TSMC’s Speciality Technologies(techtaiwan.com)
techtaiwan.com
TSMC’s Speciality Technologies
https://techtaiwan.com/20210816/tsmc-speciality-technology/
31 comments
This seems like a more gentle way to say that innovation in the sector is slowing a bit. Which would account for why "mature" nodes will be producing more revenue than the latest developments coming online.
28 nm was a turning point in cost per transistor. Previously, unit cost of transistors declined with gate lengths, but smaller finfet nodes and EUV add enough complexity that cost per transistor is actually rising on more advanced nodes.
The good thing about this is that capacity in "cutting-edge" fans will be lower and lower every year as more products find a "home" node that balances price and performance.
The good thing about this is that capacity in "cutting-edge" fans will be lower and lower every year as more products find a "home" node that balances price and performance.
I'm not sure where that idea came from. The last we saw was that N5 was slightly more expensive than N7, but at the time that info came out nobody but Apple could run N5 except for super expensive proto runs that even then were under tight competition. And even then, N5 is still cheaper than N10. Everything else is cheaper for gate count as you go down in gate size, and we should expect N5 to drop in price to under N7.
https://www.tomshardware.com/news/tsmcs-wafer-prices-reveale...
Maybe this idea came from the industry not fully internalizing the consolidation in the space and how much competition for wafers affects the initial price of a new node?
https://www.tomshardware.com/news/tsmcs-wafer-prices-reveale...
Maybe this idea came from the industry not fully internalizing the consolidation in the space and how much competition for wafers affects the initial price of a new node?
These processes are explicitly not about logic parts like CPUs or GPUs, but analog/power/mixed-signal parts, which don't benefit from smaller node sizes. Many of these parts are made on 0.12/0.18 µm or even bigger nodes because anything smaller is pointless.
> Generally, TSMC’s speciality technologies portfolio includes MEMS, CMOS Image Sensor, Embedded NVM, RF, Analog, High Voltage, and BCD-Power processes.
> Generally, TSMC’s speciality technologies portfolio includes MEMS, CMOS Image Sensor, Embedded NVM, RF, Analog, High Voltage, and BCD-Power processes.
"Innovation" in the sector always came at tremendous cost. The startup cost for the foundry for a new node has been increasing every generation (requires more wafers and time to amortize over), and the cost per wafer (on a production basis, not including the amortized development cost) has been increasing over time as well (and you're going to be competing with AMD, Nvidia, and Apple for fab space... those are deep pockets to fight against). And finally, the (per customer) development cost at each node is becoming more expensive as well.
All of this means theres tremendous value in exploiting mature processes. And it's going to continue to be this way for quite a long time based on all the trendlines and projections that the big players put out.
All of this means theres tremendous value in exploiting mature processes. And it's going to continue to be this way for quite a long time based on all the trendlines and projections that the big players put out.
It's not just the nodes AFAIK. There's huge demand for EUV machines but they are lower margin products for the company that makes them.
I don't really understand this business, or any business really, so don't take my word for it
I don't really understand this business, or any business really, so don't take my word for it
I have no digits, but it seems that at least bruteforcing the volumes is in ways with how much chips Apple + AMD + Qualcomm ship.
EUV productivity is a big problem. Currently machines are stopped every few shifts for cleaning tin from mirrors. Even without that, light sources never managed to live to the original power output promise.
There are few known ways to produce EUV light besides tin plasma.
One is in using some plasmonic quantum processes, I can't even fathom how to describe, to produce microwatts of EUV. The other is a free electron laser to produce potentially kilowatts of it non-stop.
The later approach requires to build the whole fab around a synchrotron, a commitment currently nobody can take even with current surreal capex numbers.
EUV productivity is a big problem. Currently machines are stopped every few shifts for cleaning tin from mirrors. Even without that, light sources never managed to live to the original power output promise.
There are few known ways to produce EUV light besides tin plasma.
One is in using some plasmonic quantum processes, I can't even fathom how to describe, to produce microwatts of EUV. The other is a free electron laser to produce potentially kilowatts of it non-stop.
The later approach requires to build the whole fab around a synchrotron, a commitment currently nobody can take even with current surreal capex numbers.
Equipment for ion implantation already includes mini accelerators [1] [2]. The semiconductor equipment industry in general has many machines that feel like they came out of a physics lab into a semi fab. For example, plasma dry-etching or deep-reactive ion etching. EUV litho is just one of many very interesting problems--currently the bottleneck so it's talked about a lot.
[1] https://en.wikipedia.org/wiki/Ion_implantation [2] https://www.axcelis.com/products/purion-xe-series-high-energ...
edit: add link to cool ion implanter machine pics
[1] https://en.wikipedia.org/wiki/Ion_implantation [2] https://www.axcelis.com/products/purion-xe-series-high-energ...
edit: add link to cool ion implanter machine pics
Oh they could do the latter, they just need to trust it's gonna work.
There is such a facility, not far away from me. Though not that much free space around it anymore.
https://www.xfel.eu/
https://openstreetmap.de/karte.html?zoom=17&lat=53.58825&lon...
https://www.xfel.eu/
https://openstreetmap.de/karte.html?zoom=17&lat=53.58825&lon...
Maybe we should rebuild the Large Hadron Collider as the core of a silicon fab/s
Particle pipes for semi fabs (a la oil pipelines) would be a cool bit of retrofuturist-y world building
I disagree. A lot of the upcoming requirements of transistors just do not need small transistor nodes. For example, for RF applications, power handling needs simply require larger gate lengths. There is no way you can use a 7nm finfet to realize some RF applications.
Innovation in semiconductors takes many forms, not just pushing to lower transistor nodes.
Innovation in semiconductors takes many forms, not just pushing to lower transistor nodes.
What's impressive is that TSMC chugs along doing great work with little submarine marketing on tech Twitter or HN. Same for ASML. No filler blogs about some new framework here.
There's a huge underbelly of companies nobody ever hears about holding the technological torch in various industries. Some obscure mittelstand firm in Germany will be the world leader in ball bearings or thermostats or something like that, in a thousand little niches.
TSMC is admittedly must bigger but like these types of firms, they don't need to tell everyone how important they are.
TSMC is admittedly must bigger but like these types of firms, they don't need to tell everyone how important they are.
There was a story about China finally having the metallurgical skills to make domestic ball point pens with a certain longevity that was considered a major milestone.
Little things like that seem small but are huge, because the same skills for that type of precision can also be applied to military and other areas
Little things like that seem small but are huge, because the same skills for that type of precision can also be applied to military and other areas
Marketing it the common person if you don't make the things they use isn't generally helpful. Intel has name recognition, but it isn't clear how many people calling Dell/hp/... to buy a new computer actually care if they get Intel vs AMD. Dell/hp/... need to market because those are the names the average person can influence, you call one and not the others (or go to local electronics store and choose one from several choices).
Of course I'm not sure that intel really cares that the common person knows their name. All the marketing could be aimed at the Cxx of Dell/HP/... and the reason the rest of use see those ads is because they can't be sure the few people they target.
Of course I'm not sure that intel really cares that the common person knows their name. All the marketing could be aimed at the Cxx of Dell/HP/... and the reason the rest of use see those ads is because they can't be sure the few people they target.
> Some obscure mittelstand firm in Germany will be the world leader in ball bearings or thermostats or something like that, in a thousand little niches.
The German reputation for precision engineering is certainly well-deserved. Part of this is because their economic system is designed to support these kinds of businesses. Surgical equipment is another great example.
The German reputation for precision engineering is certainly well-deserved. Part of this is because their economic system is designed to support these kinds of businesses. Surgical equipment is another great example.
Marketing to the general public has two purposes, selling consumer products and creating hype for your stock. Did the latter purpose, you need really good hope (Elon Musk) to get the public to YOLO on you. Otherwise you will just desk with experienced investors who will usually want more details than what an ad provides and do the own DD.
What is amazing is TSMC is just the integrator of all the technology from different technology company.
No. TSMC spend enormous amount of money developing their processes. To implement them in the fab, they need and thus cooperate (and co-develop) with vendors. But TSMC is far from just integrating stuff.
As an example, TSMC design their own gates, including how they are structured and created on the die. This means developing each process step.
As an example, TSMC design their own gates, including how they are structured and created on the die. This means developing each process step.
That is just not true. TSMC got ahead of the pack by developing their own tech again and again and again... Without their own tech, they might as well be UMC.
But the only reason why they would need to is for hiring. If they needed to for their business model they would but they don't sell to you or I.
I work for a hedge fund, we have almost no online presence because it just isn't relevant to what the company gets up to.
I work for a hedge fund, we have almost no online presence because it just isn't relevant to what the company gets up to.
TSMC is 100% B2B (business to business).
B2B market dynamics have almost nothing in common with consumer markets. Sales for example is 100% "direct" - you knock on doors and visit to close deals in-person. You can use Zoom et al. but it's only effective once the sale is pretty mature. Most selling is highly technical and requires a face-to-face conversation plus hands-on "proof of value" in "real-time". Such conversations "weave and dodge" like a boxing match so you have to think, respond and sell on your feet. A passive web page or data sheet PDF is merely the 1% of the conversation required to close a deal.
You (in both marketing and sales roles) literally know all the companies in the competitive and customer space, however.
As a result, broadcast advertising is pretty much a waste of money because EVERYONE in such industries has a black book with every person who matters to their product purchase decisions in it and it's not that thick either. The marketing and selling process has so many intangibles and goodwill that this is the only way it works. This is also why direct sales is so high paying as a profession.
In just one area we sell into, there are literally 100 people on the planet who know anything about the market space. The market cap happens to be $10B-$25B depending on what you include so it's well worth the effort.
I literally know ALL my competitor and customer employees by name because I've worked with most of them over a career.
So if I want to sell something, I just contact the right people, arrange a meeting and if my pitch, process and product are good, I'll get the sales. It does require work along the way but once you have it down, it's a lot of "turning the crank" mentally, emotionally and socially.
We're rolling out a new piece of semiconductor manufacturing equipment and that exactly how I closed our first sales.
Of course the product sales cycle is anywhere from 3-18 months because it's heavy capital equipment. Sign-off is generally at a VP or CEO level and an elaborate buying process exists.
Our unit prices range from $30K-$1.5M so this is never going to be something you buy with e-commerce and a credit card. At least not until our installed base is big enough. But that's only companies like TSMC - they still nit and negotiate then so it's still not a "credit card" simple purchase.
The point of all of this is ALL of TSMC's customers are just like ours (TSMC is also one of our customers): every TSMC foundry deal is starting at 6-figures and if you can't play that ante, you'll never be a TSMC customer.
Again they don't have that many customers so most marketing and sales is utterly different from anything seen in consumer markets.
And everyone comes to them because of their brand value which has been top-notch for decades now - they literally have to turn business away.
B2B market dynamics have almost nothing in common with consumer markets. Sales for example is 100% "direct" - you knock on doors and visit to close deals in-person. You can use Zoom et al. but it's only effective once the sale is pretty mature. Most selling is highly technical and requires a face-to-face conversation plus hands-on "proof of value" in "real-time". Such conversations "weave and dodge" like a boxing match so you have to think, respond and sell on your feet. A passive web page or data sheet PDF is merely the 1% of the conversation required to close a deal.
You (in both marketing and sales roles) literally know all the companies in the competitive and customer space, however.
As a result, broadcast advertising is pretty much a waste of money because EVERYONE in such industries has a black book with every person who matters to their product purchase decisions in it and it's not that thick either. The marketing and selling process has so many intangibles and goodwill that this is the only way it works. This is also why direct sales is so high paying as a profession.
In just one area we sell into, there are literally 100 people on the planet who know anything about the market space. The market cap happens to be $10B-$25B depending on what you include so it's well worth the effort.
I literally know ALL my competitor and customer employees by name because I've worked with most of them over a career.
So if I want to sell something, I just contact the right people, arrange a meeting and if my pitch, process and product are good, I'll get the sales. It does require work along the way but once you have it down, it's a lot of "turning the crank" mentally, emotionally and socially.
We're rolling out a new piece of semiconductor manufacturing equipment and that exactly how I closed our first sales.
Of course the product sales cycle is anywhere from 3-18 months because it's heavy capital equipment. Sign-off is generally at a VP or CEO level and an elaborate buying process exists.
Our unit prices range from $30K-$1.5M so this is never going to be something you buy with e-commerce and a credit card. At least not until our installed base is big enough. But that's only companies like TSMC - they still nit and negotiate then so it's still not a "credit card" simple purchase.
The point of all of this is ALL of TSMC's customers are just like ours (TSMC is also one of our customers): every TSMC foundry deal is starting at 6-figures and if you can't play that ante, you'll never be a TSMC customer.
Again they don't have that many customers so most marketing and sales is utterly different from anything seen in consumer markets.
And everyone comes to them because of their brand value which has been top-notch for decades now - they literally have to turn business away.
they're big in industry trade shows. but their clients are hardly readers of hn or tc so really not much use for them to get vanity articles published.
I doubt they do this. Why do they need it? As if somebody doesn't know who the real kingmaker of the industrial world is.
If the world will be left without modern semi manufacturing, it will rm -rf half of the NASDAQ by default.
I feel it's the Silicon Valley itself which needs to acknowledge how insignificant they are in the bigger picture of things.
If I were some big "tech" company head like of Facebook, Amazon, Google, I would've at least gave a look how much the microchip supply chain can trip their companies.
Tan Hock Eng has over the last 5 years bought almost every major network gear chipmaker, and is now holding the entirety of the "cloud" industry under gunpoint for example. The "cloud" guys might be now spending more money on switches than on CPUs.
Now imagine, somebody else will rollup something equally critical.
If the world will be left without modern semi manufacturing, it will rm -rf half of the NASDAQ by default.
I feel it's the Silicon Valley itself which needs to acknowledge how insignificant they are in the bigger picture of things.
If I were some big "tech" company head like of Facebook, Amazon, Google, I would've at least gave a look how much the microchip supply chain can trip their companies.
Tan Hock Eng has over the last 5 years bought almost every major network gear chipmaker, and is now holding the entirety of the "cloud" industry under gunpoint for example. The "cloud" guys might be now spending more money on switches than on CPUs.
Now imagine, somebody else will rollup something equally critical.
More money on switches?! Are you serious or exaggerating?
Broadcom has a stranglehold on both chips for the top of the rack, and multi-terabit ones.
Intel makes very timid inroads into high end switching chips with Tofino, and that's it.
Intel makes very timid inroads into high end switching chips with Tofino, and that's it.
Broadcom have been busy but there's still choice: https://elegantnetwork.github.io/posts/A-Summary-of-Network-...
A choice they are allowed to make N years from now, when their current Broadcom terms expire, otherwise they forfeit their preferred pricing.