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
TDP has nothing to do with the architecture you're using. The reason ARM chips don't target similar TDPs is because they don't have too. That's just RISC vs CISC.
They could. It's just not a market segment they're very interested in. And even without it, they beat most AMD+Intel chips anyway.
However, ARM is proving to be quite satisfactory for many non-demanding tasks, and no doubt will get better as technology improves. Windows and Linux already run on ARM, and mush of the challenge is not, IMHO, processor power but willingness to devote resources to making the OS run better.
That's changing now. Apple won't be the only one to get N2 at launch, and A16 (TSMC 1.6nm node) is HPC-focused and Apple definitely won't get it first (rumors say NVIDIA) and I wouldn't be surprised if they delay it until a mobile-focused version comes out.
That's why I think this deal makes sense for Apple even aside from the political considerations. HPC revenue at TSMC recently surpassed smartphones. TSMC has jacked up prices- including on Apple. Apple and TSMC are going from a symbiotic relationship where TSMC needed Apple to fund the new node and test it out first, and Apple needed TSMC for manufacturing, to one where Apple is dependent on TSMC but TSMC doesn't need Apple in the same way. TSMC is raising prices because of the AI data center boom, but Apple can't afford to get into a bidding war for wafers with AI chip companies that have chips that sell for thousands each.
Apple despises supplier relationships where Apple is not the one with the leverage and generally tries to dual-source so it can play suppliers against each other. If the second supplier isn't up to par, Apple will often work with them for years to try and make them viable (LG Display, BOE, Luxshare, etc.). It makes sense to try to do this with logic chips too. Intel will obviously give Apple a sweetheart deal, while Apple can use the existence of a credible second supplier to stop TSMC from raising prices on it again.
Last edited:
You’re not very aware of the geopolitical problem if you’re asking this question.
I can easily imagine a world where TSMC is literally blown out of the water and no longer exists as a viable fab, so if you want ANY chips you have to go with what still exists.
If China makes a move on Taiwan (which they say they're going to, and 2027 is the target year), TSMC is going away regardless of the outcome. If China is losing they'll blow it up so the west can't have it. If China is winning, the west will blow it up so China can't have it. TSMC does not survive a Chinese attack on Taiwan either way.
Ok THAT is very believable. Apple will still be getting TSMC chips, just contracted through Intel. Of course, it remains to be seen if Apple can avoid screwing THAT up.I think this has to be about M7. Apple doesn’t yet have the final version of Intel 18AP to work with, so whatever designs they are planning, they won’t be complete for months.
Also, keep in mind that 30% of Intel’s own chips are manufactured by TSMC, a partnership that is extending to N2:
Intel, AMD, and MediaTek reportedly among TSMC's 2nm early adopters — company said to have 15 customers lined up for new process tech
An unprecedented number.www.tomshardware.com
So there’s something of a natural progression or convergence here, and Apple using both manufacturers isn’t strange at all.
Mediatek IS eating their lunch. Apple leaving Qualcomm is very likely doing more damage than Apple leaving Intel because Mediatek is not slowing down.
I just don't agree with this take. TSMC exists outside of Taiwan. It will obviously be a major disruption when China invades Taiwan, but TSMC will still operate in some capacity, and just because they're invaded doesn't mean they're not going to be the premier chip fab anymore. China can blow up every fab in Taiwan, but other TSMC fabs exist outside of the country. It would take a mass civilian casualty event to wipe out the braintrust that makes TSMC as good as they are.
Sure, it does, but...
TSMC, unlike many multinational corporations, remains deeply Taiwanese at its core, with operations and talent overwhelmingly concentrated on the island. Of TSMC’s 83,000+ global employees, close to 90% are Taiwanese, including 88% of management-level staff. 87% of employees work in Taiwan. And, over 90% of TSMC’s total capacity is in Taiwan.
Of course, there's one fab in Arizona, but ... sometime ago,
-- Dr. Morris Chang
-- Dr. Cliff Hou
Apologies if this is sarcasm that has gone over my head, but that is not what I was talking about. Intel uses TSMC’s foundries the same way Apple does. It’s just that Intel also has its own foundries, which Apple can also utilize.
Intel Foundry Services (now Intel Foundry — that link has a roadmap, Apple is looking at using 18AP) launched in 2021. It is a division of Intel, but it is a separate entity.
You’ve taken this quote completely out of context. He is talking about TSMC’s first attempt to operate a US foundry in 1995 (Wafertech, in Washington state, near Portland, Oregon). He is not talking about their experience 25 years later in Arizona.
Last edited:
You should have noted the part "...some time ago" in my post, as it gives the general idea of the risks involved in doing fabrication in the US. He is still against that.
TSMC's consideration of overseas fabs is primarily driven by customer demand. However, when the "customers" are foreign governments, it becomes problematic. In certain countries, governments don't remain in power for more than four years, and the next government might overturn all the decisions made by the previous one. They usually do, don't they?
The whole of the digital economy, automotive industry, and industrial production relies on chips from TSMC. The world economy would grind to a halt without TSMC. All top corporations in the Western world would practically fail immediately...and governments.
Last edited:
Industrial production doesn't need cutting edge chips-- they could go to less advanced fabs.
Similarly, cars could go to less advanced fabs-- they'd just have to go back to being more mechanical devices and less large, rolling smart phones. As in, the ECU in my car is made using a 65nm process, and there's not really any room for improvement in how it handles engine management.
It's all the computers, phones, watches, VR, consoles, and especially AI that would really suffer. Beyond losing TSMC, there's a vast network of suppliers to TSMC that could go away.
If Taiwan gets attacked, I could see a solid decade+ of the digital devices you own going in being as good as it gets-- and AI ending.
And that's assuming a best case scenario, when it doesn't bring about the end of globalization
Good luck!Industrial production doesn't need cutting edge chips-- they could go to less advanced fabs.
Similarly, cars could go to less advanced fabs-- they'd just have to go back to being more mechanical devices and less large, rolling smart phones. As in, the ECU in my car is made using a 65nm process, and there's not really any room for improvement in how it handles engine management.
It's all the computers, phones, watches, VR, consoles, and especially AI that would really suffer. Beyond losing TSMC, there's a vast network of suppliers to TSMC that could go away.
If Taiwan gets attacked, I could see a solid decade+ of the digital devices you own going in being as good as it gets-- and AI ending.
And that's assuming a best case scenario, when it doesn't bring about the end of globalization
There are anticipated to be 3nm chips in the US by 2027 or 2028, but the critical question is whether there will be a reliable supply chain, and skilled professionals in semiconductor manufacturing. The stability of the supply chain may be influenced by international relations, particularly with key suppliers located in Asia.
No, he isn’t against it. Here are some more quotes from the same speech in December 2022:
“25 years passed and we have a new chairman, Mark Liu. And he happens to share my dream! Now you see a partial, not yet fulfilled, but progress of this dream …”
“We did learn from our experience earlier, and we are far more prepared now. And we hired almost 600 engineers here a year and a half ago, we sent them to Taiwan, and they were under training in Taiwan for one year to a year and a half. In the meantime, about the same number of Taiwan engineers underwent training in Taiwan also. So before we see a single wafer, we have more than a thousand people being trained. This I think is a very good sign that we are prepared. It’s a very good sign that my dream of 25 years ago will now be fulfilled …”
Morris Chang is a naturalized US citizen who identifies strongly as an American, who took a job in Taiwan in 1985 and then started TSMC in 1987 with government backing just as Taiwan was emerging from more than 40 years of military dictatorship (martial law was lifted in 1987).
He has also made it clear that he doesn’t think China is likely to invade Taiwan, and if they were to do so, the world would have bigger problems than just semiconductors.
Unless you have a role inside the upper echelons of the Chinese Communist Party today, you don’t know either.
It’s not just semiconductors. Something like 90% of shipping for heavy industry goes through the international waters in the Taiwan Strait. Shutting that conduit down for more than a few days would be catastrophic.
I’m not really disagreeing about TSMC’s importance, although I think you are overlooking the critical role of the ASML high-resolution lithography equipment TSMC relies upon.
I’m only disputing the assumption that you and a few others seem to be making, which is that an invasion of Taiwan is inevitable.
Nobody is more acutely aware of China’s problems with corruption in the military than Xi Jinping. Despite all the propaganda and posturing, it’s very, very unlikely he is deluded enough to believe it would be easy for China to take Taiwan or even blockade it for any extended period of time.
Last edited:
Sure, it was his dream, but he also knows that the fab in the US can't produce wafers at the price it can be done in Taiwan, but may cost at least 150%.
So before we see a single wafer, we have more than a thousand people being trained. In other words, the futility of the venture -- wasted resources on 600 additional engineers.