• Guest, The rules for the P & N subforum have been updated to prohibit "ad hominem" or personal attacks against other posters. See the full details in the post "Politics and News Rules & Guidelines."

News [CNBC] SoftBank hires Goldman Sachs to explore sale options for chip designer Arm. . .

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Mar 11, 2004
20,312
2,461
126
Huawei. Hah!
Wouldn't surprise me, but I'd also guess it wouldn't be direct.

I think an acquisition by some Chinese state owned enterprise or local governments would be interested but that could be blocked by the Japanese government and it will definitely be blocked by the American government but that won't be an issue with China's regulators ...

I wonder how Japanese regulators would react if America expressed opposition to this ...
Certainly they would.

I would imagine Japan has plenty of their own reservations about such, so they'd be more likely to be on the same page.

The whole point, and value, of ARM lies in its ability to be licensed cheaply to a vast variety of markets.

It would be very strange if it got sold to someone who reduced that - and would probably lead to a replacement being developed in reasonably fast order.
That wouldn't have to stop just because someone, even someone that has a vested interest in their own ARM designs (say Apple for instance), buys them. Which I think that might could be part of the impetus behind an offer. Someone like Apple who is becoming more dependent on ARM views it as a calculated move to make sure they keep access to the IP. They can let ARM operate largely on its own, doing design/licensing, while Apple gets first dibs at new feature sets/etc and instead of paying licensing they're profiting from it.

"Reasonably fast" would not be fast enough to avoid massive damage to the entire semiconductor industry. There's entire industries built around ARM-compatible IP, and massive amounts of effort invested into ARM-compatible software stacks. Starting over from scratch would be a huge blow.

Maybe one of the existing players like MIPS or RISC-V could step into the gap... but there would probably be a decade of chaos before it all shook out.
I'd think if China/Chinese companies were considering that they'd know its not gonna go over well so they'd likely agree to keep licensing setup in place. Which I think just about anyone buying them might get a lot of companies asking for at least that to be mandated to prevent monopoly control.

There were rumours back in 2010 that Apple was interested in buying Arm. Nothing came of that though, and it may have been idle speculation from finance types (just because Apple had a ton of cash on hand).

I think it could be just about anyone.

In just chipmakers, I'd say it'd be like this:
Nvidia (could really use a CPU so they don't have to rely on competitors' products, ARM is starting to take off in server/enterprise, and is dominant in mobile)
Intel (couldn't beat them so might as well buy them, maybe is looking to limit ARM to mobile; maybe even heding their bets in case x86 proves problematic - if nothing else it provides them easy means to produce at other foundries should they keep having issues with their own)
Samsung (just seems like a Samsung thing to do, and there were rumors they thought about buying AMD and ARM would be a bigger get for them I think)

Apple (I just don't think they'd want to own it, we saw how they are with the GPU with Imagination, but they might view it as integral to maintaining access/IP rights with them moving Macs towards ARM as well).

Qualcomm (considering they supposedly have been shopping their chip business I don't see this)
AMD (I think has good reasons to maybe want it, but don't think they're in financial position to make such a move; it would be bold, would make them not reliant on Intel's control of x86, instantly gives them mobile access, plus companies looking at ARM to deal with Intel's dominance could be partly limiting to them already and likely might increase; it should be a boon for their semi-custom business; also could be big if they require licensing of their GPU - while open sourcing the software as much as possible to placate concerns about lock-in, so kinda take the anti-Nvidia path where you let people license the hardware and then offer open source software stack)

Definitely there are others. Perhaps its Chinese backed companies (although I'd guess they'd do it less directly knowing it will be blocked, meaning the Chinese companies invest in company that then makes the bid), and perhaps that's why Softbank said they'd gotten offers (basically going "hey someone else bid more so we can turn them down keep out of that mess").

I could also see Amazon (just rolled out their own ARM design) or maybe Facebook or someone. I wouldn't think Microsoft or Google but couldn't rule them out entirely either. Other players might view it as a way back to success (IBM, maybe they look to go from Power to ARM; Dell or maybe HP to similarly try to revolutionize their product portfolio). And probably other holding companies and investment groups, etc. I could even see it being a shadow group, where maybe the NSA, Israeli defense, or someone wants deep level access or something.

Heck, I wouldn't rule out Arm & Hammer.
 
  • Like
Reactions: amd6502 and Tlh97

Qwertilot

Golden Member
Nov 28, 2013
1,475
162
106
"Reasonably fast" would not be fast enough to avoid massive damage to the entire semiconductor industry. There's entire industries built around ARM-compatible IP, and massive amounts of effort invested into ARM-compatible software stacks. Starting over from scratch would be a huge blow.

Maybe one of the existing players like MIPS or RISC-V could step into the gap... but there would probably be a decade of chaos before it all shook out.
True. Quite hard to see the value in a purchase that would cause that though, unless the goal is simply to spend money in order to watch the world burn.

Alternatively, if someone did buy it and started (seeming to) favour themselves, you'd get the alternatives strengthening over time fairly quickly.
 

DrMrLordX

Lifer
Apr 27, 2000
16,371
5,282
136
I'd think if China/Chinese companies were considering that they'd know its not gonna go over well so they'd likely agree to keep licensing setup in place. Which I think just about anyone buying them might get a lot of companies asking for at least that to be mandated to prevent monopoly control.
If you were Qualcomm, or Samsung, or Altera, or Amazon, or Fujitsu, or Marvell/Broadcom, or anyone else licensing ARM tech for big projects/widely-sold products, would you want to be licensing 90-100% of your CPU tech from HiSilicon with no alternative vendor? Rememeber, Apple is the only company selling custom cores anymore. Huawei is apparently trying to go their own way with the tech they already have licensed (see: Taishan v110), but if benchmarks of the 8c/8t Kunpeng 920 are any indicator, their custom cores aren't that good: read, below A76 despite being of the same generation. See below

True. Quite hard to see the value in a purchase that would cause that though, unless the goal is simply to spend money in order to watch the world burn.
The CCP has been laying out a lot of money on semiconductor design and research for years, with little to show for it. Their best company - Huawei's HiSilicon division - still has to license tech from a foreign company, and they are being cut off from foreign IP. The Taishan v110 is starting to look like kind of a dud (granted I have not seen the 64c Kunpeng 920 benched yet). If ARM Holdings is up for sale, it saves them a ton in future investment to just buy the gold standard in mobile computing horsepower today, even if the market value of the company plummets from foreign companies rapidly abandoning their designs - or attempting to, anyway. ARM is embedded in so many computing devices that it'll be hard to get away from it entirely.

Meanwhile, whoever buys ARM Holdings immediately has access to some top-flight CPU designs, including the stuff we haven't even seen yet. And substantial engineering resources to boot.

For the Chicoms, why continue producing a domestic "alternative" to foreign semicon IP when they can just buy one of the gold standards?
 
  • Like
Reactions: Tlh97

Doug S

Senior member
Feb 8, 2020
326
455
96
There were rumours back in 2010 that Apple was interested in buying Arm. Nothing came of that though, and it may have been idle speculation from finance types (just because Apple had a ton of cash on hand).

Apple would gain no advantage by owning ARM, they already have everything they want from them with the architecture license, and no doubt they already have a voice as far as the architecture's future directions.

It would also be an acquisition an order of magnitude larger than any they have ever made.
 
  • Like
Reactions: Tlh97 and amd6502

dullard

Elite Member
May 21, 2001
22,422
752
126
Apple would gain no advantage by owning ARM, they already have everything they want from them with the architecture license, and no doubt they already have a voice as far as the architecture's future directions.
Would Apple be able to just no longer provide licenses? What would Apple's competitors do without ARM? That is why I was asking about how long the current licenses last. It would be expensive and only a temporary blow, but Apple can afford it. Apple's lawyers are the best in the world, they win almost every case they are involved with.

I really think China wants it more than Apple though. China really, really wants to be independent from the west for processors (1). China has the money and buying ARM is a big way to that goal. They essentially already own VIA technology through a joint venture (2), have AMD Zen chip design (3), and could take control over TSMC.

(1) https://www.bbc.com/news/business-50287485
(2) https://en.wikipedia.org/wiki/Zhaoxin
(3) https://en.wikipedia.org/wiki/AMD–Chinese_joint_venture
 
Last edited:

soresu

Golden Member
Dec 19, 2014
1,420
591
136
I highly doubt that. I'm not saying it won't be broached, but I would be skeptical that it would be prevented. The US almost certainly wouldn't (Apple controlling a major tech IP would be viewed as a positive; and especially since it would give extra reason for them to block Chinese companies from using it). EU might try to block it, but I'm not even sure they would.
It is by definition anti competitive to acquire something that would essentially allow you to control a market you are currently competing in.

As the owners of ARM they could just refuse to develop new cores and that would leave those like Qualcomm, Samsung, Mediatek, Rockchip, Amlogic etc etc all in the dog house.

Allowing that level of anti competitive behaviour is basically like cutting the actual potential of a market and therefore the jobs created by said market - this much at the very least the US govmt would be interested in.

Apple could even refuse to allow uArch licenses if they owned ARM - it just isn't a win for anyone but Apple, and it isn't even much of a win for them, it's an expression of defeat implying that they can't compete without cutting off their competitors IP suppliers.
 
  • Like
Reactions: moinmoin

soresu

Golden Member
Dec 19, 2014
1,420
591
136
There were rumours back in 2010 that Apple was interested in buying Arm. Nothing came of that though, and it may have been idle speculation from finance types (just because Apple had a ton of cash on hand).

Exactly, if they were going to they would have back then when they were first investing in their custom cores after acquiring PA Semi and Keller with them.
 

Doug S

Senior member
Feb 8, 2020
326
455
96
It is by definition anti competitive to acquire something that would essentially allow you to control a market you are currently competing in.

As the owners of ARM they could just refuse to develop new cores and that would leave those like Qualcomm, Samsung, Mediatek, Rockchip, Amlogic etc etc all in the dog house.

Allowing that level of anti competitive behaviour is basically like cutting the actual potential of a market and therefore the jobs created by said market - this much at the very least the US govmt would be interested in.

Apple could even refuse to allow uArch licenses if they owned ARM - it just isn't a win for anyone but Apple, and it isn't even much of a win for them, it's an expression of defeat implying that they can't compete without cutting off their competitors IP suppliers.

The core licenses are perpetual, so even if no one was designing new ARM cores any longer Qualcomm can keep making ARM SoCs and shrinking them, they'd just be frozen on whatever core is the most recent they've licensed - A78/X1 presumably. A buyer of ARM can't pull the license Qualcomm already has.

Even if Apple didn't get into legal trouble (they would effectively control the entire mobile chip market if they controlled ARM and used it as a club to beat Samsung et al so I think they'd be in violation of the Sherman Act) it wouldn't really make that much difference.

Would Apple extending their performance lead over Android phones more than it already has make anyone here who buys Android switch? I doubt it. People who were choosing 'the fastest phone' (if there are any such people) were already buying iPhones.
 

ThatBuzzkiller

Senior member
Nov 14, 2014
979
127
106
In just chipmakers, I'd say it'd be like this:
Nvidia (could really use a CPU so they don't have to rely on competitors' products, ARM is starting to take off in server/enterprise, and is dominant in mobile)
Nvidia hates CPUs and would vastly prefer it if all valid C++ code could run on their GPUs which is why they have a massive representation on the C++ design committee standard. They'd rather spend money on improving CUDA than buying ARM to improve it just so all the other vendors could benefit off from their work. Nvidia doesn't like sharing their work so there's no chance of them buying ARM just to empower the other vendors from their exclusive work ...

Intel (couldn't beat them so might as well buy them, maybe is looking to limit ARM to mobile; maybe even heding their bets in case x86 proves problematic - if nothing else it provides them easy means to produce at other foundries should they keep having issues with their own)
Sounds more like Intel would buy ARM just to spite the other vendors rather than transitioning to it. I can see it right now, Intel shells out over $30B just to shutdown ARM's only design house so that every other vendor will have to make a duplicated effort of designing their own competitive implementations rather than relying on a centralized CPU designer like they used to ...

Not a good idea that ARM CPU designs were mostly centralized and I doubt the industry could find a replacement ...

Samsung (just seems like a Samsung thing to do, and there were rumors they thought about buying AMD and ARM would be a bigger get for them I think)
Bad idea for them to expand when they're struggling to maintain a leading edge on all of their in-house technologies. Hard enough keeping their memory and display technology as world leading while their fabs are falling behind ...

Apple (I just don't think they'd want to own it, we saw how they are with the GPU with Imagination, but they might view it as integral to maintaining access/IP rights with them moving Macs towards ARM as well).
This could trigger antitrust laws ...

Qualcomm (considering they supposedly have been shopping their chip business I don't see this)
They very much like this current arrangement since it allows them to solely focus on developing 6G while ARM subsidizes designing their CPUs ...

AMD (I think has good reasons to maybe want it, but don't think they're in financial position to make such a move; it would be bold, would make them not reliant on Intel's control of x86, instantly gives them mobile access, plus companies looking at ARM to deal with Intel's dominance could be partly limiting to them already and likely might increase; it should be a boon for their semi-custom business; also could be big if they require licensing of their GPU - while open sourcing the software as much as possible to placate concerns about lock-in, so kinda take the anti-Nvidia path where you let people license the hardware and then offer open source software stack)
AMD hates ARM too in general since they have less of a say as an ARM CPU vendor than as an x86 CPU vendor ...
 

cortexa99

Member
Jul 2, 2018
89
113
76
Huawei. Hah!
I'm afraid it's not going to happen. I've heard that reason why Softbank selling ARM is because ARM has unintended involved in politics conflict between UK(probably include US) & China. It's a long story that can't tell enough detail in a single post.

In short, ARM has a subsidiary in China(ARM-CN), which is in charge by Chinese management, but this subsidiary was trying to get rid of its parent company, the ARM-UK:

Since 2018 the ARM-CN developed some tech that suit to China but it's out of the subsidiary's range of rights, ARM-UK wanna kick out the ARM-CN's chinese ceo but the ceo refused to hand out his seat. OTOH more interesting is the chinese ceo declared months ago that ARM still keep cooperation to Huawei.

Right now you see UK decide to remove Huawei infrastructure, you know how complicated the situation is.
 

moinmoin

Golden Member
Jun 1, 2017
1,781
1,789
106
I'm afraid it's not going to happen. I've heard that reason why Softbank selling ARM is because ARM has unintended involved in politics conflict between UK(probably include US) & China. It's a long story that can't tell enough detail in a single post.

In short, ARM has a subsidiary in China(ARM-CN), which is in charge by Chinese management, but this subsidiary was trying to get rid of its parent company, the ARM-UK:

Since 2018 the ARM-CN developed some tech that suit to China but it's out of the subsidiary's range of rights, ARM-UK wanna kick out the ARM-CN's chinese ceo but the ceo refused to hand out his seat. OTOH more interesting is the chinese ceo declared months ago that ARM still keep cooperation to Huawei.

Right now you see UK decide to remove Huawei infrastructure, you know how complicated the situation is.
I'm not sure I trust these rumors outright, but the situation certainly is quite complex. Arm being British means it's not directly affected by the US - Chinese trade war, aside of its design shops in the US. SoftBank being Japanese may have been originally motivated to buy Arm thanks to Fujitsu's Fugaku supercomputer which as today's fastest supercomputer completely builds upon custom Arm chips. It may have been the national Japnese interest to have local companies control the supply chain, with SoftBank getting respective support for the acquisition. Then SoftBank from the start saw China as a major market for Arm and IoT (SoftBank's official reason for its interest in Arm). It was SoftBank's push for a more independent local Chinese subsidiary of Arm to accelerate licensing and production of Arm based products in China. Of course with the world's politics developing like it did in the last couple years this is going to make for plenty of awkward interaction between the UK HQ and the SoftBank backed Chinese subsidiary.
 
  • Like
Reactions: Tlh97

LightningZ71

Senior member
Mar 10, 2017
500
415
106
I don't really read it as a violation of anti-trust law for Apple to purchase ARM. The ARM licenses are perpetual. Once you have the license to produce a core, you can produce that core forever. If you purchased the right to modify that core, you still have that right. What you don't have access to is future iterations of the ARM core, which you didn't have before either as you were still required to license the next version.

As for stifling competition, there are alternatives on the market. While those alternatives may not be as good, its not "impossible" to switch to them. It will be quite expensive of course, as you'd need to not only switch to a different instruction set (x86, RiscV, etc) but you'd also need to recompile Android and all the applications that you're shipping with the phone.

I also see where Apple may not use the ARM IP defensively. With Apple switching the Mac line to ARM, it is in their best interests to continue to push the rest of the industry towards ARM as well. One of their issues is going to be nudging software vendors towards producing software for the new platform. Closing down the ability for competitors to cost effectively produce ARM designs will limit the appeal of targeting ARM for software development as the industry moves to a different platform.
 
  • Like
Reactions: beginner99

CluelessOne

Member
Jun 19, 2015
64
40
91
It is by definition anti competitive to acquire something that would essentially allow you to control a market you are currently competing in.

As the owners of ARM they could just refuse to develop new cores and that would leave those like Qualcomm, Samsung, Mediatek, Rockchip, Amlogic etc etc all in the dog house.

Allowing that level of anti competitive behaviour is basically like cutting the actual potential of a market and therefore the jobs created by said market - this much at the very least the US govmt would be interested in.

Apple could even refuse to allow uArch licenses if they owned ARM - it just isn't a win for anyone but Apple, and it isn't even much of a win for them, it's an expression of defeat implying that they can't compete without cutting off their competitors IP suppliers.
Isn't architecture license basically just licensing the ISA and how it communicate with other parts of the CPU? The licensee make custom cores based on the specifications? As long as there is no extension they need what impediment to use it for a few generation CPU?
 

dullard

Elite Member
May 21, 2001
22,422
752
126
How, short of starting a war?
China thinks that Taiwan is part of its control. They are currently dismantling the "one country, two systems" agreement with Hong Kong and Macau and has indicated interest to include Taiwan in that as well. China is building up military in the seas. I don't think China wants a war, they'll probably do it through economic paths (coerced financial agreements). But we can't rule out war as a possibility either.
 

Doug S

Senior member
Feb 8, 2020
326
455
96
With Apple switching the Mac line to ARM, it is in their best interests to continue to push the rest of the industry towards ARM as well. One of their issues is going to be nudging software vendors towards producing software for the new platform. Closing down the ability for competitors to cost effectively produce ARM designs will limit the appeal of targeting ARM for software development as the industry moves to a different platform.
Yes, it is in Apple's interest to see a market for ARM based Windows PCs to develop, it can only help Apple by removing any performance related objections to running x86 Windows applications on the Mac. Might even encourage some authors of Windows only applications to consider porting to Macs after they are ARM native - to gain access to the best performing ARM platform.
 

soresu

Golden Member
Dec 19, 2014
1,420
591
136
China thinks that Taiwan is part of its control. They are currently dismantling the "one country, two systems" agreement with Hong Kong and Macau and has indicated interest to include Taiwan in that as well. China is building up military in the seas. I don't think China wants a war, they'll probably do it through economic paths (coerced financial agreements). But we can't rule out war as a possibility either.
Taiwan is a completely separate state whether or not China or anybody else acknowledge it as so.

TSMC is one of its main exports and a centrepiece in the production flow of many US companies including Apple and AMD.

The govmt may be jousting with TSMC over ASML machines at present, but they are not going to allow TSMC to get into Chinese hands - there would be war if they tried, with the US because semicon power is just as important as oil and gold in todays economy.
 

ThatBuzzkiller

Senior member
Nov 14, 2014
979
127
106
The govmt may be jousting with TSMC over ASML machines at present, but they are not going to allow TSMC to get into Chinese hands - there would be war if they tried, with the US because semicon power is just as important as oil and gold in todays economy.
The semiconductor industry is far from being sustainable to that degree. In 10 years it will have to come to terms with a post-silicon world where it's reaching the limits of physics. China is destined to catch up as it has all the cards in it's hands even if they have to source their own photolithography technology too ...

I can see some OPEC nations easily supplying oil for the next 30+ years by comparison. Taiwan is going to have to find another industry to excel in or diversify because their leading advantage in logic fabrication is eroding faster than they think. They aren't even good at the other specialized fields in the semiconductor industry like NAND/DRAM memory fabrication, display panel production, telecommunications, or chip design as well. Taiwan is eventually going to have to prove themselves to be more than just a one trick pony otherwise there really isn't case for them to stay independent when a hostile state right next to them becomes technologically superior to them in every category ...
 

soresu

Golden Member
Dec 19, 2014
1,420
591
136
The semiconductor industry is far from being sustainable to that degree. In 10 years it will have to come to terms with a post-silicon world where it's reaching the limits of physics.
The silicon of today is not the silicon of yesteryear*.

Likewise foundries will adapt and take on new devices and materials like spintronics and photonics as time goes by.

Some fabs are already integrating embedded MRAM memory into their processes, so spintronics is definitely on the horizon for them

It's very unlikely that TSMC is just running on a CMOS silicon only roadmap for the future - they just don't go running their mouth off about process nodes far off into the future like Intel and their promised MESO process, which at this point sounds like it was just a PR move while they were still bringing up 10nm Ice Lake.

The fact that TSMC won't even talk about their non MBCFET next gen 3nm node despite Samsung boasting about their own MBCFET 3nm node supports the view of TSMC's tight lipped nature about the future.

*I mean silicon is still silicon, but the process nodes have changed materials in different places over the years - they are not unamenable to change.
 
  • Like
Reactions: Tlh97

ThatBuzzkiller

Senior member
Nov 14, 2014
979
127
106
Likewise foundries will adapt and take on new devices and materials like spintronics and photonics as time goes by.

Some fabs are already integrating embedded MRAM memory into their processes, so spintronics is definitely on the horizon for them

It's very unlikely that TSMC is just running on a CMOS silicon only roadmap for the future - they just don't go running their mouth off about process nodes far off into the future like Intel and their promised MESO process, which at this point sounds like it was just a PR move while they were still bringing up 10nm Ice Lake.

The fact that TSMC won't even talk about their non MBCFET next gen 3nm node despite Samsung boasting about their own MBCFET 3nm node supports the view of TSMC's tight lipped nature about the future.

*I mean silicon is still silicon, but the process nodes have changed materials in different places over the years - they are not unamenable to change.
That's not how the industry operates. A new technology is only brought about when a competing implementation is made possible ...

This is currently not the case different materials or types of logic devices compared to GAAFETs transistors. If we don't hear any plans of a radical change in the paradigm of transistor technology by the end of 2023 then TSMC and the rest of the industry should start to seriously wonder that they'll be running into a brick wall pretty soon. It was hard enough envisioning whether or not EUV lithography would've panned out over the many years of false starts ... (there's a high chance the using different materials or types of logic are also failed ideas as well)

We are starting to get a much more clearer picture about the limitations of transistor technology and it's alternatives since duds are piling up quickly ...
 

soresu

Golden Member
Dec 19, 2014
1,420
591
136
This is currently not the case different materials or types of logic devices compared to GAAFETs transistors. If we don't hear any plans of a radical change in the paradigm of transistor technology by the end of 2023 then TSMC and the rest of the industry should start to seriously wonder that they'll be running into a brick wall pretty soon.
There are any number of different logic types under development at the moment, spintronic being chief among the general direction, though there are many sub types of spintronic logic currently in development.

As I have said before - TSMC is not likely to squeal about experimental fab technology for a far off node when they will not even discuss their non MBCFET 3nm node.

The reasoning for boasting of far off processes is usually when ones current market competitiveness or performance is less than stellar, as with Samsung when they initially announced their 3nm MBCFET/Nanosheet/GAA node, and likely with Intel when they were talking about MESO when Ice Lake was still not out the door.

Rest assured - just because TSMC are not talking about such things does not mean that they are not researching them, albeit at this point I would imagine that Nanosheets and Forksheets are on their more immediate future horizons past finFET.
 
  • Like
Reactions: moinmoin

ThatBuzzkiller

Senior member
Nov 14, 2014
979
127
106
There are any number of different logic types under development at the moment, spintronic being chief among the general direction, though there are many sub types of spintronic logic currently in development.

As I have said before - TSMC is not likely to squeal about experimental fab technology for a far off node when they will not even discuss their non MBCFET 3nm node.

The reasoning for boasting of far off processes is usually when ones current market competitiveness or performance is less than stellar, as with Samsung when they initially announced their 3nm MBCFET/Nanosheet/GAA node, and likely with Intel when they were talking about MESO when Ice Lake was still not out the door.

Rest assured - just because TSMC are not talking about such things does not mean that they are not researching them, albeit at this point I would imagine that Nanosheets and Forksheets are on their more immediate future horizons past finFET.
How quiet a corporation is doesn't say anything about general industry trends. Even Intel sees GAAFETs as the new trend despite their recent performance so it's not just Samsung that's alone ...

There is no consensus or even the slightest hint of one to somehow believe that new materials or new types of logic will replace silicon. We've failed pretty hard in finding a replacement for the paradigm around silicon transistors for the last 2 decades so chances are we'll fail during the next decade as well but I guess TSMC/Taiwan had a pretty good run all things considered ...

We can't even design electronic circuits around over the concept spintronics which just goes to shows you how dire the future truly is ...

Taiwan has to prepare for a backup plan (if they even have one) in case their wells start drying up because even if they do manage to change the paradigm how much could they improve upon it ?
 

ASK THE COMMUNITY