No 3NM Intel CPU’s

Intel will not produce 3nm CPUs and will follow AMD’s strategy

Intel has controlled everything connected to CPUs, both desktop and server, for more than a decade, thanks in part to its architecture and in part to its superior manufacturing processes. But then came their 10nm, and they were overly ambitious to give it a high density, so they spent a lot of time in limbo trying to correct the flaws. As a result of Intel’s collaboration with TSMC to use their 3nm process, they are now suffering the price. Isn’t everything going as planned? 

The most bizarre thing about this whole problem is that Intel is much later than we thought. Especially considering that less than a month ago, the company had presented its new nomenclature system with dates for the different nodes. 

Is Intel, in collaboration with TSMC and their 3nm process, creating a smokescreen?  

It is no longer a secret that Intel has been forced to rely on TSMC for its next-generation Xe GPUs, notably its 6nm node. Furthermore, several speculations suggest that this will occur on some low-end CPUs without identifying the lithography process in between.

However, current speculations say that Intel would have reserved a major portion of TSMC’s output with its 3 nm process, leaving Apple the only business competing with those from Santa Clara for efficiency and performance superiority. The volume projected by TSMC is so great that, as part of an Intel TSMC cooperation for its 3 nm, the Taiwanese will develop a new dedicated manufacturing plant to meet the demand.

This plant will be Nanke 18b, which is next to Nanke 18a, which currently manufactures 5 nm chips for Apple. 

Delays costing millions of dollars, Intel caves with TSMC and copies AMD

What Intel is doing is possibly the most rational and astute move possible: it is avoiding the semiconductor market leader, asking for chip production capacity at its leading-edge node, and thereby mimicking AMD’s strategy. Except for one minor detail: all of this will take place in late 2022 or early 2023, but by then, Intel will have deployed its Intel 4 (former 7 nm) lithographic process, so we may expect to see processors from the blue giant with different nodes on the market.

What does this mean for AMD? A good question with a complex answer. According to rumors, there is no more 3 nm capacity now, and it won’t be until 2024 (ASML EUV scans are absent, they would be behind schedule, and the volume is massive). Therefore, AMD would hardly be able to produce chips with TSMC. As a result, Lisa Su’s team would have moved to Samsung to inquire about their 3nm with GAA if TSMC could no longer provide more chips for its CPUs and GPUs.

Soap opera insight? Possibly, NVIDIA has already had to switch from TSMC to Samsung, and the results in Ampere have not been horrible, but the Koreans are a little behind, and that has cost them dearly, very dearly today. Intel, for its part, has killed two birds with one stone: it secures the best node for its CPUs (and maybe GPUs) while drowning AMD, forcing it to settle for a technically and theoretically inferior node. Masterstroke?

Previous Post

Intel shows its new Xeon CPUs with 64 GB HBM2 and 5700 mm2 memory!

Next Post

Crysis Remastered Trilogy Release Date

Related Posts