Sony to leave three-way microchip development pact with NEC and Toshiba

[patsu]

I can appreciate both sides of the argument. Unfortunately I'm a total noob at this.

What is a logical next step for Sony ? Can it participate in advanced process research with TSMC and IBM (without owning any fabs) ? Perhaps in exchange for better deals ? What other possible moves are there ?

 

[Carl B]

These use to be the reasons to go internal. Problem is, all of these reasons are going by the wayside. Internal fabs don't really save money anymore, nor does need for that last mile of performance a big deal either. No one would go with foundries if trade secrets are at a major risk of being leaked. Competition from vertically integrated companies are also a non-threat, since Intel is the only true vertically integrated company today (excluding Samsung).

Well, truthfully Sony itself is vertically integrated across a number of areas as well, one being digital imaging and another being the Playstation business, if we're talking about fab needs.

This is page 3 of the original thread on this topic (begun after Nakagawa's original comments) back from February: http://forum.beyond3d.com/showthread.php?t=38609&page=3

A lot of what we might or might not otherwise go over here in this thread (at least in terms of my own arguments) is touched on there.

Indeed, since it is the default, we should view Sony's position until now as on a course towards internal 45nm build-out, with only the most recent moves indicating direction might change. I don't disagree with you that there are reasons to consider this on Sony's part, but I think ultimately it has a lot more to do with the Stringer/Nakagawa paradigm of expense reduction vs the Kutaragi (and to an extent Manabe) paradigm of pursuing a strong internal R&D and fabrication capacity.

I think frankly at the time Stringer took over and Kutaragi was removed from being head of CE and semiconductors, the company's silicon efforts reached an inflection point that is going to see a very different Sony semi strategy than otherwise may have arisen if KK pursued his IBM/Toshiba R&D alignments to their logical conclusion.

That's a vastly overplayed threat. Foundries are not bad at what they do, and it's much more likely that an internal fab company will run into trouble with its own fab than a foundry would.

I think the foundries are great - what's not to like? You want to make a chip, and they'll fab it up for you; you don't like one, you've got two other majors to choose from. And I feel that can work more or less problem-free for one gen, but after that you've basically put your fate in their hands... it's about control to an extent. Once you duck out of process R&D - that's it - you'll never get back to where you otherwise would have been. Though outsourcing production is the fiscally progressive option in the short-term, foundries work especially well in a hybrid model to supplement internal fab capacities. Becoming 100% reliant on them on the other hand puts all your eggs in a basket someone else holds. You see from that article I linked to how (relatively) little Texas Instruments will save by forgoing their own process tech; are the margins worth tossing that fundamental industry participation out the window for?

I don't think there are yes/no answers to any of this, but for myself flags go up when I fear that short-term profit growth may be determining strategy that will effect the next 10 years in the space. Stringer's basically pursuing a policy that will get him 'reelected' by meeting his promise of 5% margins... but we all knwo that sometimes the political campaign is based on the pursuit of goals that are in fact detrimental to the macro picture.

 

[Carl B]

What is a logical next step for Sony ? Can it participate in advanced process research with TSMC and IBM (without owning any fabs) ? Perhaps in exchange for better deals ? What other possible moves are there ?

In either case, whether or not Sony pulls out of fabrication research, they'll still remain actively engaged in chip design research - it is the entire premise of their semi division afterall. Think of it as going away from the Intel model more towards the NVidia model.

Sony could of course help fund external line buildouts (such as they did at East Fishkill) in return for capacity rights/guarantees, or fund R&D for process development in return for some other benefit, but I really wouldn't understand either outside of the context of a formal fab interest. And since we're here speaking in the hypothetical of a pullback from that strategy, it would have to be a very specific case to warrant it I imagine. Of course it doesn't have to be black and white; Sony may indeed trend fab-lite goign forward but still pursue joint fabs with Toshiba and/or IBM as the opportunities present themselves (and they of course will). Right now though the conversation ultimately boils down to whether or not it's worth it for Sony to maintain a strong internal knowledge base with regard to fabrication technology; for various reasons, I personally think it is of benefit, especially if you already have a research alliance structure set up to begin with.

 

[patsu]

Yes, from your discussions so far, fabs seem too important for Sony to "give up/outsource" in the simplest sense. I reckon Sony management would keep their plans flexible and will have contingencies in place.

Special deals and arrangements with external lines sound like a possible foot forward.

 

[Deusp]

Well, truthfully Sony itself is vertically integrated across a number of areas as well, one being digital imaging and another being the Playstation business, if we're talking about fab needs.

This is page 3 of the original thread on this topic (begun after Nakagawa's original comments) back from February: http://forum.beyond3d.com/showthread.php?t=38609&page=3

A lot of what we might or might not otherwise go over here in this thread (at least in terms of my own arguments) is touched on there.

Indeed, since it is the default, we should view Sony's position until now as on a course towards internal 45nm build-out, with only the most recent moves indicating direction might change. I don't disagree with you that there are reasons to consider this on Sony's part, but I think ultimately it has a lot more to do with the Stringer/Nakagawa paradigm of expense reduction vs the Kutaragi (and to an extent Manabe) paradigm of pursuing a strong internal R&D and fabrication capacity.

I think frankly at the time Stringer took over and Kutaragi was removed from being head of CE and semiconductors, the company's silicon efforts reached an inflection point that is going to see a very different Sony semi strategy than otherwise may have arisen if KK pursued his IBM/Toshiba R&D alignments to their logical conclusion.

I think your point is that you think this is shortsighted of Sony to do this, and that you think this is a Sony unique move. I don't either one is true. Everyone is going towards this direction, and the advantage shrinks every generation in relationship to the final product. There was an article a few months ago (can't find it now) that stated that eventually there will just be two fab technology camps: Intel's and IBM's, with everyone's fabs based off of one of these two companies' technology. If this is endgame result, then whether you go foundry or internal the end result may be exactly the same.

I think the foundries are great - what's not to like? You want to make a chip, and they'll fab it up for you; you don't like one, you've got two other majors to choose from. And I feel that can work more or less problem-free for one gen, but after that you've basically put your fate in their hands... it's about control to an extent. Once you duck out of process R&D - that's it - you'll never get back to where you otherwise would have been. Though outsourcing production is the fiscally progressive option in the short-term, foundries work especially well in a hybrid model to supplement internal fab capacities. Becoming 100% reliant on them on the other hand puts all your eggs in a basket someone else holds. You see from that article I linked to how (relatively) little Texas Instruments will save by forgoing their own process tech; are the margins worth tossing that fundamental industry participation out the window for?

It's not quite so simple. Foundry processes need substantial R&D to get to working properly. Move a chip from your own fabs to a foundry have their own difficulties. It's quite likely that for high-end stuff, it's either/or and not hybrid. You forget to an extent IBM, which is something of a foundry that takes chip design orders rather than a preexisting design.

I don't think there are yes/no answers to any of this, but for myself flags go up when I fear that short-term profit growth may be determining strategy that will effect the next 10 years in the space. Stringer's basically pursuing a policy that will get him 'reelected' by meeting his promise of 5% margins... but we all knwo that sometimes the political campaign is based on the pursuit of goals that are in fact detrimental to the macro picture.

I don't think the macro picture favors internal fabs either, unless you're Intel anyways. For everyone else, both the short-term and long-term benefits support outsourcing fab production.

 

[nelg]

I think that being vertically integrated works well from a c/b point when you approach full utilization and above. I am sure if we look at Cell, if it does not pervade other markets and PS3 demands are not as predicted, Sony will not be able to produce them cost effectively. Economics can be like GPU's, if you can't absorb the latency you are going to be vulnerable.

 

[standing ovation]

How about this?

[size=-2]"San Jose, Calif. — The IC business is fast becoming unaffordable for all but the wealthiest chip makers.

Rising fab and chip-design costs are creating a new elitism, with the well-heeled few able to shoulder the burden and the mass of suppliers squeezed out, according to data from one design tool vendor. The trend will become even more pronounced starting at 45-nanometer manufacturing, Synopsys Inc. predicted at last week's International Symposium on Quality Electronic Design (ISQED) here." - EETimes[/size]​

Basically the article states that as you move to a smaller design process, costs should increase exponentially. This means you will have to sell more chips just to break even. So rather than rely on pandemic sales in the future, it makes cents to start cutting costs now. Hence we are seeing companies like Texas Instruments getting out of one business in order to specialize in another.

It's all about taking advantage of someone else's expertise (and economies of scale) without having to build them yourself. That's not a bad thing, is it?

 

[Carl B]

I think your point is that you think this is shortsighted of Sony to do this, and that you think this is a Sony unique move. I don't either one is true. Everyone is going towards this direction, and the advantage shrinks every generation in relationship to the final product. There was an article a few months ago (can't find it now) that stated that eventually there will just be two fab technology camps: Intel's and IBM's, with everyone's fabs based off of one of these two companies' technology. If this is endgame result, then whether you go foundry or internal the end result may be exactly the same.

I'm a big believer in the Intel vs IBM model as well when it comes to PC/server chips, only as a counterpoint I do also see room for the Asian CMOS process expertise to live on in the CE space - IBM is frankly not all that great in that space. Hell there're lots of articles out there, each with different opinions, and some of them written by me - I'm not saying I'm right and others are wrong, but I do bias towards the continued internal view of the industry vs the other.

By the way I don't think its shortsighted per se by Sony, I'm saying that the factors driving it are not in appearance the ones that should be. Whether it ends up actually being right or not is independent of what has put it in motion. Likewise I obviously acknowledge that others are moving in this direction as well, since TI - third largest semi company - has opted to go this route as well. Which you might imagine, I also view as a dubious move on their parts...

(and they do indeed compete directly with Intel, just not on the desktop)

It's not quite so simple. Foundry processes need substantial R&D to get to working properly. Move a chip from your own fabs to a foundry have their own difficulties. It's quite likely that for high-end stuff, it's either/or and not hybrid. You forget to an extent IBM, which is something of a foundry that takes chip design orders rather than a preexisting design.

IBM is small though in terms of their industry footprint - more of a high-end boutique in terms of a foundry, and frankly not all that hot in the actual fabbing itself. What they're good at is fab technology, but even that can't be viewed as internally driven - it's possible due to the number of firms relying on them to keep going, and funding that research. Intel is a monolith, but in your Intel/IBM world remember that IBM would fall apart if not for alliance members depending on (and funding) their further research. If Sony drops out of fabbing, that's frankly a major partner of IBM's - Sony invested large large amounts to help IBM get its 65nm SOI process off the ground, and owns rights to some of the East Fishkill production line. Right now it's IBM, Chartered, AMD on one axis, and IBM, Sony, Toshiba on another. IBM needs more members of that hodgepodge rather than fewer if they are going to take on Intel in a process holding action.

And again, I view the CMOS process advancements as a separate - yet significant - factor in the CE industry as well. Toshiba, NEC, Samsung... Sony?... someone's got to keep at it.

I don't think the macro picture favors internal fabs either, unless you're Intel anyways. For everyone else, both the short-term and long-term benefits support outsourcing fab production.

It's not clearcut; if it were, everyone would be doing one, or doing the other - and right now there are a lot of companies still pursuing their own internal 45nm strategies. Frankly if all of these companies did choose to outsource, let's be honest - TSMC, UMC, and Chartered aren't going to have the 45nm capacity to accommodate them all, and that means higher prices in a supply/demand breakout. Higher prices swing the pendulum back towards internal, except once you get off that train, you don't get back on.

Again, I'm not saying I'm absolutely right on this - hell one doesn't even have to be right or wrong - the results can just be viewed as "different." But again, the action plan at Sony should be put into motion after a lengthy review process independent of cost-cutting goals associated with Stringer's end of year promise.

 

[Carl B]

I think that being vertically integrated works well from a c/b point when you approach full utilization and above. I am sure if we look at Cell, if it does not pervade other markets and PS3 demands are not as predicted, Sony will not be able to produce them cost effectively. Economics can be like GPU's, if you can't absorb the latency you are going to be vulnerable.

I agree, and the Nagasaki 65nm fab is singularly risky when compared to other fab efforts in that it is an SOI line, and thus as far as it concerns Sony geared towards the production of Cell alone. Obviously when the plans were laid, KK was in charge of CE, gaming, and semis all... and you can believe that that fab would have been cranking out Cell processors all day long. Only now has it even come on line though, so it's hard to gauge what the demand is or is not; since Sony has been sourcing from IBM at 90nm *and* has a 65nm supply option with them, previously its obvious Sony predicted that their own internal fab capacity would not be enough. Implying full utilization, implying a fab well executed. Now the picture is less clear - we know that demand is presently lower than anticipated, but we still don't know if it breaks below their own internal capacity.

How about this?

[size=-2]"San Jose, Calif. — The IC business is fast becoming unaffordable for all but the wealthiest chip makers.

Rising fab and chip-design costs are creating a new elitism, with the well-heeled few able to shoulder the burden and the mass of suppliers squeezed out, according to data from one design tool vendor. The trend will become even more pronounced starting at 45-nanometer manufacturing, Synopsys Inc. predicted at last week's International Symposium on Quality Electronic Design (ISQED) here." - EETimes[/size]​

Basically the article states that as you move to a smaller design process, costs should increase exponentially. This means you will have to sell more chips just to break even. So rather than rely on pandemic sales in the future, it makes cents to start cutting costs now. Hence we are seeing companies like Texas Instruments getting out of one business in order to specialize in another.

It's all about taking advantage of someone else's expertise (and economies of scale) without having to build them yourself. That's not a bad thing, is it?

Yeah, I read that article when it came out. Indeed, I like it - informative and paints a snapshot of the industry. But I think it's a little disingenuous of you to cite it as an indicator of what Sony should or should not do since Sony themselves is explicitly mentioned within it:

...In Lammers' view, the R&D picture is also less bleak than Williams of Synopsys paints it. "Intel is about the only logic company that can do it alone," Lammers said, but many IDMs have spread their R&D and fab costs by "following the consortium model."

One example is IBM Corp.'s manufacturing-technology alliance, which includes AMD, Chartered, Freescale, Infineon, Samsung, Sony and Toshiba. In this "fab club," the vendors plan to share the R&D costs at the 65-nm node and beyond...

And this is what this thread is all about - is Sony better off sticking in their R&D alliances in order to maintain internal KB, or are they better off going fab-lite a la TI?

By the way, Texas Instruments isn't "getting out of one to specialize in another," they're just simply getting out of one. The "other" you refer to is something they do and have been doing anyway.

 

[Capeta]

Japan's top semiconductor firms form fab consortium to compete with US and Korea.

http://www.eweek.com/article2/0,1895,1889022,00.asp

 

[Deusp]

I'm a big believer in the Intel vs IBM model as well when it comes to PC/server chips, only as a counterpoint I do also see room for the Asian CMOS process expertise to live on in the CE space - IBM is frankly not all that great in that space. Hell there're lots of articles out there, each with different opinions, and some of them written by me - I'm not saying I'm right and others are wrong, but I do bias towards the continued internal view of the industry vs the other.

By the way I don't think its shortsighted per se by Sony, I'm saying that the factors driving it are not in appearance the ones that should be. Whether it ends up actually being right or not is independent of what has put it in motion. Likewise I obviously acknowledge that others are moving in this direction as well, since TI - third largest semi company - has opted to go this route as well. Which you might imagine, I also view as a dubious move on their parts...

(and they do indeed compete directly with Intel, just not on the desktop)

I'm sure Sony had a rational debate over this new direction of theirs, and they concluded they would be better off this way. I doubt if corporate politics was the major player here, though I'm sure it played some role. It's to be seen if this move was smart or dumb. It may even show a degree of foresight some time down the road.

IBM is small though in terms of their industry footprint - more of a high-end boutique in terms of a foundry, and frankly not all that hot in the actual fabbing itself. What they're good at is fab technology, but even that can't be viewed as internally driven - it's possible due to the number of firms relying on them to keep going, and funding that research. Intel is a monolith, but in your Intel/IBM world remember that IBM would fall apart if not for alliance members depending on (and funding) their further research. If Sony drops out of fabbing, that's frankly a major partner of IBM's - Sony invested large large amounts to help IBM get its 65nm SOI process off the ground, and owns rights to some of the East Fishkill production line. Right now it's IBM, Chartered, AMD on one axis, and IBM, Sony, Toshiba on another. IBM needs more members of that hodgepodge rather than fewer if they are going to take on Intel in a process holding action.

And again, I view the CMOS process advancements as a separate - yet significant - factor in the CE industry as well. Toshiba, NEC, Samsung... Sony?... someone's got to keep at it.

Sony would continue funding IBM's fab technology with various joint research ventures and future product order. That in itself won't end anything soon. They just won't be a self-fabbing company with all the bells and whistles in terms of research. Most of the research is being carried by IBM and Intel anyways right now, with most other companies being side-shows in terms of research. Fundamental research on future processes are probably going to be so great any additional research by smaller chipmakers probably won't do much, especially when there is so much consolidation going on right now.

It's not clearcut; if it were, everyone would be doing one, or doing the other - and right now there are a lot of companies still pursuing their own internal 45nm strategies. Frankly if all of these companies did choose to outsource, let's be honest - TSMC, UMC, and Chartered aren't going to have the 45nm capacity to accommodate them all, and that means higher prices in a supply/demand breakout. Higher prices swing the pendulum back towards internal, except once you get off that train, you don't get back on.

I think that's another bogeyman. Very few products even need 45nm never mind the ability to make a 45nm chip. I also think you underestimate the foundries process tech. TSMC had 65nm up and running for a year now. 45nm is expect to hit in Q3 of this year. Plus they're going to have a giant giga-fab, capable of churning out 80,000-100,000 wafers a month. In terms of production prowess, the foundries are second only to Intel. Not to say the foundries aren't being pressed for capacity (which they are), but it seems to be most the cheap and high-volume stuff.

Again, I'm not saying I'm absolutely right on this - hell one doesn't even have to be right or wrong - the results can just be viewed as "different." But again, the action plan at Sony should be put into motion after a lengthy review process independent of cost-cutting goals associated with Stringer's end of year promise.

I'm sure that was what happened. We'll see what the end result is a few years down the road.

 

[Carl B]

Sony would continue funding IBM's fab technology with various joint research ventures and future product order. That in itself won't end anything soon. They just won't be a self-fabbing company with all the bells and whistles in terms of research. Most of the research is being carried by IBM and Intel anyways right now, with most other companies being side-shows in terms of research. Fundamental research on future processes are probably going to be so great any additional research by smaller chipmakers probably won't do much, especially when there is so much consolidation going on right now.

Just to be clear again, Sony and the rest of IBMs partners are in fact larger chipmakers than IBM itself - IBM is smaller then allof them when it comes to actual semi revenues. What IBM is, is a design house. And that's all well and good, but I think you tend to overplay IBM's fabbing footprint and underplay that of the non-PC/server side companies.

I think that's another bogeyman. Very few products even need 45nm never mind the ability to make a 45nm chip. I also think you underestimate the foundries process tech. TSMC had 65nm up and running for a year now. 45nm is expect to hit in Q3 of this year. Plus they're going to have a giant giga-fab, capable of churning out 80,000-100,000 wafers a month. In terms of production prowess, the foundries are second only to Intel. Not to say the foundries aren't being pressed for capacity (which they are), but it seems to be most the cheap and high-volume stuff.

I read that announcement of TSMC's back in the day as well, but I can point to several other companies on that list that announced 65nm as being 'ready' way back when. Fact is that one of TSMC's most high-profile complex chip categories - GPUs - have yet to see a transition en masse to 65nm, and Microsoft has had their transition to 65nm on 360 delayed as well on the foundry front.

I'm a fan of foundries again, and TSMC in particular, but like all companies their press releases tend to overplay the level to which they've landed on a certain process.

Also I'm not sure what you're implying when you say few products "need" 45nm... it's about putting twice as many chips (in the theoretical) on the same wafer, and the associated shrinks it allows in CE products, increases in battery life, increases on product margins, additional complexity (if desired), and movement to system-on-chip designs. These are *crucial* differentiating factors to CE manufacturers looking to differentiate themselves internally on the market or to third-parties looking for chip solutions.

 

[Deusp]

Just to be clear again, Sony and the rest of IBMs partners are in fact larger chipmakers than IBM itself - IBM is smaller then allof them when it comes to actual semi revenues. What IBM is, is a design house. And that's all well and good, but I think you tend to overplay IBM's fabbing footprint and underplay that of the non-PC/server side companies.

Excluding IBM then. They're special.

I read that announcement of TSMC's back in the day as well, but I can point to several other companies on that list that announced 65nm as being 'ready' way back when. Fact is that one of TSMC's most high-profile complex chip categories - GPUs - have yet to see a transition en masse to 65nm, and Microsoft has had their transition to 65nm on 360 delayed as well on the foundry front.

I'm a fan of foundries again, and TSMC in particular, but like all companies their press releases tend to overplay the level to which they've landed on a certain process.

That has more to do with the companies designing the chips and not TSMC. Some chips were made at 65nm last year, just not many and nothing particularly aggressive.

Also I'm not sure what you're implying when you say few products "need" 45nm... it's about putting twice as many chips (in the theoretical) on the same wafer, and the associated shrinks it allows in CE products, increases in battery life, increases on product margins, additional complexity (if desired), and movement to system-on-chip designs. These are *crucial* differentiating factors to CE manufacturers looking to differentiate themselves internally on the market or to third-parties looking for chip solutions.

I was referencing to the timeframe of when 45nm will first go online. Obviously, nearly everyone would move to 45nm eventually. For most of the companies now, 45nm does little more than reduce costs and only a few companies (like the GPU makers) need that shrink, and even then not any time soon. Power consumption reductions is probably not going to happen right out the gate either, given the leakage situation of 65nm and below right now. All I'm saying is that by the time they are ready and are in need of 45nm, TSMC will have it ready for a long time.

 

[Carl B]

We'll just wait and see what happens; as news arises I'm sure they'll be plenty to talk about. I think the gigafab model is definitely a strong one, and brings economies of scale that can help even out the costs external vs internal while maintaining some margin for the foundries, but nothing's set in stone just yet. It's worth noting that the individual that wrote that EETimes article you linked to in relation to escalating costs is (Lammers) is also the one that is cited in the article quote I provided above as saying consortium-style R&D efforts are a worthy alternative to the fab-lite model.

EDIT: And in that super-timely fashion news tends to hit in, EETimes has another article for us today with regard to the foundries themselves and the move to 45nm: http://eetimes.com/news/latest/showArticle.jhtml?articleID=198800817&pgno=1

Doesn't relate directly to the companies undergoing internal efforts, but it's definitely good flavor in the context of the discussion.

 

[Frank]

Well, it's a trend happening all over the semiconductor industry. Unless you have serious strategic interest in keeping at the very front and are willing to invest billions every few years, there is less and less point in trying to keep up.

In short: many players feel the possible rewards don't justify the astronomical investments anymore.

Even Philips, a very strong player from the start, and with a large part of the market share in analog and bridging components, has sold their semiconductor division. Simply because they couldn't compete with the top few players without putting their whole company on the line when things might go wrong.

Sure, like most of the other big players, their semiconductor assets were but a part of the whole, less than a quarter of the total. But they required by far the most unsure investments.


In a few years time, when they have to go below 45 nm, I expect only the top few players will survive. Everyone else will be part of one of the opportunistic conglomerates.

That is, unless Microsoft (or some company equally wealthy) decides to step in, with only long-term prospect through market domination as their goal. Or a wealthy research company like IBM discovers something interesting and decides to wager their future on being the first and the best.

 

[standing ovation]

And this is what this thread is all about - is Sony better off sticking in their R&D alliances in order to maintain internal KB, or are they better off going fab-lite a la TI?

The article seemed to put things into a business perspective -- that's why I mentioned it. Given their financial condition, I really think Sony would be better off focusing on one area.

By the way, Texas Instruments isn't "getting out of one to specialize in another," they're just simply getting out of one. The "other" you refer to is something they do and have been doing anyway.

Ah, it's semantics! How about this: Texas Instruments is getting out of one area to concentrate on another.

 

[Shifty Geezer]

In short: many players feel the possible rewards don't justify the astronomical investments anymore.

In a few years time, when they have to go below 45 nm, I expect only the top few players will survive. Everyone else will be part of one of the opportunistic conglomerates.

Isn't there a big problem here in that there won't be enough fabrication ability to satisfy demands? Unless the big producers like TSMC buy all the dropped plants like Sony's, you'll have all the chip consumers of the world wanting to use the chip manufacturers. Will there be enough production capabilitiy to go around?

 

[Carl B]

Ah, it's semantics! How about this: Texas Instruments is getting out of one area to concentrate on another.

I see them as naturally complementary though - imagine it like this, you have the car industry with its own in-house design and manufacturing capabilities; do you believe maintaining one comes at the expense of the other? They are both facets of potential differentiation and competetive advantage.

Now granted cars plants don't grow in expense at the rate fabs do, so the squeeze itself isn't exactly analogous, but I just want to emphasize that running your own fabs is hardly a 'distraction' from IC design. Whether they feel they can afford to continue investing in fab tech is another story - but no doubt if the expense was the same as five years ago, we wouldn't even be having this thread here and everyone of these companies would be maintaining the internal research efforts for the synergies they bring.

 

[Frank]

Isn't there a big problem here in that there won't be enough fabrication ability to satisfy demands? Unless the big producers like TSMC buy all the dropped plants like Sony's, you'll have all the chip consumers of the world wanting to use the chip manufacturers. Will there be enough production capabilitiy to go around?

Prices will raise, which makes it more interesing to invest, after which prices will drop again, etc. Business as usual.

 

[Shifty Geezer]

Prices will raise.

Which is why it'd make sense to invest. If everyone else is pulling out, competition for limited resources will push up fabrication outsourcing prices. Those with their own fabs ough to save a fortune. Yet apparently the thinking is that even then, it'd be cheaper to outsource than run your own 45nm fabrication.