Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Intel Delays 7-Nanometer Chips Until Late 2022 or Early 2023
- Thread starter MacRumors
- Start date
- Sort by reaction score
That must've been a hard announcement for them to make. With the very direct threat AMD has been over the last couple of years, and with Apple making the move they're making now ... an announcement like this can only further erode people's perception that Intel is going to continue to have any relevance in the future.
You're definitely right about Apple having deep access to Intel's internal roadmap but this Apple Silicon Mac transition is years in the making. The decision was made around the time the 64-bit A7 SoC was introduced.
The real problem is that the 2023 date won’t hold. If they’re already copping to it this early, it’s going to be a bumpy ride to 7nm.
If I had to bet, just based on their track record, I’d guess Intel will be on 10 until 2024-2025. TSMC is starting risk production of 3nm next year with volume in 2022. You’re talking about at least 2, maybe 3 generation lead by the time Intel nails 7nm.
Last edited:
Isn't that more a function of the GPU than the CPU? Users with the 5600 not great improvement in terms of fan noise with that GPU.Move over, Intel. You're the 2020 version of the PowerPC. Can't wait for ARM. Typing this on a 16" MBP with blaring fans just because an external monitor is plugged in.
My 16” plugged into my 4K at 60hz will sit the Radeon pulling 19 watts just looking at an idle desktop - above 80 degrees out and I’m looking at least 3000k rpm. Very normal experience with the 16” (and 15s before it), I don’t think the CPU is to blame here however
Maybe. One would hope they have learned from their mistakes in the 10nm transition.
We haven't even seen a TSMC 5nm desktop/server class CPU yet. But TSMC's execution since ~2017 has certainly been impressive. If they can keep it up, Intel will have a hard time catching up.
But what was almost lost in all the 7nm hoopla is that Intel had a record quarter and beat expectations, with strong growth particularly in the server market. I don't think they are at risk of going under anytime soon ...
Man, this is not a good look for Intel. Third party foundries? That's an earth shattering statement.
For a couple decades now, Intel has really only had one thing going for it: their process. The x86 instruction set is antiquated and crufty. They fail repeatedly at designing new architectures. AMD is making it clear that Intel can't even ship the best version of the Intel Architecture. Aside from some hobby projects, they can't seem to ship anything that isn't an x86 chip or an x86 support chip. The only thing that has kept that that company alive is that their process technology is advanced enough to keep their crippled architecture competitive.
Now it's clear their process team has let them down. Intel is going to come out of this a significantly diminished company. I'm not sure why there isn't a full on shareholder revolt. The executive leadership has one job to do-- keep the fabs running. If there's no ROI on fabrication, then it's time to shut down the capital intensive part of the business and move into IP licensing. Let TSMC give AMD a run for their money.
As it is, when Apple announced their switch it was seen as a small player taking a risky gambit. Now it looks prescient.
I wouldn't be surprised if the reason we aren't hearing about Bootcamp support for Windows Intel on Apple Silicon is because Microsoft is planning to use Macs as a test bed for wider deployment of Windows ARM and they don't want to show their hand just yet.
For a couple decades now, Intel has really only had one thing going for it: their process. The x86 instruction set is antiquated and crufty. They fail repeatedly at designing new architectures. AMD is making it clear that Intel can't even ship the best version of the Intel Architecture. Aside from some hobby projects, they can't seem to ship anything that isn't an x86 chip or an x86 support chip. The only thing that has kept that that company alive is that their process technology is advanced enough to keep their crippled architecture competitive.
Now it's clear their process team has let them down. Intel is going to come out of this a significantly diminished company. I'm not sure why there isn't a full on shareholder revolt. The executive leadership has one job to do-- keep the fabs running. If there's no ROI on fabrication, then it's time to shut down the capital intensive part of the business and move into IP licensing. Let TSMC give AMD a run for their money.
It may have been agreed that Apple would announce their switch before Intel announced their delay-- this is all bad news for Intel, but I think if Apple was seen as jumping ship after this announcement, the optics would have been worse.
As it is, when Apple announced their switch it was seen as a small player taking a risky gambit. Now it looks prescient.
I wouldn't be surprised if the reason we aren't hearing about Bootcamp support for Windows Intel on Apple Silicon is because Microsoft is planning to use Macs as a test bed for wider deployment of Windows ARM and they don't want to show their hand just yet.
Apple doesn't need to scale the process. They can just shop their chips to whoever does it best, or share their business across a couple and watch them compete for their share of the pie-- just like they do with a lot of iPhone components today.
Last edited:
Apple user can laugh at Intel over their planned ARM chips, PC user laughing with their AMD rigs. Intel chips not competitive anymore, just plain heat and heck latest comet lake board price still doesn't make sense in 2020.
The most people I know retain Intel at this time because they still use hackintosh till the Apple drop the ball.
The most people I know retain Intel at this time because they still use hackintosh till the Apple drop the ball.
Bob Mansfield. Not Srouji. Mansfield is the secret weapon behind this. Srouji is just taking the credit. This is why he left the leadership page.
I don't think this means they'll outsource manufacturing of their mainline CPUs. They now have the capability of packaging chips with different process nodes (Foveros). They'll probably outsource things like IGPUs, Atom CPUs and similar things.
Their architecture is top notch. The only reason that AMD has been able to catch up (and overtake in some areas) is TSMC's superior process node. I bet the upcoming Intel 10nm Cores will clobber AMD 7nm in per-core performance.
Their storage unit is growing strongly, and they are a becoming a leading supplier of chips for 5G basestations.
I think it's exactly the opposite. Their architecture is great, but the process technology is holding them back.
There is no other foundry that is competitive with TSMC. Samsungs nodes have inferior package density.
I don't know that I'd bank on it either, but MS needs to untether themselves from the Intel anchor. Nadella has done a good job of making them a more flexible company. They've dabbled with Windows on ARM in some niche products, but Apple Silicon is the first chance they'll have to show it really running on a fully capable ARM CPU delivered by a manufacturer they know will stand behind their product and continue it for a few generations at least.
And it's not like MS hasn't promised to support other architectures in the past, only to orphan the product later. They've always like to treat Apple as a special case that never quite reaches parity. I don't see where the risk to MS would be.
I too suspected they had macOS running on A series with the A7 at the time. When Phil Schiller described the A7 as having desktop class performance, I said ok, so they have macOS running on this thing already.
Steve Jobs was the weapon behind this. It was under his leadership that Apple acquired PA Semi and later on Intrinsity to build the semiconductor team.
Apple Is Quietly Designing and Building Its Own Silicon Empire - ExtremeTech
Apple is quietly building a semiconductor empire for itself that could, in the long-term, challenge Qualcomm and Intel on modems, and even Intel on CPUs. That last bit won't happen anytime soon, but it's not impossible.
www.extremetech.com
