Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/how-many-patents-and-copyrights-does-tsm-have-68860.18413/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

How many patents and copyrights does TSM have? 68860

Arthur Hanson

Well-known member
I know TSM has a large trove of IP, do they rely more on secrecy or legal protections for the vast amount of IP they have at all levels of the semiconductor process including specialty materials they must have created? Does TSM compartmentalize IP as much as legal protections? With IP becoming ever more valuable, will new methods and strategies come out to protect IP including deliberate subterfuge? I had to resort to doctored part numbers to counteract IP theft from my own company. This was recommended by a major supplier that said theft was rampant by some companies that put projects out to bid just to get information.
 
I think you're making too much of corporate patents. Intel has over 70,000 active patents. IBM has been the top recipient of US patents for 26 years. At the moment neither company is a paragon of market success. Execution and quality are the foundations of product leadership.
I have no doubt the value of TSM patents is far higher than Intel's. Also, working with IBM equipment didn't impress me at all compared to the competition.
 
And I'm sure your opinion is based on an objective analysis of patents from both companies, which are easily searchable online. :ROFLMAO:

Which IBM equipment?
ATMs, Data Communications, Networking and Storage as a field tech for brief periods at a range of companies, always seemed a step behind. I also did a job for a customer that I bid very, very fat I thougt and made my boss very, very happy and at the end the customer showed me IBM wanted to charge over ten times what we bid and made a staggering margin on. Also another questionable practice when it came to safety. I must note, I changed fields years ago to trade stocks.
 
ATMs, Data Communications, Networking and Storage as a field tech for brief periods at a range of companies, always seemed a step behind. I also did a job for a customer that I bid very, very fat I thougt and made my boss very, very happy and at the end the customer showed me IBM wanted to charge over ten times what we bid and made a staggering margin on. Also another questionable practice when it came to safety. I must note, I changed fields years ago to trade stocks.
I don't know how to react to this if the information is years old. I'm not much of an IBM fan lately myself, but this non-specific stuff is not convincing of anything about the current company.
 
Last edited:
Remember after GF dismantled their R&D they tried to shake down the entire industry with some classic IBM-style patent trolling? TSMC countersued and it was settled out of court, which Tom Caulfield said "secures GF’s ability to grow and is a win for the entire semiconductor industry which is at the core of today’s global economy". That settlement money has since presumably been wired into offshore bank accounts of emirati princes, because GF has done nothing interesting technologically and relies entirely on financial engineering to dupe investors. Their last CFO lasted 2 months because he probably took one look at the books, realized he was their patsy, and then almost came to fisticuffs with TC before quitting.

This is all to say patents are fake and China is right to abuse a system that has long been cynically abused by the same western companies that wrote the laws.
 
This is all to say patents are fake and China is right to abuse a system that has long been cynically abused by the same western companies that wrote the laws.
Patents are not all fake, whatever that means, but US patent laws do need an update, in my opinion. For one thing, patent trolling should be illegal, meaning patents should be owned only by companies which make or sell products relying on the patents. Defending against legitimate infringement is not trolling. Also, patent filing should include a requirement for a prototype. Finally, patent filings should require proposed infringement tests. Opportunistic accusations of infringement should be heavily penalized.

In some fields this stuff is easy, but not for semiconductors.

While China may abuse other countries/companies patents, they seem to want everyone to respect theirs. Example, in 5G technology and their submissions to industry standards groups. If China wants to be respected and included, they need to clean up their IP act.
 
I have no doubt the value of TSM patents is far higher than Intel's. Also, working with IBM equipment didn't impress me at all compared to the competition.
Based on what? Intel makes the fastest single-threaded processor in the world, and has the highest performing process, despite being "a node behind". I'm not sure how one could measure this in the first place, but let's not discount Intel's achievements so easily. Plus, by all indications they are making excellent progress with the nodes we don't know everything about.

IBM makes some of the most important equipment in the world. I mean, their mainframes average milliseconds of downtime per year, and are the backbone of many banks and financial institutions. Plus, they are on the leading edge of Quantum Computing. Not sure why they get so much dislike in this forum, because while they aren't what they were (no company is as dominant as IBM was), they make a ton of money, and have some really great technology, with incredible reliability and capability. Everyone is so fond of saying what would happen if TSMC were not a thing. What do you think would happen if IBM mainframes were suddenly taken away? If you think it wouldn't be a big thing, maybe look at who uses these machines, and how important they are.

Of course, in both cases, other companies would fill the gaps, my point being, IBM is a very important part of what goes on every day, even if it's not very visible.
 
Based on what? Intel makes the fastest single-threaded processor in the world, and has the highest performing process, despite being "a node behind". I'm not sure how one could measure this in the first place, but let's not discount Intel's achievements so easily. Plus, by all indications they are making excellent progress with the nodes we don't know everything about.

IBM makes some of the most important equipment in the world. I mean, their mainframes average milliseconds of downtime per year, and are the backbone of many banks and financial institutions. Plus, they are on the leading edge of Quantum Computing. Not sure why they get so much dislike in this forum, because while they aren't what they were (no company is as dominant as IBM was), they make a ton of money, and have some really great technology, with incredible reliability and capability. Everyone is so fond of saying what would happen if TSMC were not a thing. What do you think would happen if IBM mainframes were suddenly taken away? If you think it wouldn't be a big thing, maybe look at who uses these machines, and how important they are.

Of course, in both cases, other companies would fill the gaps, my point being, IBM is a very important part of what goes on every day, even if it's not very visible.
The IBM bashing comes from IBM claiming technical superiority off the back of technological failures/deadends. Biaxial strain, gate first, SOI, claiming they are ahead of others because they showed of a 2nm technology demonstrator in like 2019. TLDR I like the tech, but I don't like when they act smug while pretending their process technology is anything beyond a technological novelty.

I won't comment on the design side very much as it is outside my wheelhouse, but it all seems very brute force-ish to me. Oh boy they maxed out the reticle limit of four dies with huge caches, and run dozens of them as backups inside a property software blackbox for maximum reliability. Effective? Yes. Cool? Also yes. Elegant? Not so much. I suspect the main reason IBM has this market "on lock" is because it is too small for other firms to bother wasting the money trying to displace IBM and their entrenched ecosystem. Put another way, I feel like a NVIDA or an Apple could do it, but the cost of doing so would not justify the relatively meager profits they would capture even with 100% MSS. The money is simply better spent bludgeoning their competitors.

As for quantum they can squak about it all they want. But any lead they want to claim they have doesn't matter until the tech becomes viable. For example, did it matter that the Nissan Leaf was the first high volume BEV?
 
The IBM bashing comes from IBM claiming technical superiority off the back of technological failures/deadends. Biaxial strain, gate first, SOI, claiming they are ahead of others because they showed of a 2nm technology demonstrator in like 2019. TLDR I like the tech, but I don't like when they act smug while pretending their process technology is anything beyond a technological novelty.

I won't comment on the design side very much as it is outside my wheelhouse, but it all seems very brute force-ish to me. Oh boy they maxed out the reticle limit of four dies with huge caches, and run dozens of them as backups inside a property software blackbox for maximum reliability. Effective? Yes. Cool? Also yes. Elegant? Not so much. I suspect the main reason IBM has this market "on lock" is because it is too small for other firms to bother wasting the money trying to displace IBM and their entrenched ecosystem. Put another way, I feel like a NVIDA or an Apple could do it, but the cost of doing so would not justify the relatively meager profits they would capture even with 100% MSS. The money is simply better spent bludgeoning their competitors.

As for quantum they can squak about it all they want. But any lead they want to claim they have doesn't matter until the tech becomes viable. For example, did it matter that the Nissan Leaf was the first high volume BEV?
IBM isn't after elegance, they are after making the best products they can for their customers, or at least good enough to get them to upgrade enough.

Apple and NVIDIA wouldn't have the first idea how to make something that reliable, or at least none of their products even remotely suggest it. It's not that small a market, when you consider the attachment rate of products, including software, they get with those sales. Of course, it's all relative, but when you're taking about many billions of dollars a year, it's not insignificant. Compared to the iPhone market, it is small though. But, Apple has pursued smaller markets, and trying to compete with IBM in big iron is not an easy task; they are serious machines for serious companies. Like the MacIntosh????

Intel/HP kind of took a swipe at it with IA-64, but failed miserably, and I'd venture to say Intel/HP has a better chance of competing with IBM than Apple and NVIDIA. The risk of competing, and the high probability of failure is daunting to any company competing with IBM. Their mainframes are extraordinary machines, and have the software (z/OS) to go with it. MacOS? Yeah, no.

I'm just pointing out with Quantum that IBM does have patents in what could be a very important technology. Or it might be nothing. But, so might many of TSMC's patents. And it's not like IBM isn't making a lot of money, they are. And it's not like all their products are unsuccessful; they are a very successful company. If you judge them only by the fab technology, yeah, I can agree, but they are a weird company that has tentacles in a lot of things, and most are very successful.

I just read Rapidus saying their 2nm is going to be REALLY expensive to make. IBM doesn't care so much about costs, since their servers are so superior they can ask a lot of money. But, who else is going to want to spend 10x what current fab technology costs? I don't know how big a market their tech will be playing in with such high costs.
 
I think the point is TSMC has a lot of IP, which should not be surprising. The value of IP is highly variable, and number of patents is not a great indicator on its own. But no one should doubt TSMCs status as an innovator and IP generator.
 
IBM isn't after elegance, they are after making the best products they can for their customers, or at least good enough to get them to upgrade enough.
I don't disagree that brute force is good as long as it does the job at a satisfactory cost. But as an engineer I do love an elegant solution.
Apple and NVIDIA wouldn't have the first idea how to make something that reliable, or at least none of their products even remotely suggest it. It's not that small a market, when you consider the attachment rate of products, including software, they get with those sales. Of course, it's all relative, but when you're taking about many billions of dollars a year, it's not insignificant. Compared to the iPhone market, it is small though. But, Apple has pursued smaller markets, and trying to compete with IBM in big iron is not an easy task; they are serious machines for serious companies. Like the MacIntosh????
My point was I think they have the talent that is smart enough and capable enough to do it (even if they will never try to do it for a laundry list of reasons). Which isn't to say their designers are dumb, but rather my point was they weren't on the same level of mastery at the craft. But as you said it would be alot of hard work to create designs to fill that niche. Hard enough that the ROI would never be worth it. If IBM had a monopoly on mobile or AI, folks would try to break it. But for banks that will prob never switch off no matter the comp's merits, it is too steep a hill to climb for how small it is (relatively speaking anyways).
I'm just pointing out with Quantum that IBM does have patents in what could be a very important technology. Or it might be nothing. But, so might many of TSMC's patents. And it's not like IBM isn't making a lot of money, they are. And it's not like all their products are unsuccessful; they are a very successful company. If you judge them only by the fab technology, yeah, I can agree, but they are a weird company that has tentacles in a lot of things, and most are very successful.
They do. But it's not like Google, intel, and a thousand other folks don't also make advanced (at least by the current primitive standards we have) quantum computers.
I just read Rapidus saying their 2nm is going to be REALLY expensive to make. IBM doesn't care so much about costs, since their servers are so superior they can ask a lot of money. But, who else is going to want to spend 10x what current fab technology costs? I don't know how big a market their tech will be playing in with such high costs.
I think that was in reference to Japan's current most advanced processes (I think Toshiba's 40nm). Either way that is still crazy, since I don't think N3’s structural costs are more than 10x 40nm, but I could be wrong as 40nm was a LONG time ago.
 
Last edited:
I don't know how to react to this if the information is years old. I'm not much of an IBM fan lately myself, but this non-specific stuff is not convincing of anything about the current company.
It's also the culture, just like ATT going downhill. Cultures get a successful mindset that works until times change and they don't. TSM has been on the bleeding edge far longer than most.
 
I don't disagree that brute force is good as long as it does the job at a satisfactory cost. But as an engineer I do love an elegant solution.

My point was I think they have the talent that is smart enough and capable enough to do it (even if they will never try to do it for a laundry list of reasons). Which isn't to say their designers are dumb, but rather my point was they weren't on the same level of mastery at the craft. But as you said it would be alot of hard work to create designs to fill that niche. Hard enough that the ROI would never be worth it. If IBM had a monopoly on mobile or AI, folks would try to break it. But for banks that will prob never switch off no matter the comp's merits, it is too steep a hill to climb for how small it is (relatively speaking anyways).

They do. But it's not like Google, intel, and a thousand other folks don't also make advanced (at least by the current primitive standards we have) quantum computers.

I think that was in reference to Japan's current most advanced processes (I think Toshiba's 40nm). Either way that is still crazy, since I don't think N3’s structural costs are more than 10x 40nm, but I could be wrong as 40nm was a LONG time ago.
I think their mainframes are very elegant, but maybe we're looking at different aspects of it.

If we look at brute force, some of it is, based on their market, they can do that more/better than anyone else. I almost think their customers expect them not to skimp on costs, given how expensive their computers are.

But, my feeling is you are looking at it from one perspective, and I from another, and we probably both have merits in how we're looking. For example, their L2 cache/L3 cache arrangement, to me, is extremely elegant and completely innovative. I also think that having a computer that can detect errors in every instruction that is run, and take that processor off-line, and replace it without the user ever noticing is elegant. I think being down a few fractions of a second, per year, on average, is very elegant. Well, maybe not elegant, but really impressive. I think being able to remove a memory stick and having the computer replace it on the fly with a spare, with no down time, is elegant.

I think an OS that is so incredibly robust, it can make changes without a reboot, or without downtime, is elegant.

Of course, you guys have a good point with their manufacturing nonsense, and boasting. But, they do have nice tech too. Their POWER line is good at what it does, and as you can tell, I have a ton of respect for their mainframes and what they do.

We'll have to agree to disagree on other companies doing it, but then I don't even think we disagree that much. It's theoretically possible, but the biggest impediment is the super, super reliable OS. Apple hasn't come close. Neither has MS. No one has. I don't think the hardware is the biggest issue, although IBM does some advanced stuff no one else has. But, I agree, that would be far from impossible. But, the OS, and the way the whole system works together? That's a big canyon to cross.
 
I think their mainframes are very elegant, but maybe we're looking at different aspects of it.

If we look at brute force, some of it is, based on their market, they can do that more/better than anyone else. I almost think their customers expect them not to skimp on costs, given how expensive their computers are.

But, my feeling is you are looking at it from one perspective, and I from another, and we probably both have merits in how we're looking. For example, their L2 cache/L3 cache arrangement, to me, is extremely elegant and completely innovative. I also think that having a computer that can detect errors in every instruction that is run, and take that processor off-line, and replace it without the user ever noticing is elegant. I think being down a few fractions of a second, per year, on average, is very elegant. Well, maybe not elegant, but really impressive. I think being able to remove a memory stick and having the computer replace it on the fly with a spare, with no down time, is elegant.

I think an OS that is so incredibly robust, it can make changes without a reboot, or without downtime, is elegant.

Of course, you guys have a good point with their manufacturing nonsense, and boasting. But, they do have nice tech too. Their POWER line is good at what it does, and as you can tell, I have a ton of respect for their mainframes and what they do.
Once again proving why I shouldn't be making many comments on architecture ;). The elegance aspect was mostly me thinking about it from a throwing silicon at the problem aspect. However I did not know about the instruction checking thing. In short I yield to you on this point good sir. Software wise I don't have much to say positive or negative (other than I'd prefer if it was opensource, but I think that is a universal compliant for anything proprietary). As for the fab tech I agree there is alot of cool stuff. their trench eDRAM, I'm a fan of FDSOI (even if I know it is a niche thing). I think there was also certainly a romanticism to 2000s/2010s IBM semiconductor until GF dropped out at 7LPP.
We'll have to agree to disagree on other companies doing it, but then I don't even think we disagree that much. It's theoretically possible, but the biggest impediment is the super, super reliable OS. Apple hasn't come close. Neither has MS. No one has. I don't think the hardware is the biggest issue, although IBM does some advanced stuff no one else has. But, I agree, that would be far from impossible. But, the OS, and the way the whole system works together? That's a big canyon to cross.
We agree then. It won't happen, but that is because there is not enough reason to force the issue. My point as you correctly surmised was that IBM has everything tuned for a niche, but I don't think their engineers are magically smarter than everybody else in the industry.
 
IBM isn't after elegance, they are after making the best products they can for their customers, or at least good enough to get them to upgrade enough.
IBM's z-series is a unique systems-level product-line. Very innovative technology at multiple architecture layers.
Apple and NVIDIA wouldn't have the first idea how to make something that reliable, or at least none of their products even remotely suggest it.
Apple does not design enterprise systems. Nvidia doesn't either, as HPC is a different market altogether than enterprise business systems.
It's not that small a market, when you consider the attachment rate of products, including software, they get with those sales. Of course, it's all relative, but when you're taking about many billions of dollars a year, it's not insignificant. Compared to the iPhone market, it is small though. But, Apple has pursued smaller markets, and trying to compete with IBM in big iron is not an easy task; they are serious machines for serious companies. Like the MacIntosh????
You're comparing Macs to the z-series? That's silly. Macs are used in business, especially in content creations applications for example, but only as workstations or work-group application servers.
Intel/HP kind of took a swipe at it with IA-64, but failed miserably, and I'd venture to say Intel/HP has a better chance of competing with IBM than Apple and NVIDIA. The risk of competing, and the high probability of failure is daunting to any company competing with IBM. Their mainframes are extraordinary machines, and have the software (z/OS) to go with it. MacOS? Yeah, no.
Itanium was not a success, but it was aimed at the 64bit RISC Unix systems (e.g. PA-RISC), not IBM's z-series. Since all but one of the 64bit RISC processors of 20 years ago went away (the exception being the Sun/Oracle SPARC CPU), I suppose if we squint we could call Itanium a kamikazi-like success. (I don't.) In the end AMD's x64 strategy finally killed Itanium, but you know that.

MacOS? Seriously? Nope, but it doesn't try to be. Mach for hardware interface and abstraction, a highly-customized FreeBSD Unix layer for OS guts, and the Mac user interface, file management, and networking? I don't understand why you're pressing this. These ingredients are not a recipe for highly-available enterprise servers, and they weren't meant to be.
I'm just pointing out with Quantum that IBM does have patents in what could be a very important technology. Or it might be nothing. But, so might many of TSMC's patents. And it's not like IBM isn't making a lot of money, they are. And it's not like all their products are unsuccessful; they are a very successful company. If you judge them only by the fab technology, yeah, I can agree, but they are a weird company that has tentacles in a lot of things, and most are very successful.
I look at IBM's strategy as defending their gross margin, which isn't that great, considering the value they add. It's 55%. Oracle's is about 73%, to compare them to another CPU/systems/application software/cloud vendor. Put another way, if I were choosing a stock to own, it would be Oracle before IBM.
 
Back
Top