Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Almost possible today depending on which Apple Silicon Mac you have.

https://asahilinux.org/ - "production ready" (ish)
https://github.com/AppleWOA - decidedly less than production ready hehe (as in, no, you can't really use it today, but for some reason someone somewhere is actually dedicating time to attempt to make it happen)

Isn't Asahi Linux only have the basic working up to only M3-series?

 
TSMC sells chips. Apple sells finished goods.
TSMC manufactures chips. They don’t sell chips. The difference you’re describing (selling chips versus selling finished products) is the difference between Apple/Nvidia versus AMD/Intel, all four of whom are TSMC customers. Apple and Nvidia are *not* merchant silicon vendors. Apple sells consumer products and Nvidia sells business products, consumer GPUs, and DGX systems. AMD does a bit of both, but they also sell chips, unlike Nvidia.
 
Isn't Asahi Linux only have the basic working up to only M3-series?

If by the basics you mean everything except Thunderbolt, USB4/DP-alt-mode (very close to ready!), TPM (Secure Enclave Processor), and TouchID, then yeah, you're right.

For M1 and M2 series machines it's more or less ready to go minus the above features, with most things maintained upstream even!

All that aside, Thunderbolt being missing is actually a big blocker for a lot of people, myself included, but there's been rumblings that that will be the focus area after USB4/DP-alt-mode is done (my own analysis of the situation), which is as I understand it nearing completion!
 
  • Like
Reactions: hajime
It's not a bubble, more like a deflating tire. With everyone building their own AI chips, AMAZON, Microsoft, Google, Apple, etc to get out from under Nvidia, the air is being let out slowly. But make no mistake, Nvidia's days are numbered. Their moat is leaking.
 
If by the basics you mean everything except Thunderbolt, USB4/DP-alt-mode (very close to ready!), TPM (Secure Enclave Processor), and TouchID, then yeah, you're right.

For M1 and M2 series machines it's more or less ready to go minus the above features, with most things maintained upstream even!

All that aside, Thunderbolt being missing is actually a big blocker for a lot of people, myself included, but there's been rumblings that that will be the focus area after USB4/DP-alt-mode is done (my own analysis of the situation), which is as I understand it nearing completion!

Is running Linux under Parallels or vmware better?

Since there is still no M5 Pro MacBook Pro release and running Windows and Linux natively may be faster than under Parallels, maybe time to leave TC’s waiting game and look for Windows laptop.
 
Bottom line, this can work two ways....

1. Bad: to the extent that Apple has to compete for TSMC capacity, it can negatively effect Apple's chip supply chain in terms of availability and price of the chips they buy.

2. Good: Nvidia is obviously benefiting from past TSMC's gigantic investments in R&D, which is 50% of TSMC's operating expenses. That past R&D was funded by previous customers, particuarly Apple. Going forward, to the extent more companies like Nvidia, as TSMC customers, are providing more funds, Apple will benefit.
 
This isn't sports.
This really doesn't matter beyond supply chains for Apple. Something that can be addressed multiple ways.

Supply chains aren't really the thing that Apple has shown it is bad at under Cook leadership.
agreed, except for the fact that the increased competition means higher cost which will be translated to the customer.
Also securing future smaller nodes will be more challenging meaning chip progression might be a bit slower until supply is present.
 
This is Apple speaking: We have an open position you might be interested in at our Marketing department.
If you read closely, you'll hopefully realize that I was dissing Apple. Where I disagree with you is that Apple improving their quality would significantly increase their TMSC market share. I cetainly do wish Apple products (especially software) were higher quality, but I think that such a focus would rather stand in the way of their success and profit. I wish they made better products even at the cost of less profit. But that's not how they operate.
 
Last edited:
One key factor behind Nvidia's rising share of TSMC revenue is the nature of the chips it needs. AI accelerators are significantly larger, more complex, and more expensive to manufacture than Apple's A- or M-series chips. They often require leading-edge process nodes, advanced packaging techniques, and higher wafer costs, all of which translate into higher revenue per chip for TSMC. While Apple ships far higher volumes of processors overall, it requires smaller system-on-a-chip designs optimized for power efficiency and consumer devices, resulting in lower manufacturing costs per unit.

TSMC's growing reliance on AI customers could have direct implications for Apple. While it remains one of the foundry's most important customers, it is no longer the primary driver of TSMC's capacity expansion or capital expenditure decisions. Analysts say that Nvidia has effectively taken Apple's place as the scale customer that helps guide development and justify increased investment in each new leading-edge process node.
There is one flaw in this analysis — Nvidia does not use a new TSMC process node every year, unlike Apple. Grace Hopper, Blackwell, and Blackwell Ultra are built on two custom 5nm nodes (4N and 4NP). Vera Rubin will be on third-generation 3nm N3P (skipping over N3 and N3E). Nvidia probably [?] won’t use N2, so Apple’s A20/M6 production will still lead the way on 2nm, just like they did on 3nm.

Nvidia could cause a delay in Apple’s A21/M7, but there are two possible 2nm nodes for that. My guess would be Apple uses N2P, Nvidia uses A16, and everybody’s happy. Or the other way around.

Still, beyond Nvidia, there plenty of other AI heavyweights looking for a piece of TSMC’s pie, so the overall point of the article is still valid. AI silicon is changing Apple’s relationship with TSMC.
 
Relying on a foreign company manufacturing overpriced products that is at risk of losing it’s core facilities and personnel is risky at best and at worst probably stupid, not to mention not being in line with the the current ‘American First!’, ‘Made in the USA!’, Government backed Trillion dollar wave of industrialization occurring. Any American Company wanting to be a part of that wave should not rely foreign manufacturers.
 
Nvidia is selling more chips than the world has data centers to house, and there’s increasing resistance against building more. The sales volume is not sustainable.
They’re not building just for Sam Altman’s and Elon Musk’s gratification.

They’re building for total population control through predictive data by corporations ( Amazon ) and governments along with their militaries. Whose workload are increasingly being warehoused in Amazon’s AWS along with Palantir and Anduril.

Pre-cog levels of data awareness on every human being AND social credit scoring for financial and public services are the end game. There is a LOT of hardware and chip sales left in this end-game run up for Silicon Tyranny.
 
Relying on a foreign company manufacturing overpriced products that is at risk of losing it’s core facilities and personnel is risky at best and at worst probably stupid, not to mention not being in line with the the current ‘American First!’, ‘Made in the USA!’, Government backed Trillion dollar wave of industrialization occurring. Any American Company wanting to be a part of that wave should not rely foreign manufacturers.
You’re very naive. This tiny spit of land called the United States of America does not have the natural resources to go full independent from the world. America First is a marketing slogan and you fell for it.

Just like the label on your jug of laundry detergent that screams “NEW AND IMPROVED” only until you see additional tiny print below that says…

….”packaging”.
 
Last edited:
Blew my mind that Nvidia was worth $4.6T
It’s the speed in which they did it as opposed to Apple or even Microsoft.

But it does show the genius of Jensen Huang’s long range vision of making a software and framework eco-system purposed built for your hardware (CUDA) and getting that into colleges’s and universities’ Comp-Sci programs so that you have an already trained workforce before they’re even hired.

This is why AMD and particularly Intel will never catch up to Nvidia. CUDA is the de-facto and default framework for AI. And because of that Nvidia hardware sales naturally follow.
 
Apple hasn't "lost" anything, they have longstanding supply agreements and TSMC will fulfill those. Nvidia having a larger share means that other customers of TSMC will face issues ...
And given the rumor that Apple is looking at Intel 18/14A, TSMC will be looking at alternatives to keep the fabs full.
Normal business ...
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.