Intel Now Producing More 10 Nanometer Wafers Than 14 Nanometer Wafers

[Uskompuf]

Intel has recently confirmed during a recent earnings call that it is now manufacturing more 10 nm wafers than 14 nm wafers as part of its IDM 2.0 plan. This news comes four years after Intel first started shipping 10 nm products to customers however the production capacity did not exist to launch any mainstream desktop processors. Intel launched a few low-power mobile Ice Lake processors in 2019 but their 2021 flagship 11th Generation Rocket lake processors remained on 14 nm. This looks set to change as Intel highlights reduced production costs and prepares to launch 10 nm 12th Generation Alder Lake processors in Q4 2021. Intel will offer more details on their future manufacturing plans during their upcoming Intel Accelerated event.

Under IDM 2.0, our factory network continues to deliver and we are now manufacturing more 10-nanometer wafers than 14-nanometer. As 10-nanometer volumes ramp, economics are improving with 10-nanometer wafer cost 45% lower year-over-year with more to come.

View at TechPowerUp Main Site

 

[Fatalfury]

so intel probably want to extract the most out of the 10nm node atleast for like 3-5 years..
when tsmc is in 5nm and samsung in 6-8nm. we can expect intel to get 5nm by 2025 maybe.
however we need alternatives for tsmc.

 

[dj-electric]

so intel probably want to extract the most out of the 10nm node atleast for like 3-5 years..
when tsmc is in 5nm and samsung in 6-8nm. we can expect intel to get 5nm by 2025 maybe.
however we need alternatives for tsmc.

You might or might not get a very clear clarification on that in a matter of days.

 

[bug]

Ok, more wafers, check. But at what yields?

 

[watzupken]

so intel probably want to extract the most out of the 10nm node atleast for like 3-5 years..
when tsmc is in 5nm and samsung in 6-8nm. we can expect intel to get 5nm by 2025 maybe.
however we need alternatives for tsmc.

I feel Intel won't want to spend too much time on 10nm. The problem is 10nm is very late, and with 7nm (which is at least 1 year late) expected in 2023, I feel if they can, they will want to move away ASAP. Otherwise, they will have a dependency on TSMC for cutting edge nodes which is not going to be easy on them when it comes to cost.

My opinion is that Intel's 10nm is better than their 14nm for sure, but I don't think it is a lot better. Take Tiger Lake H for example, it is faster with the bigger and newer cores along with high frequency, but it comes with high power consumption. Performance is a wash, with Tiger Lake winning some, and losing some to Ryzen 5000H. Moving on to Alder Lake, the performance core count remains stagnant, and instead Intel is cramming "efficient cores" to bump up multicore numbers. You could argue that big/small CPU config is more efficient when you have light loads that don't need that much power, but,
1. Do you really need up to 8 efficient cores for light loads? it sounds to me they are just slapping more cores to bump up multicore numbers at load, which brings us to point 2,
2. With the rumored PL2 value at 228W, if you tend to run heavy loads, then it is not actually efficient when AMD's current Ryzen 5000 caps out around 140W for a 16 cores config.

Just my speculations on the state of Intel's 10nm limitations.

 

[persondb]

so intel probably want to extract the most out of the 10nm node atleast for like 3-5 years..
when tsmc is in 5nm and samsung in 6-8nm. we can expect intel to get 5nm by 2025 maybe.
however we need alternatives for tsmc.

They all continue to offer nodes even after going to newer ones. TSMC as an example has a lot of mature nodes that they still offer to clients.
It doesn't mean that Intel will continue to use 10nm for their flagship products for another 3-5 years.
They have a lot of stuff like IoT, FPGAs and etc that will be manufactured for long periods of time too. Plus their with their whole foundry goal, they can't just abandon 10nm after offering it to chip designers.
Also, Intel 5nm is very different from TSMC 5nm, closer to their 3nm. Intel competitor to TSMC 5nm will be their 7nm process.

 

[TheLostSwede]

They all continue to offer nodes even after going to newer ones. TSMC as an example has a lot of mature nodes that they still offer to clients.
It doesn't mean that Intel will continue to use 10nm for their flagship products for another 3-5 years.
They have a lot of stuff like IoT, FPGAs and etc that will be manufactured for long periods of time too. Plus their with their whole foundry goal, they can't just abandon 10nm after offering it to chip designers.
Also, Intel 5nm is very different from TSMC 5nm, closer to their 3nm. Intel competitor to TSMC 5nm will be their 7nm process.

Just a side note, Intel has been busy selling off most of its IoT stuff, or just killing it off, for some reason.

 

[Richards]

Looking good they already shipped 50 million tiger lake's.. amd can never get amount of volume. When tsmc is slow

 

[Vayra86]

Ok, more wafers, check. But at what yields?

Someone's paying attention

Cost 45% lower YoY also speaks of a pretty horrible figure at this time and prior..

 

[Ravenas]

Where are these wafers being produced???

 

[bug]

Where are these wafers being produced???

Oregon, most likely: https://en.wikipedia.org/wiki/List_of_Intel_manufacturing_sites

 

[bonehead123]

"As 10-nanometer volumes ramp, economics are improving with 10-nanometer wafer cost 45% lower year-over-year with more to come"

Sooooo... in other words, they didn't want to use these chips in mainstream products until they knew for sure they could make a killer markup/profit margin on them...

geez, whodathunkit

 

[Frank_100]

Just a side note, Intel has been busy selling off most of its IoT stuff, or just killing it off, for some reason.

because no one needs to be able to control their dishwasher from their phone.

 

[bug]

"As 10-nanometer volumes ramp, economics are improving with 10-nanometer wafer cost 45% lower year-over-year with more to come"

Sooooo... in other words, they didn't want to use these chips in mainstream products until they knew for sure they could make a killer markup/profit margin on them...

geez, whodathunkit

High cost was due to low yields. Who wants that for their mainstream products?