Tuesday, October 13th 2020

NVIDIA Reportedly Moving Ampere to 7 nm TSMC in 2021

A report straight from DigiTimes claims that NVIDIA is looking to upgrade their Ampere consumer GPUs from Samsung's 8 nm to TSMC's 7 nm. According to the source, the volume of this transition should be "very large", but most likely wouldn't reflect the entirety of Ampere's consumer-facing product stack. The report claims that TSMC has become more "friendly" to NVIDIA. This could be because TSMC now has available manufacturing capacity in 7 nm due to some of its clients moving to the company's 5 nm node, or simply because TSMC hadn't believed NVIDIA to consider Samsung as a viable foundry alternative - which it now does - and has thus lowered pricing.

There are various reasons being leveraged at this, none with substantial grounds other than "reported from industry sources". NVIDIA looking for better yields is one of the appointed reasons, as is its history as a TSMC customer. NVIDIA shouldn't have too high a cost porting its manufacturing to TSMC in terms of design changes to the silicon level so as to cater to different characteristics of TSMC's 7 nm, because the company's GA100 GPU (Ampere for the non-consumer market) is already manufactured at TSMC. The next part of this post is mere (relatively informed) speculation, so take that with a saltier disposition than what came before.
That NVIDIA is looking to tier its manufacturing process across high-end and the rest of its product stack (with 7 nm for high-end and 8 nm for the rest of it) would become a headache for themselves and for consumers, should NVIDIA just have two suppliers for the same graphics products. There would likely be need for some changes in the power delivery designs, there are a range of new quality assurance tests that have to be taken for the new silicon, and NVIDIA would set itself up for legal troubles should they just silently update the manufacturing process on high-end models - not only would early adopters be understandably miffed about their product having evolved over time, as there could be some claims regarding 8 nm-based models being bought after the 7 nm ones are launched. And if NVIDIA were to put a sticker on retail boxes updating the 8 nm to 7 nm, well, then any user could just decline to purchase any 8 nm cards, and only look for the 7 nm versions, which might leave NVIDIA with a real immovable supply problem.

No. If this report checks out, NVIDIA will likely launch the newly produced top-end Ampere cards (we're thinking RTX 3090, RTX 3080 and RTX 3070) in 7 nm Super versions, taking a page from their RTX 20-series book. The introduction of a higher-performing product built in 7 nm within a whole new series would protect NVIDIA from legal troubles while allowing them to publicly announce the transition. This would keep early adopters "happy" in the respect that this is a whole new product launch - users would receive that much better than by feeling that they were beta testers for NVIDIA's tango with Samsung, as reception for the 20-series Super cards shows. The usage of this new process would also allow NVIDIA to improve performance further over the original 30-series cards, due to lower leakage and higher potential operational frequencies - perhaps in addition to NVIDIA's 20-series strategy of trickling down bigger GPU designs.
Source: DigiTimes
Add your own comment

119 Comments on NVIDIA Reportedly Moving Ampere to 7 nm TSMC in 2021

#1
P4-630
Previous reports pointed out that Nvidia also booked 5 nm production capacity at TSMC and this is probably meant for the upcoming "Hopper" GPUs rumored to release exactly 1 year after the Ampere models. The 2-year release schedule appears to have been reduced to 1 year, since AMD is strongly signaling that it is ready to match Nvidia’s gaming GPUs this year. Team red already has plans to release Navi 21’s successor produced on the 5 nm TSMC nodes at the end of 2021, and Nvidia is most likely not willing to give the competition a head start.

www.notebookcheck.net/Report-suggests-Nvidia-could-launch-improved-RTX-3000-GPUs-produced-on-TSMC-s-7-nm-node-in-2021.497489.0.html
Posted on Reply
#2
ZoneDymo
wonder how rich the people at TSMC must be, also poor samsung
Posted on Reply
#3
Vayra86
Yep...

I'm going to sit comfortably on this 1080 for another year, it just got confirmed. Stick that 10GB POS where the sun don't shine tyvm

Smoke > Fire. Always
Posted on Reply
#4
Xaled
LoL, so Nvidia just used Samsung to make TSMC feel jealous.
How about Nvidia use both Samsung and TSMC to respond to the supposed high- demand?
Posted on Reply
#5
ViperXTR
same for me i guess, had to hold off for now
Posted on Reply
#6
Xex360
Xaled
LoL, so Nvidia just used Samsung to make TSMC feel jealous.
How about Nvidia use both Samsung and TSMC to respond to the supposed high- demand?
Don't believe so, TSMC has more capacity more that Apple and others are moving to 5nm, nVidia gambled with Samsung and hopefully will pay the price for their stupid pricing.
Posted on Reply
#7
Chrispy_
If you read between the lines, the headline could be "Nvidia so dissatisfied with Samsung's 8nm that they're already planning to abandon it halfway through the current generation"

It doesn't inspire confidence in the 3000-series on sale right now though; Just another thing to add to the 3000-series launch fiasco...
Posted on Reply
#8
EarthDog
Chrispy_
If you read between the lines, the headline could be "Nvidia so dissatisfied with Samsung's 8nm that they've already made plans to abandon it this generation"

Doesn't inspire confidence in the 3000-series people can buy right now, though....
That's certainly one way to think about it.
Posted on Reply
#9
Smaeili
Would the smaller manufacturing process bring any improvement to thermals/power?
Posted on Reply
#10
Ashtr1x
Not going to believe this. How much of an improvement will be TSMC 7nm process will do ? It's not like it will become super fast or clock high, at max it will lower the power consumption, everyone knows that since Pascal this b.s Clock Boost came so I'm thinking the TSMC shift won't magically make the GPUs boost to higher clocks, look at Ryzen 5000, it's on 7nm EUV, 7NP it didn't change the clock speed and we know Ryzen 7nm has been pushed to max out of box, and add that boosting behavior where the clocks change like GPUs. And that VRM component disaster of the AIBs on the Ampere is not at all because of the 8nm node but rather BOM.

TSMC this TSMC that, this Intel screw up gave them too much hype even to the MSM Shill media, a big joke when you actually watch der8aur video of this whole nm marketing drama. And see the Apple A series processors, it's like they came from outer space in Anandtech's Spec scores and etc, but on Load the power consumption is insanely high and then on top you have the Qcomm vs Apple Application/Software performance videos and benches all over the Youtube to tell the real life performance / bang for buck / comparison etc. And to be honest, AMD only managed to beat the Skylake uArch a 5 year old design and 14nm node from 7 years ago in 2020 with Ryzen 5000.

Also Samsung is mentioning their upcoming non custom design (no M cores unlike prev Exynos, similar to the Qcomm post 820, since 820 is the only custom they did after that 810 big fiasco) Exynos 1080 is on 5nm node. So losing up a big contract like Nvidia is insanely bad to them in this pureplay fab industry, as we move to one and only one corporation doing that.
Posted on Reply
#11
ratirt
Now that's some sort of news and people should have waited with their purchase. I'm glad I did.
Although I've got a problem with this 8nm Samsung and now 7nm TSMC. At first, NV went with the Samsung 8nm because it is cheaper and matured and wanted to skip the 7nm TSMC because it was much more expensive. I mean that's what the news is. Now NV is going with TSMC's 7nm because? ........ Is it because NAVI is coming and the numbers that have been shown by AMD are correct so it means NV need a node change to make some more performance out of the Ampere cards? So many unknowns.
Really interesting.
Posted on Reply
#12
Vya Domus
Story has it Nvidia was pissed about TSMC's prices so they "threatened" them by moving to Samsung in an attempt to drive their prices lower but they didn't budge even after that happened, so now they're coming back with their tail between their legs.

Samsung's nodes are inferior, always have been, even for mobile. That doesn't mean they shouldn't be used but not for 600+ mm^2 behemoths.
Posted on Reply
#13
dorsetknob
"YOUR RMA REQUEST IS CON-REFUSED"
Vya Domus
Story has it Nvidia was pissed about TSMC's prices so they went to Samsung in an attempt to drive their prices lower but they didn't budge even after that happened, so now they're coming back with their tail behind their legs.
Be glad that dogs with Diarrhea don't wag their tails :wtf:
Posted on Reply
#14
bug
ZoneDymo
wonder how rich the people at TSMC must be, also poor samsung
Maybe you should look into how much it costs to ramp up a new fab, before wondering about that.
But yes, they wouldn't be doing it if they would be in the red.
Posted on Reply
#15
Luminescent
Biggest jump in performance is always from a new node, ampere, big navi....all that crap is nothing without a smaller better node and TSMC knows this, just look at Intel, no matter how good the architecture still they lose because of the old "14nm" node.
TSMC should be getting a lot of credit for the advances they make.
Posted on Reply
#16
Metroid
Price and performace will stay the same, power consumption will drop a little, nvidia margin will rise a little.
Posted on Reply
#17
Colddecked
Xaled
LoL, so Nvidia just used Samsung to make TSMC feel jealous.
How about Nvidia use both Samsung and TSMC to respond to the supposed high- demand?
Use both Samsung and TSMC in the same SKU? That's impossible, they just bought the damn wafers lol... These TSMC 7nm ampere's, if true, will almost certainly carry a super tag or even 4000 series if improvements are big...
Posted on Reply
#18
Chrispy_
Vya Domus
Story has it Nvidia was pissed about TSMC's prices so they "threatened" them by moving to Samsung in an attempt to drive their prices lower but they didn't budge even after that happened, so now they're coming back with their tail between their legs.

Samsung's nodes are inferior, always have been, even for mobile. That doesn't mean they shouldn't be used but not for 600+ mm^2 + behemoths.
Samsung's nodes aren't up to TSMC's standards, but what about GloFo, are they now such a mess that nobody in the CPU or GPU business dares touch them?
Posted on Reply
#19
renz496
"The report claims that TSMC has become more "friendly" to NVIDIA."

not that TSMC is afraid of samsung but they probably try to make sure samsung to not becoming a formidable competitor in the future.
Posted on Reply
#20
PerfectWave
No space for NVIDIA at TSMC...
Nvidia told us that they cannt keep up with order of their Ampere GPU cos of high demand not cos of Samsung XD
Posted on Reply
#21
Chrispy_
Smaeili
Would the smaller manufacturing process bring any improvement to thermals/power?
Hopefully yes, but until the end products start rolling out of the fab, nobody will know - not even Nvidia.

One of the problems with Samsung 8nm is that it was leakier and less efficient than expected - way too late to change things at that point. TSMC's 5nm could be great, it could be awful. We won't know for several months yet.
Posted on Reply
#22
birdie
That NVIDIA is looking to tier its manufacturing process across high-end and the rest of its product stack (with 7 nm for high-end and 8 nm for the rest of it) would become a headache for themselves and for consumers, should NVIDIA just have two suppliers for the same graphics products. There would likely be need for some changes in the power delivery designs, there are a range of new quality assurance tests that have to be taken for the new silicon, and NVIDIA would set itself up for legal troubles should they just silently update the manufacturing process on high-end models - not only would early adopters be understandably miffed about their product having evolved over time, as there could be some claims regarding 8 nm-based models being bought after the 7 nm ones are launched. And if NVIDIA were to put a sticker on retail boxes updating the 8 nm to 7 nm, well, then any user could just decline to purchase any 8 nm cards, and only look for the 7 nm versions, which might leave NVIDIA with a real immovable supply problem.
This entire paragraph is, to put it mildly, some pure bullocks not corroborated by any historical evidence and the evidence says there have been no issues in the past whatsoever, for instance, the GeForce 9 (9xxx) series was produced using both 65 and 55nm nodes and no one cried foul. And it's not the only series which used several nodes simultaneously. As long as products are priced correctly and have proper specs, people will be OK.
Chrispy_
If you read between the lines, the headline could be "Nvidia so dissatisfied with Samsung's 8nm that they're already planning to abandon it halfway through the current generation"

It doesn't inspire confidence in the 3000-series on sale right now though; Just another thing to add to the 3000-series launch fiasco...
NVIDIA grossly underestimated the demand for the RTX 3000 series as way too many GeForce 1000 series owners decided to upgrade. "Fiasco" is purely your gall and nothing else. The new series has been extremely successful so far.
ratirt
At first, NV went with the Samsung 8nm because it is cheaper and matured and wanted to skip the 7nm TSMC because it was much more expensive.
And your official source for this statement is ...? Or you're based off some wild rumors on the net?
Posted on Reply
#23
mrkuro
Maybe that's why shortages in supply of GeForce RTX 3080 and RTX 3090 graphics cards could persist until 2021? This hopefully is not true, or else people will get mad :)
Posted on Reply
#24
gridracedriver
this could waste nVidia more time on 5nm and MCM compared to AMD
Posted on Reply
#25
Chomiq
They might pull it off for Ampere refresh, basically Super series, but I doubt they'll do it with current 30xx lineup.
Posted on Reply
Add your own comment