


Intel Struggles To Reverse AMD's Share Gains In x86 CPU Market (crn.com) 87
An anonymous reader shared this report from CRN:
CPU-tracking firm Mercury Research reported on Thursday that Intel's x86 CPU market share grew 0.3 points sequentially to 75.6 percent against AMD's 24.4 percent in the first quarter. However, AMD managed to increase its market share by 3.6 points year over year. These figures only captured the server, laptop and desktop CPU segments. When including IoT and semicustom products, AMD grew its x86 market share sequentially by 1.5 points and year over year by 0.9 points to 27.1 percent against Intel's 72.9 percent... AMD managed to gain ground on Intel in the desktop and server segments sequentially and year over year. But it was in the laptop segment where Intel eked out a sequential share gain, even though rival AMD ended up finishing the first quarter with a higher share of shipments than what it had a year ago...
While AMD mostly came out on top in the first quarter, [Mercury Research President Dean] McCarron said ARM's estimated CPU share against x86 products crossed into the double digits for the first time, growing 2.3 points sequentially to 11.9 percent. This was mainly due to a "surge" of Nvidia's Grace CPUs for servers and a large increase of Arm CPU shipments for Chromebooks.
Meanwhile, PC Gamer reports that ARM's share of the PC processor market "grew to 13.6% in the first quarter of 2025 from 10.8% in the fourth quarter of 2024." And they note the still-only-rumors that an Arm-based chip from AMD will be available as soon next year. [I]f one of the two big players in x86 does release a mainstream Arm chip for the PC, that will very significant. If it comes at about the same time as Nvidia's rumoured Arm chip for the PC, well, momentum really will be building and questioning x86's dominance will be wholly justified.
While AMD mostly came out on top in the first quarter, [Mercury Research President Dean] McCarron said ARM's estimated CPU share against x86 products crossed into the double digits for the first time, growing 2.3 points sequentially to 11.9 percent. This was mainly due to a "surge" of Nvidia's Grace CPUs for servers and a large increase of Arm CPU shipments for Chromebooks.
Meanwhile, PC Gamer reports that ARM's share of the PC processor market "grew to 13.6% in the first quarter of 2025 from 10.8% in the fourth quarter of 2024." And they note the still-only-rumors that an Arm-based chip from AMD will be available as soon next year. [I]f one of the two big players in x86 does release a mainstream Arm chip for the PC, that will very significant. If it comes at about the same time as Nvidia's rumoured Arm chip for the PC, well, momentum really will be building and questioning x86's dominance will be wholly justified.
Shhh... (Score:3, Funny)
don't say "86," someone might think you are out to get rid of a competitor.
86 (Score:2)
I never understood the reference to '86 (I wasn't born in the USA)
Is it something to do with Don Adams? (Get Smart)
Re: (Score:1)
I don't know if your question is in gest or not, but the x86 refers to the CPU-architecture that was first introduced with the Intel 8086 processor that was released in 1978.
Re: 86 (Score:3)
The question was about why 86 means to get rid of something. The hidden meaning is so convoluted to the point that 8647 is interpreted by the US Secret Services as a threat against the POTUS.
https://en.wikipedia.org/wiki/... [wikipedia.org]
https://www.nbcnews.com/politi... [nbcnews.com]
Re: (Score:2)
That interpretation didn't even cross my mind and I didn't see the AC post he responded to, so thanks for the clarification.
Re: (Score:2)
The same idiots crying about this are the same ones that said 8646.
Its a red herring and another distraction to what is actually happening to our country.
The joke above is good, but its terribly dumb that its even a joke.
Re: 86 (Score:2)
First time I have seen that term, so it has to be a quite obscure origin and more info needed behind it.
Back in '86 we had the Challenger disaster and Chernobyl disaster. A pretty bad year.
Re: (Score:2)
First time I have seen that term, so it has to be a quite obscure origin
So, so, so senseless. So, so, so self-centered. But I repeat myself.
There has been a Wikipedia page [wikipedia.org] on that term since 2010 [wikipedia.org]... So, both before Jack Posobiec said it about Biden, and also before he said people should be arrested for saying it about Drumpf. It's meant "to ban someone from an establishment" for decades. Anyone familiar with US bar culture knows it.
Re: (Score:2)
And everyone not visiting US bars because living outside the US don't know it.
It's very US centered thinking.
Re: (Score:2)
It's very US centered thinking.
This is a very US centered event. It's unreasonable to expect everyone around the world to know the term, but when the issue at hand is whether the people involved in the supposed scandal are claiming that everybody knows it means something that everyone actually involved knows it doesn't, the understanding of the term in the US is what's relevant.
Also, it's pretty easy to find out, because it's pretty well documented.
Re: (Score:2)
Here is the story at the BBC [bbc.co.uk]. When I first saw it I wondered if the reporter had been drinking - until I saw it somewhere else!
Re: (Score:2)
Re: 86 (Score:2)
Re: (Score:2)
The fact that the US media is going in a circle jerk around this idiocy just shows how controlled the US media is.
This is by no means unique to US media. Remember, for example, when European media threw a hissy fit because in the background of some video Greta Thunberg posted there was a plush toy Octopus [spiegel.de] visible?
uh (Score:4, Insightful)
Re: (Score:1)
Restaurants obviously. The FBI director is a job that doesn't concern itself with details of slang used by a tiny tiny tiny minority of the people his department investigates. You can spend your entire life working for the FBI and never encounter a "gangster" especially since this is slang from the 30s. You're more likely to know about this watching black and white movies than running or working in the FBI.
Re: (Score:2)
https://bsky.app/profile/oocco... [bsky.app]
Anyway, who knows more about gangster slang, a former FBI director, or a current gangster and president?
Re: uh (Score:1)
Very obscure for someone not from the US.
Re: (Score:2)
Get rid of Paulie. He's a rat.
Both can also mean to remove. Context should surely come into play.
Re: (Score:2)
The police radio code for homicide is 187.
In California. It references the CA Penal Code section for murder.
Note 187 was picked up as slang for murder in hip hop and rap. Same for 86 in the 1920s and 30s.
Re: 86 (Score:1, Troll)
It's a 1920s gangster culture meme. When the mob boss said to 86 someone it meant to take them 8 miles out of town and put them 6 feet under, handy that it also rhymes with nix which was also used in the same context. At some point the military started to use it in their lingo in the general sense of to get rid of something, restaurants and bars started to use it to mean to eject a patron or refuse service.
Re: (Score:2)
What morons modded this up?
Re: (Score:2)
What morons modded this up?
Someone familiar with cultural history. 86 was once used in popular culture just like 187 today.
Re: (Score:2)
Funny how all the citations I find say it’s restaurant lingo.
Re: (Score:2)
Funny how all the citations I find say it’s restaurant lingo.
I learned it from old movies as a kid, I'm of an age similar to Comey. To test your "citations" claim I went to wikipedia:
"According to Cassell's Dictionary of Slang, "to 86" also means "to kill, to murder; to execute judicially".[8][9] Other slang dictionaries confirm this definition.[10][11]"
https://en.wikipedia.org/wiki/... [wikipedia.org]
To see if the "youths" are aware for this I checked the urban dictionary. One of the definitions is::
"A bar/restaurant slang used to describe the end of something or as an actio
Re: (Score:2)
The Get Smart character was a pun on the idea.
To 86 [the item] meant to get rid of it, or to toss it over the side while I was in the Navy.
As mentioned below it had a related meaning in the food service business.
So a burial at sea too (Score:2)
To 86 [the item] meant to get rid of it, or to toss it over the side while I was in the Navy.
So a burial at sea too. :-)
867-5309 (Score:2)
I always thought the song 867-5309 was a love song .. now I find out it's about a dude who was going to murder a woman named Jenny.
Re: (Score:2)
As a verb, it’s a restaurant term meaning to refuse to serve someone, or to show them the exit.
Re: (Score:1)
As a verb, it’s a restaurant term meaning to refuse to serve someone, or to show them the exit.
True, but restaurants adopted it from 1920s/30s gang culture. In 10 years maybe restaurants will be saying to 187 that order, those customers.
Re: (Score:2)
Or Expo 86 in Vancouver, Canada?
Re: (Score:2)
I thought it was a light novel and anime [wikipedia.org]
Re: (Score:2)
And an excellent one at that.
Re: Shhh... (Score:1)
So x86 doesn't imply a cpu architecture at all ? It's random chars. Figment of our imagination. :) Ok Mr TDS. (/* random chars)
Re: (Score:2)
don't say "86," someone might think you are out to get rid of a competitor.
According to a display at Disneyland's "Soarin'" ride, '86 is related to being shot down by an F-86 aircraft.
ARM share in PC processor market (Score:2)
ARM increasing market share in the server area is no surprise, given that they appear to be just cheaper than comparable x86 hardware, at least when you do not need the maximum single-thread perfor
Re: ARM share in PC processor market (Score:2)
I'm not a "gamer", but most of my gaming is done on an ARM CPU, that is on a Mac. Same for my son.
We get to play Minecraft, Terraria and some retro games on these ARM cpus.
We also have and old Wii, PlayStations, and handheld retro consoles. The retro consoles usually have an ARM CPU.
Re: (Score:2)
I'm not a "gamer", but most of my gaming is done on an ARM CPU, that is on a Mac. Same for my son.
We get to play Minecraft, Terraria and some retro games on these ARM cpus.
We also have and old Wii, PlayStations, and handheld retro consoles. The retro consoles usually have an ARM CPU.
Sorry to interject, but the Wii and the PS3 use PowerPC. since PS4, it is X86. Depending on the handheld retro console, you would be surprised by the odball architectures used. PS1 and 2 also use some odball thing, but is not ARM.
Written from my 2018 x86 MacMini. Nothing wrong with having an ARM mac, and playing on it. Having said that, I game on mac by bootcamping + my trusty sonnet TB3 eGPU enclosure (currently Robocop Rogue City, I am a little behind the curve)
Re: (Score:2)
The only ARM console is the Switch (and now Switch 2).
Re: (Score:2)
The only ARM console is the Switch (and now Switch 2).
The PS Vita, Gameboy Advance, DS and 3DS would like to have a word with you..
Re: (Score:2)
Those were portables. Not the same thing.
None of the consoles have ever been Arm, unless you count the Ouya.
Re: (Score:2)
If "PC Gamer" writes about CPUs, I would have assumed they primarily mean PCs used for gaming, and I am surprised to read that 13% of those would be powered by ARM CPUs these days. Nintendo's switch or smartphones certainly do not qualify as PCs, and laptops with ARM CPUs are probably not widely used for gaming.
ARM increasing market share in the server area is no surprise, given that they appear to be just cheaper than comparable x86 hardware, at least when you do not need the maximum single-thread performance.
Actually, you can get insane performance from ARM. Ask fujitsu.
X86-64 on the server will go the way of the mainframe, running legacy workloads at decent speeds with decent reliability.
Re: (Score:2)
You're referring to the A64FX. It was strong in vector performance due to SVE and 512 bit-wide vector support. For certain niche applications, it was really neat. It isn't a serious threat to EPYC.
Re: (Score:2)
Actually, you can get insane performance from ARM. Ask fujitsu.
Fujitsu is a little short on credibility right now after what they did with the UK post [wikipedia.org].
ARM is zero threat to amd64 so far in HPC. All of the fastest clusters use amd64.
Compete. (Score:2)
The proper title is "Intel Struggles To Compete with AMD". Given how Intel has done business for the last four decades, it should be no surprise that once they lost their ability to be coercive that they would struggle to compete. Intel will continue losing until they change or figure out a new way to cheat, and don't expect them to change.
Re: (Score:2)
I've gone with AMD for every build I've done, even when the chips were slower than Intel. On principle, to keep competition alive.
I'll be building a new desktop later this year, and I may do something similar for the GPU. Buy AMD even though the nVidia cards are faster.
I also prefer to buy American, even though it costs more. Although there typically isn't an American option available for most things.
And? (Score:3)
I've gone with AMD for every build I've done, even when the chips were slower than Intel. On principle, to keep competition alive.
The margins for consumer products are quite thin (to keep new competitors out) which is why the "server market" is vastly more important. Buyers in this market (businesses) will not expend resources based on principle.
Re: (Score:2)
Also, AMD has the edge on both price performance and power performance. I think. Haven't deeply geeked the numbers for a while.
Re: (Score:2)
As much as I love AMD and hate NV, for gaming, the 5090 is the fastest, and there's no way around it. Especially if RT is your thing. AMD skipped the halo card market this time around and went for a "mid tier" card in the 9070XT. It's a solid card for its price range, but it does not challenge the 5090, and in many cases isn't even faster than the 7900XTX in raster performance.
Drivers aren't even the 5090's worst problem. Power delivery, on the other hand . . .
Re: (Score:2)
Intel basically did nothing for four years if not more.
The Iris Pro 5200 from 2014 benchmarked at 1200. I had an i5-4570R with that onboard GPU and it was entirely adequate.
The HD 630 from 2018 benchmarked at 1100. WTF? No gain at all in four years? What were you doing? Oh, stock buybacks.
For the record my current CPU is a Ryzen 4600G at 2400. It came out in 2020.
Re: (Score:2)
Intel did do something: struggled with process node issues.
Re: (Score:2)
Losing their monopoly grip is really a follow-on result from the main factor: clinging to their own process technology while TSMC rides the cell phone wave to massively higher volumes and thus higher capex, especially R&D.
Re: (Score:2)
What I would really like to see in X86-64 (Score:2)
Is Centaur resurrecting the CNS core as a laptop chip, compatible with Win11 (or even 12, one can dream).
Reports are contradictory on the AI side, some say it had an NPU, other say it had extra instructions (similar to intel's AVX-VNNI), but no matter what was the case, it was capable of doing a decent TOPS. All of this with 2019 TSMC 16nm tech.
Imagine the same chip in TSMC's 4nm+ with a DX12 FL12.1 PowerVR GPU, aimed at laptops...
One can dream, and we need competition both in the X86-64 market , and in the
Re: (Score:2)
Centaur is dead, and all their tech has been absorbed by VIA/Zhaoxin.
Marketing Can Not Solve Incomptence (Score:4, Interesting)
While everyone was innovation out there, Intel thought it could stick to their pre-21st century proven technologies by slapping meaningless buzzwords on their products.
Re: (Score:2)
They are meaningless? Damn, here I was avoiding Intel specifically so my computer wouldn't be AI ready and I couldn't have AI shoved down my throat.
Losing a battle that isn't worth winning (Score:2)
This battle they are struggling with isn't even a battle worth winning. Microsoft is working as fast as they can to leave the x86 platform behind.
Looking forward to Windows on Apple Silicon (Score:3)
This battle they are struggling with isn't even a battle worth winning. Microsoft is working as fast as they can to leave the x86 platform behind.
They'll be working faster once the Qualcomm exclusive ends. Native dual booting on Apple Silicon will be awesome, as it is a Parallels virtual environment is pretty good for Windows 11 ARM.
Re: (Score:2)
75% (Score:2)
I actually had no idea intel still has the large majority of x86 sales. Then why are they struggling financially compared to AMD?
Re: 75% (Score:1)
Re: (Score:2)
It might have something to do with their fabs not making the chips any longer. The current gen Core Ultras are all TSMC made. And those are what goes into laptops. And laptops are where Intel still dominates the sales. Mainly from corporate buying inertia but that dominance also still dictates what's on offer in that segment.
That leads me to conclude Intel's fabs will be sitting somewhat idle right now. Which can only be putting more pressure on margins already a lot thinner than what Intel was used to
Re: (Score:2)
Low margins.
Re: (Score:2)
Another possible reason for the large listed Intel share is the totals might not include any CPU sales for DIY builds. If those are not even included in the desktop numbers then that's going to be missing an awful lot of AMD's growth. That would then mean 100% of the numbers are only from HP, Dell and the likes. And we know those PC configs are all skewed towards Intel.
Intel did dirty business and deserves to lose (Score:2)
It's nice when the market actually, finally, punishes companies that deserve it. Dont care that intel is us based. It actually makes me wish for their failure more because i'm not a fan of us companies relying on underhanded tactics instead of quality. They're not lonely in that....unfortunately.
Risc-V will upset everything (Score:2)
At higher end performance there are lots of p
Re: (Score:2)
Raspberry Pi RP2040 has extra Risc-V cores
Minor adjust: I think it's the 2350 that does, not the 2040.
Re: Risc-V will upset everything (Score:1)
Re: (Score:3)
don't forget about China, who has strong economic and political reasons to be free from foreign control of such critical technology. Even if they can't keep up with TMSC they have a vast domestic market and could compete in the international market by being cost effective, even if they aren't the fastest.
China's fastest CPUs are based on licensed Zen cores. I agree that RISC-V will continue to make gains because of the licensing model, but not because homegrown Chinese designs push performance past other architectures in the foreseeable future — because if that happens, it won't be soon. It might happen... eventually. Also, the major limiting factor remains process tech.
IMO, China's best move is to work on next-generation process tech while getting by with what they can get cheaply until then. They co
Why would anyone buy Intel these days? (Score:2)
Why would anyone who is building a PC in 2025 pick Intel when AMD's Ryzen chips are clearly better in just about every way?
Even on the server, the AMD Epyc chips are much better than what Intel is doing.
Are there any places left where Intel chips are actually better (either overall or in terms of value for money) than AMD or are people just buying Intel because "its what they have always bought" or because "its what Dell/HP/Lenovo/etc will sell me" or "its what the sales guy at the PC store pushed me to buy
Re: (Score:2)
This is largely due to the fact laptop vendors are soldering wifi now. Intel wifi chips are stable and well supported on Linux, while everything else remains shit for linux driver support. All of the soldered wifi included in AMD and qualcum laptops drops wifi on linux every few minutes, if the OS even picks up the wifi hardware at all. Meanwhile youre AX211's and BE200's work great.
This is a moot issue on a LAN wired desktop though,
Intel struggles (Score:2)
That really sums it up. They have screwed over their customers time and again, they have failed to adjust their tech, they are behind on concepts, security, etc.
They thought they were invulnerable. There is a price to pay for that level of arrogance and incompetence. They are beginning (!) to pay that price now. I wish them a slow and painful death, which they deserve.
Low-End (Score:4, Insightful)
I haven't bought i3/5/7/9 for a decade (all Ryzen/Zen in that class) but I do buy Intel's low-power parts like the Atom descendants, n100 "Celeron" etc. Totally usable 12-32GB nodes in under 14W.
Their naming system is so broken I can't do anything without going to the ARC datasheets. "You call this a Celeron but it's a completely different lineage than what you used to call Celeron."
Does it have hardware vm? How about AVX512? What can I do with frequency scaling?
They are quite good and compatible if you DYOR.
The RPI5 is also quite good and a little cheaper. I have at least one in production with dual nvme. RISC-V is in the lab but doesn't yet have the price/performance sweet spot that Intel vendors have. We'll be ready when that inflection point hits and will take a small hit for the open ISA.
AMD doesn't seem too interested in this market though their new chip will have an auxiliary RISC-V core. Might be cool for development if QEMU can access it.
snooze (Score:2)
You snooze, you lose. Intel snoozed. Next question, please.
LNL (Score:2)