Home » Intel will have an advantage over TSMC, AMD and NVIDIA in chips after 10 years

Intel will have an advantage over TSMC, AMD and NVIDIA in chips after 10 years

We’ve heard a lot about Intel’s roadmap and its Angstrom lithography processes, which will debut with the 20A as such and that will mean two very clear improvements over its rivals. On the sidewalk opposite, we have TSMCwho gave more details about their most revolutionary lithographic process: their 2nm or N2which will be in inferior conditions compared to its great rival, being relegated to second or third place if Samsung manages to catch up in density.

We open a new chapter in the saga and the race for nanometers, where today we know that TSMC has problems with its 2nm, still under study. As seen for years, the Taiwanese operate a bit like Intel: they launch a lithographic node and it is updated and improved, or it goes to other types of chips. This time it wasn’t going to be any different, but upgrading would instead involve something key.

TSMC and its Nanosheet transistors for 2nm or N2, won’t they be up to Intel?


Besides the density, energy and speed improvements as such, TSMC needs to take a giant leap that it didn’t want to take with the current 3nm that are on the exit ramp or tape. The 2nm will have two key features which, unfortunately for TSMC, will have to be resolved in two phases of the same process, the so-called N2 and its update with N2P improvements.

The N2 will bring with it and not without problems the Nanosheet or GAA transistors commonly called, which have been and still are a fringing problem to be solved. These transistors are horizontal layers that improve energy leakage by being in contact with the materials on all four sides of it. Instead, the call Rear power rail Taiwanese will not be ready for this knot and that is the whole problem, because it will be necessary to wait for the N2P to see them in action.

The transition from N2 to N2P and its consequences

TSMC 2nm N2

Backside Power Rail is TSMC’s answer to PowerVia from Intel, with the problem that in the 20A node Intel makes the jump to GAA with its RibbonFET and also adds this technology to the knot all at once, all in 2024, very very close and as we have always said, the blues have put their foot on the accelerator and do not intend to lower it.

Well, N2P will be a leap forward for Backside Power Rail as it will increase performance and reduce power consumption of each transistor, but although it was commented by TSMC itself, there are only small company notes on the reasons for the delay. It seems, BPR adds additional steps and processes at the node which must be avoided in the first phase of the 2 nm.


Therefore, Intel will have an advantage over AMD and NVIDIA (in case both continue with TSMC) for CPU and GPU after more than 10 years in the shade and multiple delays in its nodes. The blue giant has set sail for a cruise and it looks like TSMC has only been fending off the agony of being overtaken by its rival, sheltering itself until physical limit of FinFET transistors and allow more time to save the SHGs, which they partially achieved.

About the author


Add Comment

Click here to post a comment

Your email address will not be published. Required fields are marked *