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.
Five Years After Apple Broke Up With Intel, Intel is Begging for Money
- Thread starter MacRumors
- Start date
- Sort by reaction score
As someone in IT, I can verify this. We have a senior exec where I work who was using an Intel PC and hated the awful battery life and general slowness of the computer. We got him an ARM PC a few months ago and he hasn't looked back since and has praised it to everyone in the company. We're actually looking to see if we could move the majority of our Windows users from x86 to ARM within the next 2-3 years as we upgrade them from older machines. Hell, the ARM machines are actually cheaper than x86 equivalents, even after customizing their storage and RAM to fit what we need.
It's the software that's our concern, not even the hardware. But Microsoft recently announced that 90% of apps used on ARM PC's are natively running on the architecture and not through an emulation/compatibility later, so that issue may be resolved for us in the next year or two, barring exceptions.
I was very surprised when I read it. It sounds like something a tabloid would post.
Sure they did.
Apple demonstrated that you could go faster with less power and still run the same types of software using RISC and other companies took note. Intel tried to deny it, but there was no denying it.
It wasn’t as if everyone didn’t know it was POSSIBLE to do it. It’s just Apple showed it was feasible and marketable and provided advantages for the way that personal computing was headed: portable with extended battery life. The key was that it REQUIRES unified memory and SSD to work. Because RISC is memory intensive, the architecture had to be able to address all memory so that you didn’t have to keep swapping it around between cache, memory, cpu and video. Apple spent 10 years on the A chips mastering this setup before they brought it to the M1 as a live demonstration for the entry level computers.
That approach is admittedly limiting on the high end, but the mass market isn’t the high end. It’s the mid and low end and ARM dominates that segment now in all form factors.
To be fair it seems like Apple poured far more money into the division than Intel did, and they had both a direct negative financial motivation (qualcomm’s cuts of the profit of the Apple devices their chips are in) and a single customer to design for (internal), both big catalysts to get the work done
I agree with those who state Intel's only real value to Apple is as a secondary chip fab, though I expect this would be a very long-term investment with a very slow ramp.
TSMC's US plants are fabricating older Apple Silicon generations that use older and larger processes. Can Intel also handle this, or are the designs locked into TSMC's processes? If Apple's silicon designs are designed exclusively around TSMC's fabrication processes, having to bi-furcate them to support both TSMC and Intel would likely be too much of a pain for Apple unless they are forced to by the US or Chinese governments.
TSMC's US plants are fabricating older Apple Silicon generations that use older and larger processes. Can Intel also handle this, or are the designs locked into TSMC's processes? If Apple's silicon designs are designed exclusively around TSMC's fabrication processes, having to bi-furcate them to support both TSMC and Intel would likely be too much of a pain for Apple unless they are forced to by the US or Chinese governments.
Imagine telling someone in 1995 that in 2025 Intel would be looking for a lifeline from Apple.
Intel’s failures are mostly poor business decisions with improper investment and improper strategy. There are still a lot of very capable engineers and scientists on staff. Intel’s current issues are 100% a failure of the C-suite and the board that oversees it.
Anyhow, I can see Apple asking Intel to engineer some stuff, leveraging Intel’s deep, highly expert, but under-utilized bench.
Anyhow, I can see Apple asking Intel to engineer some stuff, leveraging Intel’s deep, highly expert, but under-utilized bench.
Intel doesn't need any more money, they got Softbank, Nvidia and let's not forget the government.
What Intel needs is customers for their processes, they failed with 18A to attract anyone "big", now they appear to be focussing solely on 14A, but you need your fabs pretty full to be competitive.
Would 14A makes sense for Apple? maybe, but then we are either dealing with 2nd source (bad idea, no 2 foundries, the companies, are "compatible", requires different design and will have performance differences), or, some specific products/chips are only made at Intel.
Betting on 14A for specific products - huge risks, Intel has no track record ...
And where does that leave the relationship with TSMC? TSMC has executed brilliantly, why leave them?
Intel needs to be successful to the degree of being able to support government, military etc, but the doesn't keep even their existing fabs full, leave alone the planned ones.
Also, Intel themselves are outsourcing to TSMC, so the chip groups know that Intels foundries are in trouble.
Big hurdles, and I think that their new CEO is in way over his head, it was a mistake to kick Gelsinger out.
What Intel needs is customers for their processes, they failed with 18A to attract anyone "big", now they appear to be focussing solely on 14A, but you need your fabs pretty full to be competitive.
Would 14A makes sense for Apple? maybe, but then we are either dealing with 2nd source (bad idea, no 2 foundries, the companies, are "compatible", requires different design and will have performance differences), or, some specific products/chips are only made at Intel.
Betting on 14A for specific products - huge risks, Intel has no track record ...
And where does that leave the relationship with TSMC? TSMC has executed brilliantly, why leave them?
Intel needs to be successful to the degree of being able to support government, military etc, but the doesn't keep even their existing fabs full, leave alone the planned ones.
Also, Intel themselves are outsourcing to TSMC, so the chip groups know that Intels foundries are in trouble.
Big hurdles, and I think that their new CEO is in way over his head, it was a mistake to kick Gelsinger out.
IF that were going to happen, we have way bigger problems than chip supply ...
We need Intel to survive. Intel needs to show us that it wants to survive.
of course they are, you have to design around the process and the same, let's say 5nm process is different at TSMC, different at Samsung, different at Intel. Requires at a minimum design changes and a complete different mask set. In addition, the differences between processes also mean differences in performance.
Remember when Apple sourced the A chip (can't remember which one) from both Samsung and TSMC? That din't go well ...
There was a lot going on with that relationship, more than just being a customer, and Intel was the primary benefactor. Apple paid top dollar for guaranteed inventory and first access to high end chips, which left the other PC vendors sitting around waiting for yields to increase (and subsequently, prices to drop). There was also component integration/board reference designs that Apple took from Intel in the first few generations. And, Apple championed Intel’s EFI architecture, which was again a more expensive option then the BIOS that other manufacturers were sticking with. All parts of a relationship that Intel already had in-house, and Apple paid for, and didn’t really cost Intel anything extra.
Switching from PPC to intel was a necessary move at the time. And clearly Apple already knew that it could come to that. Hence having Mac OS X already compiled and running side by side the whole time on intel CPU.
Since they had a roadmap with Motorola and IBM (G4-G5). They let it go as far as it could. Since it did perform better than intel, and was a major differentiator between the Mac and PC. Much easier to say it's better when they run completely different CPU's. And of course that the G4 was pound for pound better than intel (to a point). And so to the G5 (to a point). But, the laptop is really why the change needed to happen. No way to put a G5 into a laptop effectively. I'm sure they could do it, but it would be like 1.3GHz and a thick boy.
If anything, they could have picked AMD. First to 64bit with the Athlon and performant enough to match intel in the areas that mattered. Plus they ended up owning ATi for GPU's, so a win-win on hardware. Maybe they don't own the performance crown for a bit in every category. But, eventually they would have been "here".
The move to ARM however I would argue was the best thing they ever did. They don't have to wait for anyone else to come up with something "new" in a CPU or something that would benefit Apple specifically. They can go their own way in any direction Apple sees fit. If they feel they need a better GPU. They can build it. If they feel they need more performance. They can build it. And starting with a low power design. Lets them scale up to higher power systems much easer than say getting an EPIC or Xeon chip to work in a small tower. Which you would then need a discrete GPU to fit into (a la nMac Pro). And be dependent on yet another vendor making something that works well for you and your particular use cases. M series chips is the best move IMHO than the move to x86-64 or picking between ATi(AMD) or nVidia.
I run a Windows VM with 8GB of ram assigned to it. Runs well enough for its specific use case. More is better but it's not a fail.
I have a PC running Linux and this MacBook Air. They are used for different purposes so there isn't much overlap. Libreoffice runs on both.
The PC is running an AMD chip though. Intel's integrated graphics were a sad joke when I was shopping.
If this MBA still works when Apple is done supporting it, (and possibly sooner if Apple insists on jamming the AI down my throat) it will go to Linux as well.
Shame on intel because apple gave them a grace period when they were working on their own in house chips should’ve spend time innovating since then and here are the results failure