While in the PC world, if we are talking about TSMC manufacturing nodes, 7nm is the norm, in smartphones we have had 5nm chips for a long time. It’s true that on PC, we’ll have to wait until 2022 to see 5nm hardware designs in the form of real PC hardware in our hands. But the design process for any processor takes years, and is not aimed at the next node to hit the market, but two manufacturing nodes to come. This is why the greats of the industry have hid themselves from the sight of curious projects conceived around future manufacturing nodes which are in the middle of their conception and which seek to surprise us in a few years.
Intel is one of the first customers for TSMC’s 3nm
According to the Japanese economic newspaper Nikkei, in its version for the entire Asian market. TSMC’s 3nm node will not be released exclusively by Apple as happened with its 5nm node. Rather, Intel will be added to the equation. But it will share the first with Intel. We do not know at the moment what kind of components will send TSMC to manufacture under the 3nm node, the only thing we do know is that according to Nikkei Asia, they will be launched in the second half of 2022. Both in the case of the component made for Intel as it will go to Apple.
Taking out a new node with a smartphone chip is normal, since these are smaller than those usually used in PCs. Which rem inds me of the question of what kind of processor or component Intel ordered. In addition, it must be a component that has a significant sales volume, since According to the original sources, Intel ordered more wafers than Apple.
The news does not surprise us, as it has appeared several times in recent months. For the time being we don’t know what intel intends to make under TSMC’s 3nm node. This offers a density twice that of transistors, combined with the possibility of achieving clock speeds between 10% and 15% higher and with reduced consumption between 25% and 30%, all this compared to the node 5 nm.
Are AMD and NVIDIA excluded?
In the case of NVIDIA, we can figure this out because nodes when released don’t get along with big chips like NVIDIA GPUs have been in recent years. Moreover, if the rumors are true, the next architecture change will take place well in 2020 with its Lovelace architecture, which could use TSMC’s 5nm node. Since Nvidia is launching new GPU architectures with a rate of 2 years, for its 3 nm design it will be necessary to wait until 2024 at the earliest.
In AMD’s case it’s stranger, its CCD chips are generally small and similar in this aspect to smartphone SoCs, so AMD shouldn’t have any problems accessing the 3nm node for a possible Zen 5. Okay Of course, is that the rollout date of the 3nm TSMC node for Intel and Apple is 2022, the launch year of Zen 4 which will use the 5nm node.
They are therefore not really left out, it is a phased deployment where each designer will have access to TSMC production at the appropriate time. Of course, you’ll start with the smallest and easiest chips to make and then move on to the more complex ones. How it goes in each new manufacturing node.