TheYayAreaLiving 🎗️
Suspended
I really hope MacBook Air 15" is coming with a 3nm Chip.
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 Orders Entire Supply of TSMC's 3nm Chips for iPhone 15 Pro and M3 Macs
- Thread starter MacRumors
- Start date
- Sort by reaction score
Your take is wrong. N3 has been in full-scale production since the end of 2022, as has been publicly announced by TSMC. There are no announced customers for N3 chips, and of all the possible customers capable of purchasing all those wafers, there's only one that keeps this sort of activity secret: Apple.
Obviously not. The A17, I think you're likely right. The other two, not so much.
I think 3nm will deliver a lot less than people think! Even TSMC only claim 15% speed improvement at the same power output and to get 30% efficiency improvement, it will have to run at the same speeds as the 5nm.....people just don't read the facts 😏
Apple has reportedly secured all available orders for N3, TSMC's first-generation 3-nanometer process that is likely to be used in the upcoming iPhone 15 Pro lineup as well as new MacBooks scheduled for launch in the second half of 2023.
According to a paywalled DigiTimes report, Apple has procured 100% of the initial N3 supply, which is said to have a high yield, despite the higher costs involved and the decline in the foundry's utilization rate in the first half of 2023. Mass production of TSMC's 3nm process began in late December, and the foundry has scaled up process capacity at a gradual pace with monthly output set to reach 45,000 wafers in March, according to the report's sources.
Apple is widely expected to adopt TSMC's 3nm technology this year for the A17 Bionic chip likely to power the iPhone 15 Pro and iPhone 15 Pro Max models. The 3nm technology is said to deliver a 35% power efficiency improvement over 4nm, which was used to make the A16 Bionic chip for the iPhone 14 Pro and Pro Max.
The latter two iPhone models were the first smartphones to feature chips built on the 4nm process, and it looks like Apple is again attempting to be first to market with models based on the latest cutting-edge semiconductor technology.
Apple plans to release a new MacBook Air in the second half of 2023, and it may be equipped with a 3nm chip, according to a January report from DigiTimes. However, display industry analyst Ross Young in December claimed that a 15-inch MacBook Air would be released in the first half of 2023. If DigiTimes' outlook turns out to be accurate, then perhaps both 13-inch and 15-inch MacBook Airs with M3 chips based on 3nm technology will launch in the second half of 2023 instead.
Looking further ahead, Apple analyst Ming-Chi Kuo believes 14 and 16-inch MacBook Pros coming in 2024 will feature M3 Pro and M3 Max chips that are built on TSMC's 3nm process. MacBook Pro models with the M3 Pro and M3 Max chips will go into mass production in the first half of 2024, according to Kuo.
The 3nm technology will offer improved performance and better power efficiency compared to the current chips manufactured on a 5-nanometer process, including the M2 Pro found in Apple's current high-end Mac mini and the M2 Pro and M2 Max used in its latest 14 and 16-inch MacBook Pro models.
TSMC is poised to move to N3E – an enhanced version of N3, its first-generation 3nm technology – to commercial production in the second half of this year, and Apple will be the first customer to adopt the process, according to another report this week from DigiTimes. Nikkei Asia reported in September that Apple could adopt N3E for devices launching as soon as this year, but we've not seen any other reports corroborating this roadmap.
Article Link: Apple Orders Entire Supply of TSMC's 3nm Chips for iPhone 15 Pro and M3 Macs
You are looking at benchmarks. I was talking about efficiency. You can look at various tests of efficiency, and you can see how A14/15 was way way ahead of SD888/SD8gen1. Apple was gaining significant efficiency on every generation prior to the A16 while Qualcomm literally was just sidestepping from SD865 onward, until SD8+gen1.
If Qualcomm can catch up/overtake Apple just by using the same fab as Apple, aka TSMC, that means Apple's core design is not up to snuff.
TSMC: How much would you like to buy?
Apple: Yes
Apple: Yes
To me, the key wording in this post is "100%".
This implies that Apple wasn't planning on eating all the cookies. Something changed, and now they are.
TSMC had Intel bail on 3nm (after they paid a premium for it), and they had 3nm capacity they wanted to fill. Articles in the past noted TSMC wanting other companies to hop on 3nm (N3B, if not mistaken - a more difficult node than N3E will be) but couldn't find them. That would mean they fell back to Apple as plan B - probably sharing some savings with Apple to entice them to buy more than they originally asked for.
That means more 3nm products, sooner! Savings won't be passed on to customers - maybe bigger sales incentives over the holidays.
Unleash the 12" Macbook.
This implies that Apple wasn't planning on eating all the cookies. Something changed, and now they are.
TSMC had Intel bail on 3nm (after they paid a premium for it), and they had 3nm capacity they wanted to fill. Articles in the past noted TSMC wanting other companies to hop on 3nm (N3B, if not mistaken - a more difficult node than N3E will be) but couldn't find them. That would mean they fell back to Apple as plan B - probably sharing some savings with Apple to entice them to buy more than they originally asked for.
That means more 3nm products, sooner! Savings won't be passed on to customers - maybe bigger sales incentives over the holidays.
Unleash the 12" Macbook.
Oh I think it’s pretty safe to assume the customer will absorb ‘all’ the cost. Just means I best get a Mac Studio in the next year as it’ll be a bargain compared to what I ‘think’ it’s price will go up to With an M3.
People really don't appreciate how screwed up Apple was by TSMC's N3 delay. The A16 GPU core is basically the A15 GPU core. The original GPU core for the A16 had to be put on hold because it was for N3, and would not fit in a N5/N4 chip. Expect *two* generations of advance in the A17 GPU, in addition to the benefits of N3. It's going to be a major step forwards.
Note also that the big advantage of N3 is area, not performance, which plays to GPU strengths. I would expect the A17 and the M3 to both feature more GPU cores than previous chips, which will significantly enhance GPU performance beyond the individual core performance increases. The interesting question is whether the low-end chips (A17/M3) will also have more CPU cores, and what kinds? The problem with doing that is that while the cores shrink a lot with the N3 process, the L2$ does not, as SRAM scales down almost not at all (~5%) with N3, and not at all with N3E. You don't want to add CPU cores without the cache, obviously. I'm guessing not for the A17 (maybe 2 more E cores?), and yes for the M3, but that's a low confidence guess.
BTW, as someone else posted, Apple never had a 5 year lead. It was maybe 2-2.5 years, which was still monstrous. It might be down to 1 year now. Or not. Wait for the A17/M3, then we'll see. Apple's in the worst position they've been in since before the A7, I think, with the A16/M2, but that should all be reversed with the next chips.
Last edited:
Possible, but I think not. Do you have expected wafer starts/month numbers for N3B and N3E? That would be interesting info, and would point us in the right direction, though you still wouldn't know for sure until you had a pretty good idea of how big each A17 was going to be.
You're completely missing the 70% area improvement. Apple is likely to use that to great advantage, in more GPU cores, probably more CPU and NPU cores, etc. That's going to have a big impact.
This is the wrong take. In what world would a customer pay a premium to buy more goods than they want from a vendor? That's like saying you can buy one iPod for $100, or 2 for $220.
Apple's M3 supply is not getting more expensive, it's getting cheaper.
Last edited:
My guess is that Apple has already prototyped the A17 SoC and may have started to to prototype the M3 Series SoC. It'll be interesting to see if Samsung can get their 3 nm fabs up and going soon so the Qualcomm designed Snapdragon SoC and radio modem chips will be manufactured by Samsung.
As it should be. What you see is what you pay. Never understood why NA is lagging behind on this. Drives me nuts
If I'm reading it right, N3 will give a bit of a bump to efficiency per watt over N5. N3E, the generation after benefits on higher power performance - Something I'm presuming we wouldn't want on a phone anyway. Leave that to laptops when warranted.
www.anandtech.com
AnandTech Forums: Technology, Hardware, Software, and Deals
Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
www.anandtech.com
N3E is not a generation after N3 (aka N3B). It's an offshoot of N3, and the same generation.If I'm reading it right, N3 will give a bit of a bump to efficiency per watt over N5. N3E, the generation after benefits on higher power performance - Something I'm presuming we wouldn't want on a phone anyway. Leave that to laptops when warranted.
AnandTech Forums: Technology, Hardware, Software, and Deals
Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
N3E is generally superior to N3. It is better in both efficiency and performance, though only by a little bit. But it is inferior in two important ways. The first is area - it has slightly relaxed rules, meaning that features are a bit larger. So depending on what you're building, it's a wash, or N3E is a bit better than N3. (If you were building pure SRAM, N3 would be slightly better, maybe, but who does that? Well, AMD, a little bit. It's a corner case.)
But.
The other advantage N3 has is that it's already in volume production. That means the first finished wafers are due soon, within the next month or so. If you want N3E, well, you can't have it yet! So if you want, say, an M3 Mac before the fall, you only have one choice. The iphone, on the other hand, might be a good target for N3E. (Or, maybe not. It depends on how many wafers TSMC can make, and how quickly, when they do ramp up N3E.)
BTW, your source is a good one. Anandtech's articles on this are pretty reliable. The numbers in their table provide specifics for what I wrote above.
You jest but it's also hilarious that a process like lets say TSMC N7, has a 52 nm fin height and 30 nm fin pitch. The marketing names are so hilariously divorced from what goes on inside. Ok, they're supposed to point to "minimum feature size" they say, but one feature of that length doesn't help you if the rest of the billions of transistors on the chip are nothing like that.
So contrary to a growing popular belief, we have many dimensions to shrink in before we're at the true limits of silicon, as the name is far removed from the 3d picture.
I was hoping that Apple would enter the server market with the M-chips but it seems the volume just isn't available to make this happen yet.
You mean, until the software stops supporting it. We already could be using computer hardware much longer if not for lack of software support.