Very interesting. This relaunches GF as a possible new foundry that can adopt the latest PP. New actor on the market = better production capacity = lower prices and faster advancements (none can sleep on its advantage, Intel knows).
I doubt that. I don't think it was a lack of IP that led to GlobalFoundries' abandonment of pursuing the latest node shrinks. I think it had more to do with capital expenditure and the amount of revenue derived from it. My guess is that GlobalFoundries would have been happier with a cash settlement of the lawsuit rather than a cross-licensing agreement, but after analysis and negotiation with TSMC they decided it was too dangerous to try to pursue that through the courts.
Guess we'll find out id due time, but as others pointed out capital requirements for smaller silicon are getting exponential. IP isn't the problem when fabs to use it cost billions.
You can't change your mind about how much money you have. TSMC and Samsung by now are at least 3 years ahead of GF even if they didn't sack all those people after abandoning 7nm. Customer confidence is so VERY low that most of leading edge customers would rather wait for TSMC and Samsung to expand their fabs.
My take as well. I suspect TSMC was about to eat GloFlos lunch in court. They offered a cross licensing deal witch honestly GloFlo is not going to compete with TSMC so why the heck not as they can no longer take you to court. Its win for TSMC in my mind. Globalfoundries would have to invest a ton of money to compete again on any process better than 12nm.
GloFo competes with TSMC in the majority of it market share. If you look at TSMC quarterly reports while 7nm brings the most revenue, nodes from 16/12 28 and 40/45 still brings almost half of revenue combined. I believe TSMC is in position that it knows that the patents given to GloFo will not give them competitive edge over them in those nodes beyond what they already have. While going with legal procedures will just hurt both companies more than benefiting the since in case any company is ordered to stop distribution of infringing products it's huge blow costing billions - too risky for either of them.
This still ends up being a win-win for both companies and even customers. GF ends the litigation on a positive note and the ability to enhance its current foundries further. TSMC gets access to various SOI related patents which can still be helpful for future generations of nodes.
Customers win as there is *potential* for the development tools to become more cross-fab friendly. In other words, designs are engineered toward a more optimal common set of design rules after TSMC and GF start incorporating each other's ideas. This is a big if but it would save plenty of money in the development cycle for clients who want foundry independence.
Cross licensing at this level is always a win-win for companies. It's worked well for AMD and Intel since the 80's...at least until Intel started those Athlon-era shenanigans.
In the end, I think most of these companies 'know' when some of their IP is so broad that it really should be considered fair use. Like when there is literally only one way to do something and denying someone else the ability to do it would be straight up anti-competitive.
it doesn't has to do with GloFo 7nm production, they pretty much got to risk production until they understood they are too small without major customers to justify transition to EUV high capacity production for economical reasons - EUV just cost a lot while manufacturing volume is down. GloFo was in bad situation it didn't make any profit and the owner just got enough of throwing piles of many on them without any returns so they got rid of many of their business and focused on some specialty nodes just to stay profitable, I don't see them even considering returning to the leading node manufacturing in the coming years it just cost to much for them and with their size it does not justify the investment. staying in the technology manufacturing frontier cost a lot that why from over dozen of manufactures in the leading node about two decades ago we are left with just 3 today, you either make it big and invest big cash for R&D and expanding or you fall behind.
You're a "very stable genius" ain't yahs? [snicker]
""In regards to today's quashing of the disputes, we reached out to Patrick Moorhead, Founder, President & Principal Analyst at boutique tech analyst firm Moor Insights & Strategy, for some perspective:
"I believe this agreement is good for the industry as more companies sharing patents and IP accelerates innovation. As GlobalFoundries initiated the lawsuit and is in less diverse markets, I think it likely got the better deal."
In either case, burying the hatchet is a win for both companies, as the litigation would have drug on for years across several jurisdictions. ""[/i]
Only time will tell if GF got the better deal. They will need to come up with billions to start working on a new EUV node like 7nm or 5nm. If they do in fact do that then maybe this works out for them however I don't think they will so I truly believe TSMC got the better deal.
Why are you calling me names? This is an adult site, please if you are not one at least act like one.
they will need tens of billions to compete in the leading node, they might try to compete not in the leading node but 1 or 2 gens behind in the future but as is seems the race is over for them, in this industry you either make it big get huge profit and lead on or you fall behind, from dozens of manufactures in the leading node in the past we are left with just 3 today.
IMHO good on them BOTH, the way the world is heading, we ALL or at least those making the products we want/need have to and should be working together in one way or another, even if it means they are "competitors" making their own sales and such.
Now only if Intel Nvidia Apple and such would do the same,
This whole thing has been some kind of weird chess play. I guess it might have been trying to get access to those patents so it could keep up with its current status of "slightly older architectures" a few years from now? If so, I'm actually impressed with their forwardthinkingness. Or maybe they hoped for some kind of cash settlement, which would have been petty. I do wonder if there was something behind the scenes, like if AMD wanted to jump entirely to TSMC but couldn't because of patent issues, and somehow negotiated with GloFo to have this all done. (We'll hear about it in the next while, if so.)
I don't think we need to be lawyers to deduce that this was the true purpose of GloFo's lawsuit against TSMC all along. Since they apparently own less (and/or less important) patents than TSMC this should be a big win for them.
It's most likely part of putting GF on the market, to remove any uncertainties that might have existed, now GF has been slimmed and any legal question-marks have been straightened out for any potential buyers.
Was thinking since last year they should have licensed 7nm TSMC as it is DUV and could probably be done on their scanners.
Since PS5 / XBOX / Navi1+2 / Zen 2+3 all run on this node it would benefit everyone:
- AMD sells more and gets market - TSMC can't sell more at max cap. so they don*t lose business so they could make extra $ - GF can sell 7nm - Customers get products
AMD has to get dual source. I love Mr. Su as a CEO but I have the feeling she is not agressive enough in taking a risk to get market share and too focused on margins - which was ok when AMD was starved and sucess was not sure but now I think market and mind share is more improtant than max. margin rather than have a good margin and more share as share is the base for future business.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
27 Comments
Back to Article
CiccioB - Monday, October 28, 2019 - link
Very interesting. This relaunches GF as a possible new foundry that can adopt the latest PP.New actor on the market = better production capacity = lower prices and faster advancements (none can sleep on its advantage, Intel knows).
Yojimbo - Monday, October 28, 2019 - link
I doubt that. I don't think it was a lack of IP that led to GlobalFoundries' abandonment of pursuing the latest node shrinks. I think it had more to do with capital expenditure and the amount of revenue derived from it. My guess is that GlobalFoundries would have been happier with a cash settlement of the lawsuit rather than a cross-licensing agreement, but after analysis and negotiation with TSMC they decided it was too dangerous to try to pursue that through the courts.NICOXIS - Tuesday, October 29, 2019 - link
Maybe they changed their minds seeing how much business TSMC is getting.Frenetic Pony - Tuesday, October 29, 2019 - link
Guess we'll find out id due time, but as others pointed out capital requirements for smaller silicon are getting exponential. IP isn't the problem when fabs to use it cost billions.levizx - Wednesday, October 30, 2019 - link
You can't change your mind about how much money you have. TSMC and Samsung by now are at least 3 years ahead of GF even if they didn't sack all those people after abandoning 7nm.Customer confidence is so VERY low that most of leading edge customers would rather wait for TSMC and Samsung to expand their fabs.
FreckledTrout - Tuesday, October 29, 2019 - link
My take as well. I suspect TSMC was about to eat GloFlos lunch in court. They offered a cross licensing deal witch honestly GloFlo is not going to compete with TSMC so why the heck not as they can no longer take you to court. Its win for TSMC in my mind. Globalfoundries would have to invest a ton of money to compete again on any process better than 12nm.Eliadbu - Tuesday, October 29, 2019 - link
GloFo competes with TSMC in the majority of it market share. If you look at TSMC quarterly reports while 7nm brings the most revenue, nodes from 16/12 28 and 40/45 still brings almost half of revenue combined. I believe TSMC is in position that it knows that the patents given to GloFo will not give them competitive edge over them in those nodes beyond what they already have. While going with legal procedures will just hurt both companies more than benefiting the since in case any company is ordered to stop distribution of infringing products it's huge blow costing billions - too risky for either of them.Kevin G - Tuesday, October 29, 2019 - link
This still ends up being a win-win for both companies and even customers. GF ends the litigation on a positive note and the ability to enhance its current foundries further. TSMC gets access to various SOI related patents which can still be helpful for future generations of nodes.Customers win as there is *potential* for the development tools to become more cross-fab friendly. In other words, designs are engineered toward a more optimal common set of design rules after TSMC and GF start incorporating each other's ideas. This is a big if but it would save plenty of money in the development cycle for clients who want foundry independence.
Samus - Wednesday, October 30, 2019 - link
Cross licensing at this level is always a win-win for companies. It's worked well for AMD and Intel since the 80's...at least until Intel started those Athlon-era shenanigans.In the end, I think most of these companies 'know' when some of their IP is so broad that it really should be considered fair use. Like when there is literally only one way to do something and denying someone else the ability to do it would be straight up anti-competitive.
Enter, the patent trolls.
levizx - Wednesday, October 30, 2019 - link
That's DUMB, if TSMC can win easily, why would they agree to cross-licence ? They are not stupid.outsideloop - Monday, October 28, 2019 - link
Lisa Su to TSMC: How about you guys help our other fab make 7nm chiplets, so we can all put Intel out of business??Foeketijn - Tuesday, October 29, 2019 - link
That would also be my guess.Eliadbu - Wednesday, October 30, 2019 - link
it doesn't has to do with GloFo 7nm production, they pretty much got to risk production until they understood they are too small without major customers to justify transition to EUV high capacity production for economical reasons - EUV just cost a lot while manufacturing volume is down. GloFo was in bad situation it didn't make any profit and the owner just got enough of throwing piles of many on them without any returns so they got rid of many of their business and focused on some specialty nodes just to stay profitable, I don't see them even considering returning to the leading node manufacturing in the coming years it just cost to much for them and with their size it does not justify the investment. staying in the technology manufacturing frontier cost a lot that why from over dozen of manufactures in the leading node about two decades ago we are left with just 3 today, you either make it big and invest big cash for R&D and expanding or you fall behind.eek2121 - Monday, October 28, 2019 - link
Saw this coming the second GlobalFoundries filed suit.FreckledTrout - Tuesday, October 29, 2019 - link
Same... this is a TSMC win.Smell This - Tuesday, October 29, 2019 - link
You're a "very stable genius" ain't yahs?
[snicker]
""In regards to today's quashing of the disputes, we reached out to Patrick Moorhead, Founder, President & Principal Analyst at boutique tech analyst firm Moor Insights & Strategy, for some perspective:
"I believe this agreement is good for the industry as more companies sharing patents and IP accelerates innovation. As GlobalFoundries initiated the lawsuit and is in less diverse markets, I think it likely got the better deal."
In either case, burying the hatchet is a win for both companies, as the litigation would have drug on for years across several jurisdictions. ""[/i]
Smell This - Tuesday, October 29, 2019 - link
sorry about that *tag* ---!$$@@!*&! no edit feature!
FreckledTrout - Tuesday, October 29, 2019 - link
Only time will tell if GF got the better deal. They will need to come up with billions to start working on a new EUV node like 7nm or 5nm. If they do in fact do that then maybe this works out for them however I don't think they will so I truly believe TSMC got the better deal.Why are you calling me names? This is an adult site, please if you are not one at least act like one.
Smell This - Tuesday, October 29, 2019 - link
Play the victim card, much?
You were called-out on your FUD __ you swung and missed.
bji - Tuesday, October 29, 2019 - link
Do you even know the meaning of the term 'FUD'? Doesn't seem like it ...Eliadbu - Wednesday, October 30, 2019 - link
they will need tens of billions to compete in the leading node, they might try to compete not in the leading node but 1 or 2 gens behind in the future but as is seems the race is over for them, in this industry you either make it big get huge profit and lead on or you fall behind, from dozens of manufactures in the leading node in the past we are left with just 3 today.Dragonstongue - Tuesday, October 29, 2019 - link
IMHO good on them BOTH, the way the world is heading, we ALL or at least those making the products we want/need have to and should be working together in one way or another, even if it means they are "competitors" making their own sales and such.Now only if Intel Nvidia Apple and such would do the same,
beginner99 - Tuesday, October 29, 2019 - link
That is a surprising quick and nice ending to this topic.mkozakewich - Wednesday, October 30, 2019 - link
This whole thing has been some kind of weird chess play. I guess it might have been trying to get access to those patents so it could keep up with its current status of "slightly older architectures" a few years from now? If so, I'm actually impressed with their forwardthinkingness. Or maybe they hoped for some kind of cash settlement, which would have been petty. I do wonder if there was something behind the scenes, like if AMD wanted to jump entirely to TSMC but couldn't because of patent issues, and somehow negotiated with GloFo to have this all done. (We'll hear about it in the next while, if so.)Santoval - Wednesday, October 30, 2019 - link
I don't think we need to be lawyers to deduce that this was the true purpose of GloFo's lawsuit against TSMC all along. Since they apparently own less (and/or less important) patents than TSMC this should be a big win for them.Zoolook13 - Friday, November 1, 2019 - link
It's most likely part of putting GF on the market, to remove any uncertainties that might have existed, now GF has been slimmed and any legal question-marks have been straightened out for any potential buyers.Matthias B V - Saturday, March 6, 2021 - link
Was thinking since last year they should have licensed 7nm TSMC as it is DUV and could probably be done on their scanners.Since PS5 / XBOX / Navi1+2 / Zen 2+3 all run on this node it would benefit everyone:
- AMD sells more and gets market
- TSMC can't sell more at max cap. so they don*t lose business so they could make extra $
- GF can sell 7nm
- Customers get products
AMD has to get dual source. I love Mr. Su as a CEO but I have the feeling she is not agressive enough in taking a risk to get market share and too focused on margins - which was ok when AMD was starved and sucess was not sure but now I think market and mind share is more improtant than max. margin rather than have a good margin and more share as share is the base for future business.