Listen to the audio version of the article
“Intel Foundry” probably won’t be an original name for a project that aims to be a milestone for a new phase of the silicon industry. In fact it is not at all, given that it is a “renaming” of the previous Intel Foundry Services production activity. But the announcement that the Californian giant announced in San Jose in the presence of various representatives of the elite of the US technology industry (including Satya Nadella, Sam Altman and Rene Haas, respectively the CEOs of Microsoft, Open AI and Arm) and figures of prominent institutions (among the speakers there was in fact the United States Secretary of Commerce, Gina Raimondo) is of great importance. And for some precise reasons. The first is easy to understand: it is the first foundry for the production of semiconductors designed for the era of artificial intelligence. And in the words of Intel’s CEO, Pat Gelsinger, the strategic thinking behind the project can be clearly read: «AI is profoundly transforming the world and the way in which we conceive both technology and the silicon that powers it», the manager commented in a note, adding how this transformation creates “unprecedented” opportunities for the designers of the most innovative chips and for factories, such as Intel Foundry, which are created to dedicate themselves to this new frontier. The credentials that the new creature can bring into play are weighty and include all generations of foundry processes, a significant number of customers using the ASAT (Advanced System Assembly and Test) functionalities and an estimated turnover, between wafers and advanced packaging, which will exceed 15 billion dollars.
The project with Microsoft
Satya Nadella’s keynote speech served to formalize the design of a new chip to be produced with Intel 18A (1.8 nanometer) process technology starting in the second half of 2024. Microsoft did not reveal any additional details in merit but the launch dated November 2023 comes to mind, when the Redmond company announced two new custom processors designed specifically for its artificial intelligence workloads. And yesterday’s appears as the ideal new piece of a strategy aimed at increasing computational capacity to be dedicated to applications driven by algorithms. The imprint on the new partnership with Intel Foundry was given by the CEO of Microsoft himself, speaking of a “very exciting platform change that will radically transform the productivity of every single organization and of the entire sector” and of the need to “a reliable supply of the most advanced, high-performance, high-quality semiconductors” to realize this vision.
The new roadmap with 1.4 nanometer process technology
For Intel, the baptism of the “new” foundry was also an opportunity to make public the details of the new product roadmap, which is enriched by Intel 14A process technology (at 1.4 nanometers) and various system evolutions Intel 3 (3 nanometers), Intel 18A and Intel 3-T, the latter optimized with silicon electrical feed-through connections for advanced 3D packaging designs. The ambitious action plan called “5N4Y” (five nodes in four years) is therefore confirmed on track and it will be up to the 1.8 and 1.4 nanometer chips (which will presumably debut in 2026) to mark the new step forward in terms of innovation of the production technologies of the Santa Clara company, which will make full use of the High-Na EUV lithographic system and which will rely not only on the Foundry designed for AI but also on the future Fabs that will arise in Europe.