TSMC to Begin Mass Production of 5nm Chips in 2020

[AleksandarK]

According to industry sources over at DigiTimes, TSMC will begin mass production of its 5 nm node in March 2020, when companies using the 5 nm PDK can tape out their designs and integrate them into future products. Going into volume production two years after the 7 nm node, 5 nm is trying to put Moore's Law back on track again.

Built using the Extreme Ultra-Violet lithography (also known as EUV), 5 nm node is supposed to utilize existing FinFET transistors along with many improvements in speed, power and density when compared to existing 7 nm node. Speed is supposed to increase by around 15%, while density will improve by as much as 80%, which is excellent news for everyone. Noticeable power reduction is also present and it is now possible to have about 30% reduction in power consumption, while also enjoying additional speed and density improvements that new node brings.


View at TechPowerUp Main Site

 

[RH92]

TSMC is on a mission , they just need to increase their production capacity .

 

[Deleted member 158293]

Looks like great progress, unfortunately it also means even more price increases if no other fabs can compete with them.

 

[knobling]

The increased density is almost useless anymore for high power chips as the silicon can't transfer the even more concentrated heat away any faster.

 

[notb]

The increased density is almost useless anymore for high power chips as the silicon can't transfer the even more concentrated heat away any faster.

In case you haven't noticed: large, highly clocked desktop CPU are not the only thing one can make from silicon.
5nm will be welcome in mobile CPUs (both ARM and x86) and in IoT.

 

[Turmania]

Will AMD entertain us at 05/05/2020?

 

[Space Lynx]

Will AMD entertain us at 05/05/2020?

no, ryzen 4000 is already all taped out and its on 7nm EUV node... this 5nm EUV is probably for cell phone chips at first

 

[zlobby]

no, ryzen 4000 is already all taped out

Like, totally.

 

[Turmania]

no, ryzen 4000 is already all taped out and its on 7nm EUV node... this 5nm EUV is probably for cell phone chips at first

it does not have to be a CPU it can be a GPU...

 

[windwhirl]

it does not have to be a CPU it can be a GPU...

Nope. RDNA2 is locked for 7nm+...

AMD Updates Roadmaps to Lock RDNA2 and Zen 3 onto 7nm+, with 2020 Launch Window

AMD updated its technology roadmaps to reflect a 2020 launch window for its upcoming CPU and graphics architectures, "Zen 3" and RDNA2. The two will be based on 7 nm+ , which is AMD-speak for the 7 nanometer EUV silicon fabrication process at TSMC, that promises a significant 20 percent increase...

www.techpowerup.com

 

[Prima.Vera]

LOL Intel... 14nm++++++++++++++++++++++ ))))

 

[eidairaman1]

Id like to see a fab refined

 

[R-T-B]

Id like to see a fab refined

You... want to clarify what that is supposed to mean or you just going to leave us hanging?

Fabs are certainly being retrofitted as we speak if their market target date is next year.

 

[quadibloc]

And here I thought it was their 7nm+, the version of 7nm that used EUV, that was coming in 2020, and we would have to wait for 2021 for 5nm. So this is an ambitious schedule.

 

[Ravenlord]

"...7 nm node, 5 nm is trying to put Moore's Law back on track again. "

Please... stop this bullshit with Moore's Law. Theory which was very lucky for him and which was met by people moving on his track.

 

[BorgOvermind]

TSMC is on a mission , they just need to increase their production capacity .

If they would invest much more into mass production there would be less $ for innovation, which in the long term is bad.

 

[notb]

If they would invest much more into mass production there would be less $ for innovation, which in the long term is bad.

Incorrect.

R&D costs (unallocated) don't depend on how much of the product you sell, while manufacturing costs do (allocated).

The more you spend on R&D, the more you have to sell afterwards.
Or backwards: expected gross profit limits how much you can spend on R&D.

What makes TSMC's situation special is that there isn't much competition.
They can make their target revenue with high price and low volume.

 

[Super XP]

This is great. I wonder if AMD is aiming ZEN3 for 5nm? Or going to a more mature 7nm+ is the better option?

 

[windwhirl]

This is great. I wonder if AMD is aiming ZEN3 for 5nm? Or going to a more mature 7nm+ is the better option?

Nope. Zen 3 is going 7nm+. Perhaps with Zen 4

AMD Updates Roadmaps to Lock RDNA2 and Zen 3 onto 7nm+, with 2020 Launch Window

AMD updated its technology roadmaps to reflect a 2020 launch window for its upcoming CPU and graphics architectures, "Zen 3" and RDNA2. The two will be based on 7 nm+ , which is AMD-speak for the 7 nanometer EUV silicon fabrication process at TSMC, that promises a significant 20 percent increase...

www.techpowerup.com

 

[RH92]

If they would invest much more into mass production there would be less $ for innovation, which in the long term is bad.

I don't think this is how things work ! More production capacity means more income therefore more potential budget for innovation , TSMC is barely keeping up with the demand right now and demand will only grow stronger with upcoming next gen consoles etc , so it will be foolish to not increase the production capacity and let all that money go ( they already lost Nvidia ! ) . This should have 0 impact on innovation budget , quite the opposite this should be beneficial in the long term .

 

[BorgOvermind]

I'm not saying they should not increase production. I'm saying they must not start to ignore innovations and do budget cuts on that area.