' over Zen 4? And I'm not debating how Zen 5 "feels", but rather disputing the OP's claim that this (rumored) price cut somehow bodes trouble for Intel.


Except for the fact that we've heard the same rumors for every AMD CPU launch of the last two decades: "AMD's about the take the OEM markets by storm".

Click to expand...

So if a new socket is not released it's not a "real upgrade"? Is that your way of telling us that intel's approach to platform longevity is right? We don't need to "feel" anything. You have the official numbers, we will have third party numbers soon and it seems that this time the x3D chips will release in quick succession after the regular CPUs (at least around the same time as Arrow Lake in oct-dec, possibly sooner).

And yes, the price is really bad for Intel if it turns out to be real. It is a fact that their architecture, process node and on-die packaging are very expensive, very much so compared to AMD's approach. Why is this even something you need to dispute? It's something very obvious.

As for the OEM "rumours" you've mentioned, they don't exist. We've always heard the opposite (things like limited supply have always been a problem in the laptop market). Maybe you are confusing AMD's public statements with something else.

m3tavision said:

Another strawman argument...

Techspot is NOT saying it doesn't bode well for iNTEL

Click to expand...

Maybe read threads before replying to them? The claim wasn't from Techspot, but from post #3, which I quoted clearly in my reply.

Puiu said:

So if a new socket is not released it's not a "real upgrade"?

Click to expand...

You're the one who claiming the "real" upgrade was laptops, not desktops. Do you stand by that remark?

Puiu said:

Is that your way of telling us that intel's approach to platform longevity is right?

Click to expand...

I'm flattered you're asking my opinion when world's greatest philosopher's have never been able to agree on the definition of right vs. wrong. I don't know if Intel is "right", but their approach is unquestionably more popular among consumers. Only the (small) DIY market cares about upgrade paths -- most people simply buy a new machine to upgrade.

Puiu said:

And yes, the price but is really bad for Intel

Click to expand...

I've already explained why the exact opposite is true. This move is unlikely to garner AMD market share and will significantly depress their total revenues.

Puiu said:

As for the OEM "rumours" you've mentioned, they don't exist. We've always heard the opposite

Click to expand...

Sigh, no, we haven't. Every AMD launch for decades has seen forums flooded with fanboys claiming that THIS is the launch where AMD finally has Intel by the short hairs. Every. Single. Launch.

Puiu said:

(things like limited supply have always been a problem in the laptop market).

Click to expand...

They've often been a problem in the desktop market too. AMD has many times before led Intel on price/performance ratios, but missed grabbing that golden ring because of supply-side shortages.

Endymio said:

Maybe read threads before replying to them? The claim wasn't from Techspot, but from post #3, which I quoted clearly in my reply.


You're the one who claiming the "real" upgrade was laptops, not desktops. Do you stand by that remark?


I'm flattered you're asking my opinion when world's greatest philosopher's have never been able to agree on the definition of right vs. wrong. I don't know if Intel is "right", but their approach is unquestionably more popular among consumers. Only the (small) DIY market cares about upgrade paths -- most people simply buy a new machine to upgrade.


I've already explained why the exact opposite is true. This move is unlikely to garner AMD market share and will significantly depress their total revenues.


Sigh, no, we haven't. Every AMD launch for decades has seen forums flooded with fanboys claiming that THIS is the launch where AMD finally has Intel by the short hairs. Every. Single. Launch.


They've often been a problem in the desktop market too. AMD has many times before led Intel on price/performance ratios, but missed grabbing that golden ring because of supply-side shortages.

Click to expand...

Exactly...
So why are you concerned about iNTEL stocks or implying zen5 isnt good, bcz business won't be buying AM5 chips? When that socket is for Personal Computers, not business.

All your points are based on Business and then quoting AM5 chip pricing. Don't you see how jaded and incorrect you are..? You seem overly concern about iNTEL business sector and trying to compare that to AMD's DIY sector.


Zen5 will destroy iNTEL's architecture, OEMs already know this.

Endymio said:

AMD has many times before led Intel on price/performance ratios, but missed grabbing that golden ring because of supply-side shortages.

Click to expand...

Here's one main problem: proof of these so called "AMD capacity problems" have almost always been non-existent. In other words, AMD had no capacity problems, it was just explanation why OEMs didn't use AMD.

In fact, few yars ago when Intel fcked 10nm tech, Intel really had capacity problems, very serious ones. What did OEMs do? Instead using AMD, they just acceptee that waiting Intel CPUs for 6 months is OK.

So this whole "no OEM for AMD because AMD has no capacity" -reasoning is total BS.

HardReset said:

Here's one main problem: proof of these so called "AMD capacity problems" have almost always been non-existent. In other words, AMD had no capacity problems, it was just explanation why OEMs didn't use AMD.

Click to expand...

AMD has had countless capacity problems, but I'm really referring to the period in which their chips were made by Open Foundries or themselves, rather than TSMC: periods you are perhaps too young to remember.

In any case, it's not the point, which is that this launch is no different than any other -- a swarm of zealots line up world wide, orient themselves to directly face AMD's Santa Clara headquarters, bow three times, and solemnly pronounce that, yes, this will be the time AMD seizes control of the OEM channel.
[HEADING=2][/HEADING]

Endymio said:

AMD has had countless capacity problems, but I'm really referring to the period in which their chips were made by Open Foundries or themselves, rather than TSMC: periods you are perhaps too young to remember.

In any case, it's not the point, which is that this launch is no different than any other -- a swarm of zealots line up world wide, orient themselves to directly face AMD's Santa Clara headquarters, bow three times, and solemnly pronounce that, yes, this will be the time AMD seizes control of the OEM channel.
[HEADING=2][/HEADING]

Click to expand...

Global Foundries you mean. For example GlobalFoundries 14LPP, that AMD used for quite many products, is exactly same process that Samsung used. GlobalFoundries 14nm combined with Samsung 14nm capacity does not sound quite small.

And yes, I have read rumors about AMD capacity "problems". Only problem with that BS was always fact that it rarely was seen anywhere. When Intel was supposed to have capacity problems, it clearly meant bad shortage for months.

In other words, explaining why OEMs don't use AMD because capacity has been mostly BS. Other way around, OEMs seem to love Intel when they have really bad capacity problems.

Lel guys take it easy, it only 10% performance, money or no money
almost nothing to see here

HardReset said:

And yes, I have read rumors about AMD capacity "problems". Only problem with that BS was always fact that it rarely was seen anywhere.

Click to expand...

Err, what?

Sep 2003: "Athlon 64 in short supply .... "There are only 100,000 (chips available) worldwide in Q4," said James Chen, the head of Acer's PC line. "The quantity is too small."...."

Aug 2004: "Channel supply shortages in filling orders for AMD Athlon processors may undermine AMD's opportunity to boost its market share..."

Dec 2005: "AMD Acknowledges Desktop CPU Shortages .... The chip maker sent a letter to its partners stating that several of its desktop processor models would continue to be difficult to obtain ... [it] specified the Athlon 64 3500, 3800, the dual-core X2 3800 as well as the lower end Sempron 2600 and 2800 as being in tight supply...."

Oct 2006: "Dell CEO speaks out on AMD shortages .... up to 600,000 Opteron orders backlogged..."

Jun 2010: "AMD Phenom II X6 CPUs in tight supply...AMD has issued a caution on orders placed for these processors..."

Aug 2017: "AMD GPU Shortages to Continue....AMD's APU and GPU will continue to lack inventory in 2H2017 due to Global Foundries' limited capacities...."

Jan 2021: AMD Chip Shortage Caused by Packaging Issues...all TSMC's partners at full capacity..."

Oct 2023: "OEM GPD Accuses AMD of Violating Supply Contract .. GPD cannot get enough Ryzen 7 processors from AMD to fulfill orders...."