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.
Apple and Intel Rumored to Partner on Mac Chips Again in a New Way
- Thread starter MacRumors
- Start date
- Sort by reaction score
Back when Intel was competitive their density was actually a little better than TSMC's. I'm really hoping that they can make 18A happen.
Just don't let them design the damn thing.
Just don't let them design the damn thing.
That is interesting, I am not sure if Intel can do 2nm any cheaper then TSMC, if they can, they may just have a chance at being a supplier for Apple. Can't hurt having another alternative to TSMC for manufacturing chips
It’ll have to be someone else if it’s going to be on par with what we’d be losing. Even staring into the face of one of their large customers leaving, even THAT wasn’t enough to wake them up and have them do anything but the mediocre job they continued to do. There were several opportunities in the past that could have set them on the track where they’re making the kinds of chips that people want today, but they made the wrong choice every time.
There’s nothing showing that they’ve ever started to make thee right choices yet. Fabrication in North America saddled with Intel would just yield inferior products that still don’t meet the lowered requirements we’d have to set to help them appear to be successful.
How would they do that if they're only being contracted to manufacture somebody else's design? This isn't them developing a CPU from scratch.Not to worry. Intel, as they ALWAYS do, will miss performance and TDP goals. They can’t NOT do that, it’s in their DNA. I’m sure the contract contains “If you’re unable to deliver the required product, this deal is nullified and we’ll continue to use TSMC.” If this rumor is true, in about a year we’ll be reading about how Intel failed again.
In the end, it means nothing for Apple because Intel won’t deliver and Apple will just continue to do what they’ve been doing.
How would Intel fail? Similarly to the way they have over and over again. If they can’t produce a design that they know VERY very well in a way that meets performance and efficiency goals, what’s the chances of them hitting someone else’s goals with someone else’s designs?
Intel: “So, here are the first test batches, what do you think?”
Customer: “Am I reading this right? They require 400W of power?”
Intel: “Yup! Wait, don’t ALL CPU’s require 400W of power? We thought that 40W number was a misprint. There’s NO way a chip can run on that much right? I mean we make our own chips and 40 watts wouldn’t even prefetch the cache.”
Diversifying with Intel would be like diversifying my professional soccer team by bringing in some 3rd grade dodgeball players. I mean, sure I can CALL it diversifying, but it is, in no real way, diversification.
What’s funny, is Intel HAD an impressively Strong ARM game.
Looked out across the industry and, unlike everyone else, came to the conclusion that “ARM just isn’t the future”.If it fails, Apple will have wasted all the resources they put into it.
If it succeeds, Apple will be paying more for the same product.
In either case, bad for $AAPL. Good for $INTC, at least in the short term.
Intel 18A should be at least if not more expensive than TSMC N2 ( none of this stuff is ‘nm’) . 18A has gate all around and backside power ( flipping chip over and printing g more stuff is an extra manufacturing step . It does not lower costs ) . TSMC has pushed back back and is only doing gate all around for N2 . TSMC has more EUV fab machines than Intel also. Regardless of fab process being good or not Intel can’t do TSMC volume. Intel isn’t out to be the low cost , max volume leader. They need to be a high margin leader to chips for somewhat unique performance . [ Intel under invested in acquiring EUV equipment for so long they a never going to catch TSMC with the current high production tech . Samsung and TSMC bought lots more for years. Intel bought more DUV tech and the end of range with zero external customers to sell that capacity to as they transitioned off. Bone move on multiple de
Imension but sure did inflate the stock price for more than several years. The Wall Street crowd lapped it up as whittled future away. ]
TSMC N3P should cost less than 18A and likely wins on maturity yield metrics . There was a N4C fab process that went for lower costs ( and long term availability ) late in the cycle. There is probably a N3c coming also .
Intel fab has more of a chance because demand on paper with AI hysteria is going to outstrip TSMC capacity and Samsung is still somewhat fumbling the ball.
18A has ramped with some yield hiccups. If they don’t find some 14A customers then revisions of 18A will be ‘stuck’ also. What would likely see is variations with different industry target optimizations. There is a breadth’ of offerings that Intel is missing in addition to doing battle just at the blazing leading edge.
Intel has there reaserch fab in the USA ( Oregon) . TSMC does not. The USA fabs are running 12-20 months behind . If Taiwan goes then the lead research falls into an extremely deep pothole that would takes many years to fix. ( not just missing production capacity)
Occasionally run? Sounds like a perfect use case for a cloud vm for doing builds.
Err. The plain Mn chips grossly outsell the rest of the line up. Numerous preenetstions Apple reports that the MBA and base MBP are the two best selling Mac’s. That is where the bulk of the volume comes from. Nevermind the the timing here is a bit odd (**)
[ if it is suppose to be majority of the M series die variants/types, then it still doesn’t make sense For example if Pro/Max were both composed from the same chiplet build blocks there is not much left of variations ]
If anything it would make more sense to move the lower volume stuff to Intel ( e.g,, the private cloud compute server chip ( with chiclets from Broadcom as a component ) , Ultra , Max ) . Similarly if more real chiclets in M-series in 2027 then some chiplet could be Intel and others TSMC and perhaps Intel packages them together .
Lowest end in terms of performance , they can do that now with iPad AIr chips coming from TSMC AZ .
What in the world does this have to do with the rumor even in the slightest?
P.S. Intel 18A is in high volume manufacturing now in Q4 2025. Apple picking up 18A for a volume product in mid2027 is 18 months later. Minimally there be an ‘improved’ 18 variant by them or would be looking for second round of products to roll out onto mature process by then . By mis 2027 TSMc would be well on path to scaling up their A16 fab process. ( n2 design rules with backside power )
Last edited:
China have repeatedly said that they intend to take over Taiwan in the future - using military force if necessary. If that happens and Taiwan looks to fall to China, the west will destroy every single TSMC plant on their way out just to prevent China from controlling the world's chip manufacturing.
TSMC are currently the only company in the world with the ability to produce chips for Apple and nVidia. Neither company make their own chips - they just design them. If Apple doesn't have alternative suppliers, Apple ceases to exist.
So the options are to start moving TSMC to the US or another country like Australia (which will take a decade), or to work with someone like Intel who have processing capacity, but design terrible chips. There is nothing stopping Intel from hiring TSMC employees to help improve their processes.
This is a geopolitical problem, not a financial problem.
Makes sense for Apple to explore their options. Like how they keep hoping for BOE to be able to match Samsung's displays so that they aren't as reliant as Samsung.
The elephant in the room, or one of them, at least, is that the chance of there being a war in Taiwan in the next 10-20 years is not that low. Maybe less than fifty percent... but not comfortably near zero either. If you are planning for the next multiple decades, you have to take that into consideration.
The elephant in the room, or one of them, at least, is that the chance of there being a war in Taiwan in the next 10-20 years is not that low. Maybe less than fifty percent... but not comfortably near zero either. If you are planning for the next multiple decades, you have to take that into consideration.