Intel could return to Apple computers in 2027
(theverge.com)134 points by DamnInteresting 2 days ago
134 points by DamnInteresting 2 days ago
> chips built on their process node were materially worse
Would they be considered the same chip in the same design? Or same chip with slightly different design? Or different design altogether?
Genuinely curious because sometimes to support both Windows and Linux code would need to follow OS distinct paths in many locations even if the compiler is supposed to support cross arch, resulting in binaries that are supposed to be the same but behaves differently due to necessity. Is this the case here? Does TSMC and Samsung support different format of chip design equivalent of code?
Most likely: Same design, marketed as the same chip. Might see differences in power consumption and thermals but not in a way that will affect normal users. Apple has relatively few designs that they use across different products so that would further mask differences. If the Intel-fabbed variant of a chip ran 5% hotter but Apple put it into Apple TV then basically nobody would notice or care.
Yeah, I remember hearing NVidia did the same thing via Moore’s Law is Dead podcast. At this point it seems incredibly unlikely Intel will unseat TSMC anytime soon. TSMC has proven time and time again it is the only fab capable of producing leading edge nodes at the capacity and quality required by the likes of Apple, NVidia, and AMD. It also has substantially deeper pockets than Intel to continue to invest in staying number one.
I think if Intel is to stand a chance it’ll be via gaining momentum and market via “good enough” nodes and not cutting edge, essentially taking a page out of TSMCs playbook from the late 2000s and early 2010s. It needs more capital than it can raise, and time, both of which are hard to come by.
> TSMC has proven time and time again it is the only fab capable of producing leading edge nodes at the capacity and quality [...]. It also has substantially deeper pockets than Intel to continue to invest in staying number one.
Before about 2016, you could have said the same about Intel. They were generally considered process technology leaders. They were > a year ahead in shipping products with their latest 14nm node. Similarly their previous 22nm node. There had been several occasions over the previous decades where manufacturers stumbled, not as spectacularly as Intel's decade of malaise, but definitely nodes scrapped level.
So, things can change quite quickly. Intel's 18A node is likely to be "better" than TSMC's current N3x nodes (it is denser and better performing on paper) and will ship before N2, putting Intel momentarily in the lead for process technology again for a quarter or so, and it was first with some technologies like BSPD (TSMC won't do that until A16). Yields are a question, and N2 will be coming out which probably re-takes the lead... but this is quite a turnaround from late 2010s situation, right?
The big thing Intel needs is a working foundry pipeline, because there is so much money in high performance silicon that's not x86 these days. It has always been thought their CPU design teams were very close to fabrication which was thought to be something of an advantage for them. It's likely that has also made their process more difficult for outsiders. They've tried and failed several times to get this going and get external design wins, and just never done well even when their manufacturing was doing really well. Including this latest effort (https://overclock3d.net/news/misc/intel-may-cancel-its-18a-l...). Still, it's not impossible, and I'm sure TSMC considers this one of its biggest risks if Intel can boot a self-sustaining foundry business.
> Before about 2016, you could have said the same about Intel.
To be fair, Intel is finally making the major manufacturing equipment investments needed to catch up.
They previously did stock buybacks and acquisitions of other companies that went nowhere instead of investing in the EUV manufacturing equipment TSMC used. Now they have the more advanced version of EUV in production.
> Intel has reported processing over 30,000 wafers in a single quarter using High-NA EUV exposure, achieving simplified manufacturing by reducing the required steps for a particular layer from 40 to fewer than 10,
https://www.techpowerup.com/342239/intels-advanced-packaging...
Apple doesn't buy its manufacturing partners.
If you can make something with the quality Apple wants, they will buy you a manufacturing line in exchange for price, quality and production level guarantees.
Since Intel needs large customers for its Fab and capital investments, it would be a good deal for them even if the profit margin is much lower than they would prefer.
The most obvious thing would be Intel making security processor modules. Get the supply chain for those onshore, from the US point of view.
Doesn’t require the absolute latest processes.
I think this is a very astute comment.
It reminded me that for a while all SIM everywhere seemed to come from one european chip plant, although now I say it I wonder if they were just the assembly & packaging and fabrication was offshore?
In both cases (tpm and sim) the cynic would say it's only deciding which economy owns the back-door.
It was Gemplus. The backstory about how CIA and NSA got control over it is fascinating.
You are probably referring to NXP (formerly Philips) and Infineon (formerly Siemens), both of which have produced crypto processors, smartcards (including SIMs) and other secure elements for a really long time. Infineon is/was actually a really common supplier for the little 20-pin TPM/LPC modules.
Yes i think you're right. The nexus of sim, smart card and tpm seems strong. I e used thales and Luna (now also thales) HSM which are in hypothesis glamorous, but ultimately remarkably pedestrian secure devices. I wonder if they include logic from these companies. Supply chain behind FIPS120 class stuff would be an interesting story.
What, in your mind, is a "security processor module"? As far as I'm aware, there is no such entity in Apple systems; security functionality is on the same die as the CPU/GPU. (Which is a good thing; it means that communications between the CPU and that security processor cannot easily be intercepted or interfered with.)
Apple has more than enough resources to sample check chip deliveries for being manipulated.
The funniest outcome would be Apple throwing so much money at Intel Foundry that they end up monopolizing the leading-edge nodes, like they do at TSMC, leaving the rest of Intel to fight for scraps on their own production lines. I guess Intel also uses TSMC now but... yeah, as mentioned.
Both are objectively true, though. IFS would finally stand on it's own legs with a customer at Apple's scale, and Intel has the required IP and know-how to provide a stopgap RISC chip to embedded and datacenter customers that Apple usually ignores.
The "nightmare scenario" of Apple buying out the entirety of 14A to fabricate ARM chips is more-or-less what Pat Gelsinger spent his tenure trying to arrange.
Why would Intel design and make a RISC-V chips? Fundamentally they don't have any inherent advantages over x86 or ARM whatsoever(for datacenter atleast). It makes no sense for Intel to make embedded chips on 14A or non several generations old process either (margins on them are pitiful anyway and won't sustain Intel R&D spending).
Also x86 provides a huge moat to Intel/AMD which allows them to charge much higher prices than if they had to compete with anyone else.
Yes, and to put this in perspective: TSMC is valued around 8x higher than Intel at the moment. If Intel could become a major competitor to TSMC, I don't think they'd worry about Apple monopolizing leading edge nodes.
If Intel becomes the leading foundry, even if their x86 chips are a little behind Apple, they'll still be ahead of AMD. Apple start shipping 3nm back in 2023. It's looking like AMD will get there in another year. If Intel becomes the leading foundry and they're 12-18 months behind Apple, that'll still put them 18-24 months ahead of AMD.
Plus, it's important to think about the symbiotic relationship between TSMC and Apple. Apple can commit to large orders which gives TSMC the ability to invest. If Intel can get that business away from AMD, it means that TSMC won't have the same ability to push the envelope. Without Apple to pay top dollar for early access, will TSMC have the ROI necessary to keep moving as fast as they have been?
I don't think Intel would be concerned about Apple getting the latest Intel Foundry nodes before x86 does. It'd be a win for investors and ultimately a win for their x86 chips too. TSMC has benefitted from being able to invest in improvements and have Apple pay top dollar for it. If TSMC loses that, it also means that AMD, Nvidia, Qualcomm, and other Intel competitors lose the ability to ride the Apple-TSMC coattails.
>> If Intel becomes the leading foundry, even if their x86 chips are a little behind Apple, they'll still be ahead of AMD. Apple start shipping 3nm back in 2023. It's looking like AMD will get there in another year.
No? AMD is beating Intel in power and performance. It's true they will only reach 3nm for desktop next year with Zen 6, but they're beating Intel which is already at a smaller node. In essence AMD is lagging on process because they can. They're being very strategic while Intel is struggling to catch up. Zen 7 is going to be my next build, and it may be my last x86.
I trust TSMC more than Intel, and I can't help but wonder if this is related https://news.ycombinator.com/item?id=46045236
If one executive hire was all it took to steal a whole process node, TSMC could never have built the lead it has (nor Intel the lead it had in previous decades).
Semiconductor manufacturing is an exercise in blood, tears and careful note taking, not Magic Secrets.
Semi-clickbait headline: as others have pointed out, this is just about who fabs the Apple-designed chips, not a contemplated return to x86. Still pretty interesting if true; would be cool to see Apple diversify from TSMC. I'm rooting for Intel!
Yeah, this is pure clickbait. And it's barely even new "news" anyway. Apple has been actively testing Intel's 18A process, and has been in talks with Samsung as well.
There are only three advanced chip manufacturers in the world (TSMC, Samsung, and Intel), so of course Apple has been in talks with all of them.
The whole story is based on a tweet from an analyst anyway. "Company that designs its own chips might use one of the three chip manufacturers in the world, based on my hunch." hmmmmmmmmm... Sounds more like someone just loaded up on Intel calls to me. (Not that the reality of the situation isn't real and obvious.)
Returning as a US on shore manufacturer of Apple designed chips, and apparently not the leading edge ones. This feels like making Trump happy while Apple keeps full control.
> don’t see it as a bad thing for Intel
Isn't this a ringing success for their strategy of separating chip design from fabrication?
Maybe, it's unclear at the moment.
Apple is known to be one of the kings of putting their suppliers over a barrel. There's a good chance this is mainly a move to negotiate a better deal with TSMC, and even if it's not, the chance that Intel gets a boat load of profit out of it is very small.
And historically when fabs have been separated from a business, it's always been in a way to shed a capital intensive albatross. In that case, they're normally loaded up with so much debt in the divorce that they were essentially never intended to succeed or continue to keep up, but instead just barely stay afloat on the already capitalized investment.
> the chance that Intel gets a boat load of profit out of it is very small
Why? TSMC seems to be doing ok. It’s worked with RAM and SSD suppliers the same way and they seem to be doing ok too. So does Foxconn. Apple has been known to subsidise leading edge nodes in exchange for priority or temporary exclusivity, and is absolutely ruthless, but it does not prevent its partners from being successful.
> And historically when fabs have been separated from a business, it's always been in a way to shed a capital intensive albatross
That is true. But there are other factors that might be worth considering. First, Apple hates being dependent on a single supplier (which is a single point of failure). Then, hedging risks related to the security situation in Taiwan makes sense. Whether it means subsidising a new TSMC plant in the West or subsidising a new Intel plant might not be that huge a difference. Finally, it might apply some gentle and friendly pressure on TSMC by threatening to shift some production to a competitor.
Whether all this makes sense or not depends on quantitative and qualitative analysis based on data we don’t really have.
I'm kinda surprised this deal seems to be 18A still. While progress, 14A is what really matters, as Intel has more or less been threatening to just give up if they don't get a large 14A commitment(unless that's changed recently).
Though, if this goes well, it stands to reason Apple may be that needed commitment.
This might make Trump happy, but more importantly it makes the military happy. No military leader with any intelligence is happy with so much critical for war supply chain being in China or Taiwan - the two are showing signs of going to war with each other in the future. (it may not happen, but there are enough signs military and political leaders should worry). Military leaders are not just the US, leaders in Europe are happy too - they may not trust the US but the US is still at least a second option (and Trump is 79: he is statistically unlikely to live much longer)
Those in the federal government that spy on American citizens are also happy that they are getting to leverage the $11.1 billion of taxpayer dollars they gave to Intel.
If Intel make the CPUs in the USA are they going to be shipped to China for final assembly?
Packaging is something Intel does very well and they have facilities is the USA. The rumor is many companies might do final assembly with Intel.
https://www.eetimes.com/intels-embarrassment-of-riches-advan...
Intel opened a packaging facility in New Mexico in January that’s supposed to their largest.
https://newsroom.intel.com/intel-foundry/updates-intel-10-la...
Why not? A single Boeing 747-8F could carry 10M-50M chips in a single trip.
There have already been heists of trucks carrying GPU's, I could see something like this happening.
The value of that airplane would be astronomical. I would split it up into dozens of flights just to reduce risk if one of them had a mechanical problem.
Dear leader bought Intel, Tim Apple needs to appease dear leader to avoid tariffs. 2+2
Well I sure hope not. I'll never buy a mac with an Intel processor again
Perhaps the headline should have been changed when the post was made here.
This link has much more information:
> Apple has used version 0.9.1 of process design kit (PDK) designed for Intel 18AP node. With performance, density, power, and every other metric going according to plan, Intel could become Apple's source of advanced node production in 2027... The 18A-P node enhances Intel's 18A by incorporating RibbonFET and PowerVia technologies, which offer better performance and energy efficiency. Compared to the regular 18A node, these improvements include newly designed low-threshold voltage components, optimized elements to reduce leakage, and refined ribbon width specifications, all aimed at boosting performance-per-watt metrics.
https://www.techpowerup.com/343423/intel-could-manufacture-a...
Remember that Apple previously dual sourced SOCs from both TSMC and Samsung before dropping Samsung when they fell behind and chips built on their process node were materially worse.
This is trial production, not a done deal. Intel has to deliver on their promises.
The good news for Intel is that Apple has a long history of paying up front for dedicated manufacturing lines once a manufacturing partner proves that they can hit Apple's QC metrics and price point.