The race for 2nm chips will start in 2025. And it will be the fiercest of all

2nm integrated circuits are going to hit the market in style during 2025. Users know that nanometers have lost much of their usefulness, and that they actually represent a category of semiconductors. In fact, no longer faithfully reflect the length of the logic gates or another physical parameter, such as the distance between the transistors. Each chip manufacturer handles them very freely, which prevents users from directly comparing the lithographs they try to “sell” us.

Be that as it may, during the year that we are about to start, TSMC, Intel and Samsung are going to tackle large-scale manufacturing of their 2nm chips. Of course, each of these companies is in a very different situation. The Taiwanese TSMC leads the integrated circuit manufacturing market with a market share of approximately 60% and is going to close a very successful 2024. Furthermore, according to DigiTimes AsiaApple, NVIDIA, AMD, MediaTek and Qualcomm have already expressed to this company their intention to formalize orders for its 2nm integration technology.

Samsung’s starting point is less comfortable than TSMC’s. This South Korean company is reevaluating its expansion plans and adjusting its workforce in order to face the future with the greatest possible guarantees. According to Gartner, its revenue fell by 37.5% in 2023 compared to what it reached in 2022, and 2025 is presented as a year in which there will be a probable global slowdown in the semiconductor industry. In any case, Samsung is ready to start large-scale production of 2nm chips.

2nm represents a great opportunity for Intel and Samsung

Kye Hyun Kyung, the former CEO of Samsung’s semiconductor division, predicted last year that his company will surpass TSMC and its other competitors (a clear reference to Intel) over the next five years. His a priori statements might seem like bravado to us. In fact, Pat Gelsinger, the former CEO of Intel, predicted essentially the same thing, but, logically, in favor of his own company.

The countdown on nuclear fusion begins: we already know where and when the first commercial plant will be ready

According to Intel, its 18A node has reached the necessary maturity to enter large-scale production in 2025

Ben Sell, vice president of technology development at Intel, stated at the beginning of last September that its 18A node has reached the necessary maturity to enter large-scale production in 2025 and will benefit from the resources that will be reassigned from node 20A. He leitmotiv of this change of strategy Dave Zinsser explained itthe company’s CFO: Intel will skip the commercialization of the 20A node in order to save $500 million. In the difficult situation that this company is currently going through, this cost cut seems a reasonable option.

In any case, it is important that we do not overlook that semiconductor manufacturers in general, and TSMC, Intel and Samsung in particular, handle with relative discretion one of the main challenges they face: the per-wafer performance of their lithography. more advanced at the beginning can clearly be improved. And they need to achieve at least 70% performance on their 2nm nodes to ensure the profitability of this integration technology and attract more customers. This is the real goal. We’ll see who gets there first.

Image | TSMC

In techopiniones | Intel’s big bet on laptops is here. This is what their Intel Core Ultra 200V ‘Lunar Lake’ chips look like