AMD Hits Highest-Ever x86 CPU Market Share in Q1 2024 Across Desktop and Server

[ARF]

AMD and Intel are about 50/50 among us enthusiasts but its a very very small piece of the market pie.

Nope. 80% AMD, 20% intel.

https://www.reddit.com/r/Amd/comments/16y3y09

 

[NC37]

I've bought 400 HP Probook/Elite Books over the past 3 years all Intel hardware. The amount of issues I've had with the Intel wireless chipset is mind blogging. This is them connecting to Cisco APs in which Cisco certified the chipset. I would certainly be opening to exploring AMD if they came with a Qualcomm chipset and I'm looking into the Qualcomm Elite chips coming out soon. I budget to replace 20% of my fleet at a 5 year depreciation.

Those are HPs too. HP is the only brand I recommend people against. Have not worked on one in the last 20 yrs that had decent internals. All of them tend to suffer from similar issues too. Not to mention their printers are just as bad. Course, I don't find a lot of printers reliable these days outside of ones with the printheads on the cartridges (which is the older tech cheap ones). The rest, always clogs or have issues that require more work than they are worth.

 

[Daven]

Nope. 80% AMD, 20% intel.

View attachment 347085
https://www.reddit.com/r/Amd/comments/16y3y09

Sure. I agree its probably 80-20 like you said. Wasn’t sure of the current split. Unfortunately this is why such a huge AMD DIY advantage doesn’t make much of a dent in the overall x86 market.

 

[Panther_Seraphin]

Looking at these graphs, it really bothers me that AMD got arrogant and raised its prices on Zen 4. If they want the gold, they need to make sure Zen 5 outperforms anything Intel can overclock, and then price it cheaper. That would explode their market share. AM5 + Zen 4 + DDR5 was so much higher priced than AM4/Zen 3, it slowed their market gains momentum, and only sold reasonably well due to its performance.

AMD needs a strategy rethink. Time to make modest profits to gain market share, then start turning the screws after you have the majority.

Yes. If AMD had smart shareholders, the best strategy is to limit profits to gain as large market share as possible.

Is it in today corporate world even possible to execute any real business strategy?

Have you even taken a look at AMD financials? In the last 4 quarters they have made the same amount of proft as Intel did in 1. Intel then proceeded 6 months later to make 5x order of magnitude more profit in that quarter alone.

AMD doesnt really have the ability to compete directly in a price war where AMD drops, Nvidia/Intel Drops, AMD drops again etc etc etc. They are still recovering from the Bulldozer era in terms of a war chest all the while still trying to invest in their R+D as much as possible.

AMD currently is competing in their only way they can which is pricing their products in relation to their costs with a moderate profit built in and then forcing Intel/Nvidia to compete in pricing.

 

[Tek-Check]

Nope. 80% AMD, 20% intel.

Why data from Week 39 last year when there is fresh data from Week 16 2024?
It's almost 90/10 split, but this is German DIY market, so a small sample.

Intel then proceeded 6 months later to make 5x order of magnitude more profit in that quarter alone.

Don't forget to say than Intel has lost whopping $24 billion dollars in revenues from its peak a few years ago. Context is everything.

I prefer to take AMD's word over some random internet guy who is a representative of no-one knows who...

You can test it by yourself rather than showing embarrassing ignorance of professional reviewers.

 

[davidburke30]

AMD CPUs have been my go-to for years. The Athlon XP 2500 was the one that started it all. I continue purchasing AMD graphics cards, but if they fail to deliver soon, I'll have to stop.

 

[chrcoluk]

Looking at these graphs, it really bothers me that AMD got arrogant and raised its prices on Zen 4. If they want the gold, they need to make sure Zen 5 outperforms anything Intel can overclock, and then price it cheaper. That would explode their market share. AM5 + Zen 4 + DDR5 was so much higher priced than AM4/Zen 3, it slowed their market gains momentum, and only sold reasonably well due to its performance.

AMD needs a strategy rethink. Time to make modest profits to gain market share, then start turning the screws after you have the majority.

I hope it bothers you only because it means you paid more, dont get personally attached to them, they just a company, they wont care about you.

They also will only be able to make finite products, so it may well be this is their most profitable strategy, remember they make tons for consoles.

 

[Xaled]

Despite shortcomings in its products, Intel still has shady tactics to maintain its market share dominance, including... Ahe, murky contracts with OEMs and governments. Depending on perspective, these practices could be interpreted as illegal, but I'll leave that judgment to others. The connection between Intel and Huawei has only recently come to light, then.. perhaps there's still room for optimism.

Most important shady agreement is between Intel and Microsoft. Many people didn't buy AMD CPUs because according to Microsoft their PC was not compatible with win11. However, that incompatibility could be fixed in Bios.

 

[FoulOnWhite]

I wonder if Arrow lake-s is killer if all the Intel detractactors will switch or not. I switch between either as i don't care as long as it is the best. No point sticking to one or the other cause you don't like the other, that is stupid imo. Neither company gives a hoot if you personally like them or not, all they want is your money. I am not a "fan" of either, i think they both have pluses and minuses, just depends which is best when i decide to buy. Should be the same for all of you.

 

[starfals]

I went from i5-2500k to 7800X3D, and if things are going this way with Intel, i might never go back to them again lol. Time will show of course, but so far AMD=awesome!

 

[stimpy88]

Most important shady agreement is between Intel and Microsoft. Many people didn't buy AMD CPUs because according to Microsoft their PC was not compatible with win11. However, that incompatibility could be fixed in Bios.

Wintel is a thing, and probably always will be. We all know the CPU scheduler in Windows "just works" on Intel, and gets fixed quickly if it doesn't, but it still has issues when it comes to AMD, which AMD had to release a band-aid driver for.

 

[ratirt]

AMD is gaining ground in market share but in order for the share to favor AMD CPU's, the next CPU must show its value and performance.
You don't beat intel in a year or two and get 50% share. it will take time for AMD to get close to Intel and I can bet it is not happening this year. Not the next year as well I think.
Slowly AMD will get there if they keep moving forward with advancement.

 

[SL2]

Slowly AMD will get there if they keep moving forward with advancement.

Yup, that's what happened at first, 2016 to 2020 shows a nice improvement.

After that they've had chip shortages, and I have no idea if the pandemic affected AMD differently. Also, desktop competition got a lot better since 2021. It took AMD 3 years to get back to where they once were

 

[Panther_Seraphin]

AMD is gaining ground in market share but in order for the share to favor AMD CPU's, the next CPU must show its value and performance.
You don't beat intel in a year or two and get 50% share. it will take time for AMD to get close to Intel and I can bet it is not happening this year. Not the next year as well I think.
Slowly AMD will get there if they keep moving forward with advancement.

Intel4 is a big thing holding them back. Remember 13/14th gen Intels are on 7nm while AMD is currently on 4nm.

Arrow Lake has a big undertaking as it not only has to get the power efficency under control but it will be facing off against Zen 5 parts in the very near future.

 

[Tek-Check]

After that they've had chip shortages, and I have no idea if the pandemic affected AMD differently. Also, desktop competition got a lot better since 2021. It took AMD 3 years to get back to where they once were

During the pandemic, they massively expanded EPYC penetration due to prioritising volume at TSMC for server chips. Remember, they cancelled Threadripper vanilla focusing everything to get a foothold in server. This strategy paid off, as once you sign long-term contracts with hyperscalers, it will be for many years. It was a window of lucrative opportunities. It wasn't difficult to roughly maintain Zen3 desktop provision for a while and come back even stronger with Zen4 portfolio.

 

[ratirt]

Intel4 is a big thing holding them back. Remember 13/14th gen Intels are on 7nm while AMD is currently on 4nm.

Arrow Lake has a big undertaking as it not only has to get the power efficency under control but it will be facing off against Zen 5 parts in the very near future.

If all that has been said about the Zen 5 turns out to be true, Intel will have a very hard time.
Arrow lake will have to be a huge leap for Intel in performance and power efficiency. I have my doubts for both to be good so my bet is Arrow lake will be performance oriented for the cost of power but might be noticeably better than 13/14th gen. Time will tell.

 

[stimpy88]

Yup, that's what happened at first, 2016 to 2020 shows a nice improvement.

After that they've had chip shortages, and I have no idea if the pandemic affected AMD differently. Also, desktop competition got a lot better since 2021. It took AMD 3 years to get back to where they once were
View attachment 347200

That graph should both embarrass and motivate AMD. AMD really needs to look at prices as well as their position on 3D cache, as they are just too expensive, and more generally, lacking value especially when you look at the entire AM5 platform cost.

Intel will have more (inferior) cores than AMD in their chips very soon, so AMD with their 16 (superior) cores won't look so good from a marketing point of view, and you just know Intel has some very dirty marketing lined up to make their products look better against AMDs, so as AMD seems to have stagnated with their push for more cores on the desktop, they WILL have to lower prices, and even think twice about pushing their 8 core parts as some kind of premium choice, as Intel will be offering more cores at a lower price.

For the record, I'm 100% sure that Zen 5 will be superior to anything Intel releases, but AMD's cache starved designs won't look good against Intel in gaming, which is what drives the consumer market.

It's time for AMD to stop messing around with the 3D cache band aid cash-grab, and just add more and better cache to the CPU itself, without glueing cache over the top of it, hindering performance from a thermal and clockspeed perspective, AMD also need to sort their awful memory controller out too. But AMD won't do that, as they are amateur, slightly naïve and are not cutthroat enough to ACTUALLY take Intel on and definitively beat them.


If I was AMD, I would simplify the lineup, and create only 2 product lines, Gaming and Creator/Productivity (Threadripper remains as HEDT). Ditch the 3D cache with the expensive manufacturing costs, and long time to market, as well as the consumer cash-grab associated with it, and lower prices.

I would push the 8 core version as my low-cost gateway drug entry point for gamers and offer just 2 CPU's to choose from, an 8 core and a 16 core premium gaming chip, with a fair price gap between them. When buying these 2 gaming CPU's I would include a voucher for a rebate on AMD graphics cards.

Then for the Creator/Productivity line I would create a low-cost budget 6 core CPU (from defective 8 core parts) aimed at Internet cafe's and people who only want a simple but performant low-cost computer.

Then a mid-range 12 core CPU, aimed at mid-range users for a bit of everything, gaming/productivity/content creation, made from die harvested defective 16 core parts.

Then a 16 core high-end productivity CPU, same as the 16 core gamer CPU but lower clocked.

I would sell these at a lower margin for a few years, take the hit until market share was showing good growth, (AMD has shown no real market share growth for years) and then I would slowly start increasing margin as I look back at Intel as I surpassed their market share.

I would also make moving to a quad-channel memory design a priority for AM6, which would support 32 core CPU's.

 

[ARF]

Intel will have more (inferior) cores than AMD in their chips very soon, so AMD with their 16 (superior) cores won't look so good from a marketing point of view, and you just know Intel has some very dirty marketing lined up to make their products look better against AMDs, so as AMD seems to have stagnated with their push for more cores on the desktop, they WILL have to lower prices, and even think twice about pushing their 8 core parts as some kind of premium choice, as Intel will be offering more cores at a lower price.

AMD will still have dual-way HT active. Maybe AMD needs to redesign its Zen core and enable 4-way HT? So that a 16-core CPU would have 64 workers.

If I was AMD, I would simplify the lineup, and create only 2 product lines, Gaming and Creator/Productivity

There are slow CPUs needed for basic everyday use cases, which should not be labeled "productivity".

Ditch the 3D cache

Why would they? It enables faster gaming CPUs?

I would also make moving to a quad-channel memory design a priority for AM6, which would support 32 core CPU's.

You don't need faster DDR, because it has little influence on the performance, not worth it.

 

[stimpy88]

AMD will still have dual-way HT active. Maybe AMD needs to redesign its Zen core and enable 4-way HT? So that a 16-core CPU would have 64 workers.

View attachment 347349



There are slow CPUs needed for basic everyday use cases, which should not be labeled "productivity".



Why would they? It enables faster gaming CPUs?



You don't need faster DDR, because it has little influence on the performance, not worth it.

1.) HT makes no sense anymore. Think about it.
2.) Why, it's a name, change it if your uncomfortable
3.) Read to understand why it's there, and then read why Intel doesn't need it.
4.) Incorrect, see answer 3

 

[FoulOnWhite]

3D Cache was a band aid, without it, they would have nothing to compete with Intel. They push 3D cache, do you see them hooting about anything but that, wonder why that is........

Most CPU users are not gamers, so if you don't game, why would you buy a CPU with the 3D cache? Intel CPUs are better at non gaming stuff than AMD's so guess why a lot of non gamers still buy Intel?

 

[Panther_Seraphin]

Hyperthreading is one of those that on paper is a good thing as you maximise the usage of each core but from a security perspective its a massive liability and I still think there is more to come from its mechanics abuse. Hence the movement away from both AMD and Intel. Personally would prefer HT start being removed Zen5/15th gen and be completely gone by Zen6/16th Gen just from a security perspective.

3D Cache was a band aid, without it, they would have nothing to compete with Intel. They push 3D cache, do you see them hooting about anything but that, wonder why that is........

3DVcache makes perfect sense in other areas than just standard PCs. AI/HPC/Database etc in the server space benefit massively from the 3DVcache. General PC use yeah it adds nothing to most people but for gamers/productivity aspects it makes perfect sense.

Also productivity wise there is basically nothing between a 14900K and a 7950x but the 14900k would have between a 5-10% lead in gaming. The 7950x would be 10% more efficent in most things. So I would say a toss up depending on what you do.

Currently DDR5 is technically quad channel on desktops due to the rework of the layout of data bus lines in the standard. Each DIMM is effectivley two 32bit channels vs perviously a single 64bit channel. With >32 cores being possible in a non HEDT platform will we see a return of six or eight channel memory on a standard mobo? Definately can see six channels being viable due to space required for 8 channels in mATX and ATX layouts being extreme.

During the pandemic, they massively expanded EPYC penetration due to prioritising volume at TSMC for server chips. Remember, they cancelled Threadripper vanilla focusing everything to get a foothold in server. This strategy paid off, as once you sign long-term contracts with hyperscalers, it will be for many years. It was a window of lucrative opportunities. It wasn't difficult to roughly maintain Zen3 desktop provision for a while and come back even stronger with Zen4 portfolio.

AMD were onto a blinder with their chiplet designs vs monolithic in this aspect. They basically order as many CCDs from TSMC as they can produce and they can then just decide on the fly if its an EPYC/Threadripper/Ryzen part by changing the I/O die in final manufacturing. If for some reason EPYC didnt fly, they could have done Threadripper and if that was failing they could just send them down to Ryzen parts.

 

[kapone32]

3D Cache was a band aid, without it, they would have nothing to compete with Intel. They push 3D cache, do you see them hooting about anything but that, wonder why that is........

Most CPU users are not gamers, so if you don't game, why would you buy a CPU with the 3D cache? Intel CPUs are better at non gaming stuff than AMD's so guess why a lot of non gamers still buy Intel?

Wow. Without V cache let's look at the 7950X

16 cores at 5.6 GHz, sipping 1/3 of the power vs a 14900K.

MB Compatibility means no MB+CPU upgrade for the next chip but MB or CPU upgrade.

New baseline profile makes the 14700k faster than a 14900k in some scenarios. That is like saying a 7700 is faster than a 7900X.

That is just a quick round up.

What Vcache has done is make 1% lows a thing of the past.

 

[fb020997]

I’m still waiting for the day that I get issued an AMD machine at work. I think Intel still really has the corporate world locked up.

My company (midsized bank here in Italy) buys AMD-only since the release of Lenovo’s T14 Gen 1 models with Zen 2 APUs, and bought quite a bit of Zen+ T495 laptops for a trial vs T490 ones. They unfortunately had to buy ~100 Gen1 intel ones and 20 Gen2 intels during covid, but Gen3-4 and future ones are and will be AMD’s. Waaaaaay better thermals, battery and performance vs. the competition

 

[SL2]

3D Cache was a band aid

Wow, you're more impressed by -KS binning than 3D cache? That's what Intel is hooting about, they have the hottest running CPU, and now customers have to pay the price for it..

If Intel came up with 3D cache first (and I'm not counting Broadwell) you'd never call it a band aid.

3D cache was an EPYC feature first, and later on AMD found out that it worked well in games.

Speaking of band aids..

Rocket lake was the band aid from hell, no further explanation needed.

Raptor lake was a band aid, as Meteor lake was supposed to launch after Alder lake.

Then they postponed Meteor, and threw in Raptor.

Finally, Intel scrapped Meteor for desktop and pasted in Raptor again.

::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid:: ::bandaid::

New baseline profile makes the 14700k faster than a 14900k in some scenarios. That is like saying a 7700 is faster than a 7900X.

Mmm, yeah, reminds me of the sweet days of Rocket lake, where a 10700K beats a 11900K, and a 11900K is the same as a 11700K.

Truly the golden days of Intel, and now they're back!

On a serious note, if anyone has been living under a rocket rock. Those 9900K users must have been pretty happy with their choice.

 

[Tek-Check]

Most important shady agreement is between Intel and Microsoft. Many people didn't buy AMD CPUs because according to Microsoft their PC was not compatible with win11. However, that incompatibility could be fixed in Bios.

Eh? What is this about?
It's Microsoft that has a real problem to move planetary crowd to Windows 11. Adoption rate is only 30%.
Nothing to do with AMD CPUs.