• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.
  • The forums have been upgraded with support for dark mode. By default it will follow the setting on your system/browser. You may override it by scrolling to the end of the page and clicking the gears icon.

AMD Trims Q3 Forecast, $1 Billion Missing, Client Processor Revenue down 40%, Halved Quarter-over-Quarter

Su was asked ~1Q22, doing the worst scalping/mining frenzy in our history - "how is AMD" - her response: "*outstanding* year for AMD with record annual revenue and profitability." Yeah, she had stated that doing a time for many consumers were being gouged by scalpers, hunted by miners, and/or suffering from COVID lockdowns. But, Su thought it was still all "outstanding" even when she knew that many consumers were struggling and hurting.


Well, there is a happy ending after all, and Su is now concerned.

Reap what you sow!
AGAIN as stated before Dr Lisa Su is NOT a friend to the consumer base. She has betrayed the trust of customer base for the shekel and IMHO is just 1 step lower than the Sleeze ball Jensen. I seen her business moves for the past 2 years and they have more and more using the US as their cash cow, because people still believe that they are the Under Dog.

As stated before AMD is price gouging their newest generation of product for the amount of performance you are getting over the previous generation. People do not have the money for this crap.
 
Yeah, it's obvious really. If you have 90W of energy to spend, how do you get more performance? By spending those 90W more efficiently. It's literally the what the word means, after all. Of course any specific iteration of a design can be tuned towards being more or less efficient by changing other variables, but that's another question entirely - and where people get sidetracked, as first Intel went to 125W240W TDPs for K-SKU CPUs, and now AMD is following right in their footsteps with 170W230W high X-SKU CPUs - which are still architecturally more efficient than previous generations (for both brands), but are tuned to have higher absolute power consumptions, making their absolute efficiency (perf/W) into a more complicated comparison.

Still, IMO, we need CPU makers to step down to more sensible power levels at stock. 95-125W is fine for the high end, with an easily selectable cTDP-up mode for those wanting more. The performance differences are so small it really doesn't matter, and the power savings are meaningful at scale, while anyone really wanting those extra 5% of performance could get them easily still.

You and I value different things, but efficiency is still the route to get there. I don't really care about power consumption within reason (600W 4090s are outside of reason IMO).

But all of these systems are tunable within their own power/performance curves to what a knowledgeable end user desires, and their performance limits are pretty close albeit with different strengths/weaknesses is my bet. If you want low power consumption, you can tune to that. If you want absolute performance, you can tune to that too.
 
LG didn't exit the phone business because of market cap, they weren't making any money there whilst many of their other businesses were doing just fine. But I'm (almost) sure you already knew that?

Your point may have been somewhat valid if they went belly up & didn't honor say warranty on them, but they're still delivering software updates what after 2 years? I know because I have a G8x thinQ & it's doing just fine!

Yeah they weren't going to go bankrupt but their phone business was, which is what I was going to buy. I wouldn't buy one of their TVs is that was getting slaughtered either.
My S8 still got an update like a month ago. That's like 5.5yrs.
 
make some, lose some...

surge & purge....

What goes around comes around (and usually bites you in the arse on the return trip !)....

Forget but never forgive......

this is how the market works, Capitalism 101 at it's finest :D

HOWEVER, all the companies that became soooo accustomed overpricing everything during the pandemic at outrageous scapler's prices are now feeling the pain of those decisions....

Pain, when used properly, a wonderful motivator, it is !
To complete your list,

It's a win-win situation! (we win twice, you lose)

Tomorrow won't be a pretty day for their stock but I expect a partial rebound next week.
I expect the same, that's a common pattern, so let's wait and see. A seasoned financial analyst who has paid attention on many such occasions before (like financial announcements and reports) would also be able to tell when and by how much, and make lots of money from it.
 
You and I value different things, but efficiency is still the route to get there. I don't really care about power consumption within reason (600W 4090s are outside of reason IMO).

But all of these systems are tunable within their own power/performance curves to what a knowledgeable end user desires, and their performance limits are pretty close albeit with different strengths/weaknesses is my bet. If you want low power consumption, you can tune to that. If you want absolute performance, you can tune to that too.
That's true - but all the more of an argument for the standard configuration to strike more of a balance, especially as performance-hungry enthusiasts are far more likely to know how to tune things than anyone else. And across hundreds of thousands if not millions of CPUs sold to everyone else every year, the difference in overall energy usage would be meaningful, while the performance difference wouldn't be noticeable for most of these users.
 
That's true - but all the more of an argument for the standard configuration to strike more of a balance, especially as performance-hungry enthusiasts are far more likely to know how to tune things than anyone else. And across hundreds of thousands if not millions of CPUs sold to everyone else every year, the difference in overall energy usage would be meaningful, while the performance difference wouldn't be noticeable for most of these users.

Well they need some kind of standard to measure against, I agree, or at least an agreed on method of measuring performance of a CPU with a specific set of platform / power / frequency points.

I'm not sure what people expect to happen when they stick these chips in a motherboard called 'Apex' or 'Godlike'. Those boards will run the CPUs as hard as they can.

Not all boards do that (90W difference between the NZXT N5 and the Godlike or Unify with 12900K at 4.9Ghz):

1665185628147.png
 
This shows each manufacturer's share of the passmark sample as of October 1. Mobile CPUs are currently the mainstay of the client sector, and open data was already showing signs of depression.

I posted this elsewhere last week and the kind folks politely explained how wrong I was, using 2021 data.



スクリーンショット 2022-10-08 061740.png
 
That's true - but all the more of an argument for the standard configuration to strike more of a balance, especially as performance-hungry enthusiasts are far more likely to know how to tune things than anyone else. And across hundreds of thousands if not millions of CPUs sold to everyone else every year, the difference in overall energy usage would be meaningful, while the performance difference wouldn't be noticeable for most of these users.

About time everyone realizes pushing chips to edge of oblivion at the factory is stupid and doubling it down outside is even more doubly so.
 
This shows each manufacturer's share of the passmark sample as of October 1. Mobile CPUs are currently the mainstay of the client sector, and open data was already showing signs of depression.

I posted this elsewhere last week and the kind folks politely explained how wrong I was, using 2021 data.



View attachment 264561
I don't think this is an accurate reflection of market share. First, Alder Lake was launched last year and has been around for 10 months now. The new AMD gen just launched. So what are the AMD 20% CPU samples made up of? Second, perhaps AMD users that run whatever Passmark benchmark this is based on, already did that in 2021, and the majority of the benchmarks run were done by Alder Lake users in 2022?
 
I don't think this is an accurate reflection of market share. First, Alder Lake was launched last year and has been around for 10 months now. The new AMD gen just launched. So what are the AMD 20% CPU samples made up of? Second, perhaps AMD users that run whatever Passmark benchmark this is based on, already did that in 2021, and the majority of the benchmarks run were done by Alder Lake users in 2022?

My tally is based on the launch date of each SKUs. All the Intel's and AMD's CPUs in "2022 gen." category are announced at the same date, CES 2022 in January:
Intel: Alder lake-H/HX/P/U **does not include desktop model
AMD: Rembrandt (Ryzen 6000 series) and Barcelo (Ryzen 5xx5U) **Barcelo's internal code is same to Cezanne (Ryzen 5xx0U/H), both shares "CZN"
 
My tally is based on the launch date of each SKUs. All the Intel's and AMD's CPUs in "2022 gen." category are announced at the same date, CES 2022 in January:
Intel: Alder lake-H/HX/P/U **does not include desktop model
AMD: Rembrandt (Ryzen 6000 series) and Barcelo (Ryzen 5xx5U) **Barcelo's internal code is same to Cezanne (Ryzen 5xx0U/H), both shares "CZN"
Of course Intel will outpace AMD if that is your tally. Intel new platform has already launched, whereas AMD was still on the old 5xxx platform. If you are buying a new laptop, would you buy one with a processor of the latest platform, or a refresh of one that was launched 2 years ago?
 
My tally is based on the launch date of each SKUs. All the Intel's and AMD's CPUs in "2022 gen." category are announced at the same date, CES 2022 in January:
Intel: Alder lake-H/HX/P/U **does not include desktop model
AMD: Rembrandt (Ryzen 6000 series) and Barcelo (Ryzen 5xx5U) **Barcelo's internal code is same to Cezanne (Ryzen 5xx0U/H), both shares "CZN"

So, your chart isn't wrong exactly, but not sure what it's supposed to prove that can't be shown from market share stats.

Intel has been gaining desktop market share since Q3 2020. This is one of the things I suspect a lot people don't know, but the information is out there.

Intel's Mobile share started losing share in Q2 2021, after AMD dipped a bit in 2020. That dip was probably due to insufficient supply, Intel was in its element churning out mass quantities of CPUs in 2020. Intel's server segment has been losing market share since Q3 2018.
 
So, your chart isn't wrong exactly, but not sure what it's supposed to prove that can't be shown from market share stats.

Intel has been gaining desktop market share since Q3 2020. This is one of the things I suspect a lot people don't know, but the information is out there.

Intel's Mobile share started losing share in Q2 2021, after AMD dipped a bit in 2020. That dip was probably due to insufficient supply, Intel was in its element churning out mass quantities of CPUs in 2020. Intel's server segment has been losing market share since Q3 2018.

Mercury Research reports market share through Q2, with AMD's share of the mobile market reaching an all-time high. And the change in AMD's client segment from Q2 to Q3 was -53%.

I would speculate that AMD's market share growth in Q2 may have been due to cheap inventory clearance of older generation CPUs. A search for "New Arrival" on Amazon shows that Intel has 2/3 of the latest generation, while AMD has 2/3 of the older generation.

Amazon (US) - Laptops - Intel - Any New Arrival
116: Alder lake
59: Tiger lake
10: Atom (Gemini / Jasper lake)
6: Old generations (before 10th)

Amazon (US) - Laptops - AMD - Any New Arrival
7: Rembrandt (Zen3+)
1: Barcelo (Zen3)
8: Cezanne (Zen3)
4: Lucienne (Zen2)
5: Old generations (before Ryzen 4000)


Lisa Su said "Significant inventory correction across the PC supply chain," to explain the significant drop in sales this quarter.
 
AMD will probably continue to gain market share from Intel. Especially in the server space.

The issue is the whole market is shrinking significantly. There is excess inventory, ultimately the excess demand from Covid has "gone".

Additionally, high inflation is restricting consumers budgets and pretty much no one with a modern CPU or GPU from last few years has much incentive to upgrade.

I'd expect chip sales to go below 2019 levels in the next year or two.
 
They're probably around or slightly below that level. This is what 3-4 quarter in a row where market is shrinking?
 
So, your chart isn't wrong exactly, but not sure what it's supposed to prove that can't be shown from market share stats.

Intel has been gaining desktop market share since Q3 2020. This is one of the things I suspect a lot people don't know, but the information is out there.

Intel's Mobile share started losing share in Q2 2021, after AMD dipped a bit in 2020. That dip was probably due to insufficient supply, Intel was in its element churning out mass quantities of CPUs in 2020. Intel's server segment has been losing market share since Q3 2018.
Tom's has data up to Q2 2022 now, too.
 
The PC component industry is out of touch with its user base. That's the problem. Specifically though, I see it as a GPU manufacturing problem. PSU, RAM, Storage options; these items, arguably, have far longer lifespans. But with GPU's (I feel) we're being pushed to upgrade, and companies are inventing/creating new gimmicks to justify their new, expensive products. Yeah, I mean Nvidia.

They'll need to readjust expectations. And we as consumers ought to vote with our wallets. Hypocrisy from a 2080ti owner, I know, but that was my one extravagant splurge.

Out of touch for sure, like who in the hell wants to upgrade from a AM4 to a AM5 platform when the motherboards are like $250+ DDR5 requirement. In fact by the looks of it sales are really good for AM4 still today.

The cost is crazy to upgrade even more so when i could just pick up a 5800X3D and call it a day.

They never expected this and at the very least should of had DDR4 compatibility and actually made the IHS right in the 1st place, as i know i don't want to deal with the bad cooling with only one real option to keep it cool is to void the warranty.

I gett he feeling people are settling in on whats good enough. Even if i was to replace my other PC i still build another AM4 platform, AM5 is just not worth it.
 
Out of touch for sure, like who in the hell wants to upgrade from a AM4 to a AM5 platform when the motherboards are like $250+ DDR5 requirement. In fact by the looks of it sales are really good for AM4 still today.

The cost is crazy to upgrade even more so when i could just pick up a 5800X3D and call it a day.

They never expected this and at the very least should of had DDR4 compatibility and actually made the IHS right in the 1st place, as i know i don't want to deal with the bad cooling with only one real option to keep it cool is to void the warranty.

I gett he feeling people are settling in on whats good enough. Even if i was to replace my other PC i still build another AM4 platform, AM5 is just not worth it.

To add, If you look into the details of the AM5 platform / chipset / motherboard features, really most of the major feature sets that sets AM5 apart from AM4 are only available to the very high end motherboards so far.

As example, the lower end X670 motherboards don't have a PCI 5 x16 slot, they're using PCIe 4 x16 apparently to save on cost. Most of them will simply have a PCIe 4 x16 and one PCIe 5 x4 to CPU.

I'm also seeing that none of the B650 boards will have PCIe 5 x16 based on AMDs high level diagrams of B650 connectivity. It appears this is at AMDs behest since there's no technical reason I can think of why it couldn't exist as that connectivity comes from the CPU, not the chipset.

To make matters a bit worse, the bandwidth to the chipset on all of the AMD Zen 4 chipsets is half of what it is to a Z690 or Z790. You could literally saturate that chipset<->CPU bus with a single PCIe 4 x4 m.2 SSD. So all of those connections its capable of driving are nice, as long as you only use a couple of those devices at once. Meanwhile most of the PCIe 5 lanes are not used.

Basically unless you get a $450+ motherboard, a bunch of what you pay for with Zen 4 \ AM5 is going to waste.

So with that backdrop, it makes total sense that the two best selling AM5 motherboards at Microcenter are a $699 X670E Crosshair Hero and the $999 Crosshair Extreme. I pulled this up for multiple locations across the country and these are always the two most popular :

1665249473328.png
 
Out of touch for sure, like who in the hell wants to upgrade from a AM4 to a AM5 platform when the motherboards are like $250+ DDR5 requirement. In fact by the looks of it sales are really good for AM4 still today.

Theoretically it's possible that AMD motherboard vendors are selling lots of AM4 motherboards, but you still need to drop a processor into them. What are these people doing? Dropping in previously owned CPUs?

AMD's Client business missed by $1B in revenue. That means they didn't sell a bunch of Zen 3 (Ryzen 5000 series) parts: both desktop and mobile.
 
It should be noted that revenue for this year is down for nearly everyone in this industry.. AMD is hardly alone.
True but some financial analysts are claiming that AMD is likely to have made a loss for Q3 2022. There is a big difference between loosing revenue and remaining profitable, and less revenue leading to losses.

It might be worth noting that Intel made a loss for Q2 2022, and they will be reporting on Q3 2022 on 27 October. AMD detailed numbers for the quarter come out on 1 November. Watch this space, I guess.
 
I'm also seeing that none of the B650 boards will have PCIe 5 x16 based on AMDs high level diagrams of B650 connectivity. It appears this is at AMDs behest since there's no technical reason I can think of why it couldn't exist as that connectivity comes from the CPU, not the chipset.
No, it's a cost issue. PCIe 5.0 requires more PCB layers and higher quality board materials. AMD fully allows for PCIe 5.0 on B650 if board makers want to implement it - it just doesn't make sense in terms of price.
To make matters a bit worse, the bandwidth to the chipset on all of the AMD Zen 4 chipsets is half of what it is to a Z690 or Z790. You could literally saturate that chipset<->CPU bus with a single PCIe 4 x4 m.2 SSD. So all of those connections its capable of driving are nice, as long as you only use a couple of those devices at once. Meanwhile most of the PCIe 5 lanes are not used.
That's true - but hardly relevant in real world use cases. No, you won't have a good time running concurrent sequential benchmarks on several chipset-connected SSDs, but for anything else, that's plenty of bandwidth even if you have, say, a 4k120 capture card + the SSD that capture is being written to both connected through the chipset. Very few real world use cases actually utilize anywhere near a full PCIe 4.0x4 link for any significant amount of time.

Of course, there's also a major argument to be made for PCIe 5.0 being pointless in and of itself, for consumer use cases. Current GPUs are barely held back by PCIe 3.0x16, and literally nothing is limited by a PCIe 4.0x16 link. No GPU launching in the useful lifetime of these boards will be meaningfully bottlenecked by PCIe 4.0x16 compared to 5.0x16.
 
No, it's a cost issue. PCIe 5.0 requires more PCB layers and higher quality board materials. AMD fully allows for PCIe 5.0 on B650 if board makers want to implement it - it just doesn't make sense in terms of price.

Do they? Have you looked at AMDs official B650 block diagram? If they allow it and it's up to the manufacturer, why would that diagram show it as PCIe 4 x16?

That's true - but hardly relevant in real world use cases. No, you won't have a good time running concurrent sequential benchmarks on several chipset-connected SSDs, but for anything else, that's plenty of bandwidth even if you have, say, a 4k120 capture card + the SSD that capture is being written to both connected through the chipset. Very few real world use cases actually utilize anywhere near a full PCIe 4.0x4 link for any significant amount of time.

So PCIe 5 isn't really useful is the argument, but we need to pay for it anyway right.

In any case, I don't really agree with you. There's a super common daily use for high chipset bandwidth. It's called doing a backup.
 
There's almost zero chance for that, they'd have to lose in the console segment/GPU's & Xilinx for a (total) net loss!

Based on their preliminary revenue numbers (a table in post #1), they are up 29% YoY in total revenue as three out of their four major business units reported gains. It's really the Client business (CPUs) where their forecast missed badly.

Gross margin took a dip (300-400 basis points at first glance) but it seems unlikely they would have a net loss. A deceleration in profit growth is inevitable though.
 
Back
Top