News Intel Says First 10nm Desktop CPUs Land in Second Half of 2021

InvalidError

Titan
Moderator
When Intel says one day that it is ramping down plans for 10nm in favor of fast-tracking 7nm, then 7nm trips over its feet, we get an announcement of probable 10nm desktop chips that haven't been on any roadmap I have seen before after it seemed like Intel's intent was to skip desktop straight to 7nm.

Intel's senior management must be really sweating now, possibly two more years of being stuck in process quicksand.
 

JayNor

Honorable
May 31, 2019
456
102
10,860
I've seen a few mentions of a Tiger Lake-H chip, which would be a 10nm part for desktop, using an beefed up power spec for a laptop chips. So, I doubt Alder Lake-S will be their first 10nm desktop chips, as this article states. Did Swan actually say that?
 

watzupken

Reputable
Mar 16, 2020
1,175
660
6,070
I've seen a few mentions of a Tiger Lake-H chip, which would be a 10nm part for desktop, using an beefed up power spec for a laptop chips. So, I doubt Alder Lake-S will be their first 10nm desktop chips, as this article states. Did Swan actually say that?
Alder Lake is the first 10nm desktop part. Between now to Alder Lake launch, we have Comet Lake and Rocket Lake, both which are based on 14nm fab utilizing the existing 1200 socket.
 

watzupken

Reputable
Mar 16, 2020
1,175
660
6,070
When Intel says one day that it is ramping down plans for 10nm in favor of fast-tracking 7nm, then 7nm trips over its feet, we get an announcement of probable 10nm desktop chips that haven't been on any roadmap I have seen before after it seemed like Intel's intent was to skip desktop straight to 7nm.

Intel's senior management must be really sweating now, possibly two more years of being stuck in process quicksand.
With each trip up, they are extending their pain. So this 2 years of process quicksand may potentially become longer. If 7nm is indeed in trouble just as 10nm, then I presume it they may have to make compromises for it to work.
 

D1v1n3D

Distinguished
May 8, 2015
17
12
18,515
This is going to make AMD 4xxx even more juicy so long as they pull off better gaming performance than current Intel gen as next years 5xxx series will be 5nm so this means Intel wont have PCIe 4 or 5 gen anytime soon, goes the same for DDR5 support. GO AMD lol, I never liked Intel's tactics of business, now karma is catching up to them.
 

InvalidError

Titan
Moderator
this means Intel wont have PCIe 4 or 5 gen anytime soon
10th-gen was supposed to get PCIe4 but Intel ran into unforeseen issues and had to postpone that to 11th-gen. Unless Intel has yet more unforeseen complications, LGA1200 should get PCIe4 for CPU-fed lanes when 11th-gen launches. Where PCIe5 is concerned, I'm skeptical about that making it to consumer-friendly motherboards any time soon due to how difficult maintaining signal integrity through a socket + PCIe or M.2 slot at 32Gbps is going to be.
 
Intel desperately needs 10nm CPUs to fend off AMD's increasing market share in the desktop PC market, but they aren't coming soon.

Intel Says First 10nm Desktop CPUs Land in Second Half of 2021 : Read more
AMD's continued push with its 7nm Ryzen processors has caught Intel flat footed as it continues to lose market share in the desktop PC market,
Any link to back this up?!?!
Maybe this should be spelled "in the "enthusiast" desktop PC market" ?
Which is a niche of a niche.

Because all we normal people have access to is intel's quarterly reports and there intel has doubled their net income in the last two years,so either they are selling the same amount of CPUs for double the price or their market share is way way way higher than anybody else's and they would have no issues with losing some of this doubled market share.
 
According to Intel's announcement today, we won't see the new chips until the second half of 2021.
Was anyone actually expecting otherwise? They were obviously not going to release both Rocket Lake and Alder Lake on the desktop within a year of Comet Lake. That would make little sense, so no one should have been expecting those processors to appear in the first half of the year. Most likely, we won't see Rocket Lake until early next year, and Alder Lake likely won't be coming until the tail end of next year.

Any link to back this up?!?!
Maybe this should be spelled "in the "enthusiast" desktop PC market" ?
Which is a niche of a niche.

Because all we normal people have access to is intel's quarterly reports and there intel has doubled their net income in the last two years,so either they are selling the same amount of CPUs for double the price or their market share is way way way higher than anybody else's and they would have no issues with losing some of this doubled market share.
Of course Intel's income was going to improve once they finally started releasing processors with higher core-counts after spending the better part of a decade stagnating on four cores with only minor performance improvements each year. Until AMD started releasing competitive processors with higher core counts, and Intel followed suit, people saw little reason to upgrade, as the new processors were more or less the same as what they already had. Now that more capable processors are available in every price segment, people see more reason to upgrade. So sure, Intel's processor sales are going to be up, but AMD's have undoubtedly climbed by an even greater percentage over where they were previously, and as such, Intel has been losing a share of the market. And that goes even more so for the server and HEDT processor markets, where Intel had more or less no competition whatsoever for a number of years.

Sure, Intel still holds a much larger share of the market, as companies are not going to jump to AMD in droves overnight, nor does AMD likely have the production capability to make that happen in the short term, and the article's claim that Intel "desperately" needs 10nm desktop CPUs to ward off AMD is an exaggeration, but AMD is certainly being far more competitive in every market than they had been for years. And perhaps just as important, they are building positive mindshare, that is likely to continue their growth for some time to come.

And while Intel might be happy to see increased sales in the short-term, I imagine they would have preferred to trickle out core-count increases at a slower rate, rather than needing to effectively slash the prices of their processors by more than half over the course of a couple years. A lot of this increased income is likely future sales that have been shifted forward.
 

thisisaname

Distinguished
Feb 6, 2009
929
518
19,760
Only last year the plan was
3UQVWpDMouCdErNzacEMq-650-80.png.webp


I would not bet they can stick with this timetable, even if you just push everything back another year or two.
 

jkflipflop98

Distinguished
The real problem here is the clueless executives.

Traditionally, Intel has Pathfinding > PTD > Manufacturing. Pathfinding is research, PTD is development, then finally on to the HVM fabs to make the parts for consumers. Historically PTD is tasked with development of two processes like 14nm that's ready to go out the door so we start on 10nm to get it ready.

PTD is currently tasked to capacity with development of six radically different processes as well as being forced to also function as a HVM facility for 14nm. Even with mighty D1X, we simply don't have the manpower to handle all of this. And every month they shovel something new on our plate.

So there ya go, there's the real reason we're behind for once. Don't listen to people like "cryoburner" up there ^ they don't have a clue what they're talking about.
 
Last edited:
So there ya go, there's the real reason we're behind for once. Don't listen to people like "cryoburner" up there ^ they don't have a clue what they're talking about.
What are you even talking about? You clearly didn't even read my post if you think I mentioned anything about why Intel is behind on their release schedule.

As a quick summary, my points were that...

1) The article makes it sound like people were expecting Intel to release three generations of desktop CPUs within a single year of one another, but that wouldn't make much sense from a business standpoint, so obviously they never planned to do that.

2) The claims of the person I quoted about Intel only losing market share in the enthusiast desktop PC market are not really accurate, even if they are still well ahead and arguably fine from a revenue standpoint.

3) They were undoubtedly pushed to increase core counts at any given price point, and would have preferred to roll-out more cores at a slower rate to maximize profitability had the lack of competition continued.

At no point in that post did I mention anything about why Intel might be behind on their process roadmap. It's like you only read the first sentence of my post and assumed what the rest was about. If you actually work for Intel, and can't even bother taking a minute to read a post before dismissing it, then perhaps that might be indicative of where the company's problems actually lie. But hey, we might as well just blame the upper management. It's clearly someone else's problem. : P
 
The real problem here is the clueless executives.

Traditionally, Intel has Pathfinding > PTD > Manufacturing. Pathfinding is research, PTD is development, then finally on to the HVM fabs to make the parts for consumers. Historically PTD is tasked with development of two processes like 14nm that's ready to go out the door so we start on 10nm to get it ready.

PTD is currently tasked to capacity with development of six radically different processes as well as being forced to also function as a HVM facility for 14nm. Even with mighty D1X, we simply don't have the manpower to handle all of this. And every month they shovel something new on our plate.

So there ya go, there's the real reason we're behind for once. Don't listen to people like "cryoburner" up there ^ they don't have a clue what they're talking about.

We and we're.... So you work for Intel then? It sounds like the upper management needs to be engineers not business people because no engineer would spread development that thin. Similar thing with AVX-512 which was a terrible idea to introduce into consumer chips while still on 14nm based monolithic designs. Its like you say hey we can have more cores or AVX-512 that virtually nobody uses outside of servers and the management said we don't need more cores lets implement AVX-512. Freaking clueless.
 

JayNor

Honorable
May 31, 2019
456
102
10,860
Which 7nm CPUs were delayed? I only see Meteor Lake and Granite Rapids CPUs listed on wikichip as being planned for Intel 7nm and both have always been listed as in 2022.
The only CPUs I've seen on Intel's roadmaps for 2021 are Alder Lake, Goldmont, Rocket Lake and Sapphire Rapids and all are 10nm except the 14nm Rocket Lake.

So, looks to me like the only 7nm CPU delays are for parts that were already scheduled out in 2022.