Intel has established a new division, Intel Foundry, with the purpose of manufacturing chips for other companies, having Microsoft as its largest client to date.
Microsoft is planning to create its own sophisticated semiconductor chip and has chosen Intel Foundry for its production. This long-term agreement is estimated to be worth around 15 billion US dollars.
The chip designed by Microsoft will utilize Intel’s 18A chip manufacturing process, which is expected to be introduced later this year. This was announced by Intel on Wednesday during an Intel Foundry event. Along with this, Intel shared an ambitious product roadmap aiming to become the world’s second-leading chipmaker by the year 2030.
In order to accomplish this aim, Intel requires major clients like Microsoft. The Intel Foundry’s main objective is to design chips suitable for the “AI era,” as mentioned by company representatives.
Some individuals posit that Microsoft’s agreement with Intel is an attempt to lessen its dependency on Nvidia. Despite this, a series of collaborations between Microsoft and Nvidia have been announced over the past few years. These include a 2022 partnership to construct a massive AI supercomputer, an early 2023 collaboration on the industrial metaverse, and a mid-2023 cooperation on large-scale enterprise AI.
The agreement provides Microsoft’s business with access to a US-based foundry. According to independent chip industry analyst Mike Demler, the tech giant might be aiming to reduce its dependency on another significant chip manufacturer, Taiwan Semiconductor Manufacturing Co. (TSMC).
In November, Microsoft revealed two specifically designed chips: the Azure Maia AI Accelerator, tailored for AI workloads, and the Arm-based Azure Cobalt CPU, designed for running general-purpose computing processes in the cloud. The deal with Intel is primarily concerned with the future production of these two newly introduced chips, says Demler.
According to Glenn O’Donnell, Vice President and Research Director at Forrester, Microsoft will probably design additional chips beyond the AI and cloud-specific ones. “This is merely the beginning, as it will need further chips to cater to other purposes in its cloud operations as well as processing AI at the edge and within PCs,” he stated.
O’Donnell consented that the revelation had negligible connections with Nvidia. The pact with Intel isn’t about Nvidia as Nvidia doesn’t fabricate chips, he stated. “This doesn’t affect Nvidia adversely,” he commented. “These are separate partnerships.”
Microsoft will persistently utilize Nvidia GPUs besides its proprietary chips, O’Donnell conveyed. “Though Microsoft’s AI aspirations might slightly undercut Nvidia’s superiority in driving AI, Nvidia will persist as the powerhouse in AI.”
Intel’s cascade of chip disclosures this week signifies a renewed endeavor to construct a foundry ecosystem advocating external clients, Demler articulated. In October 2022, Intel proclaimed an in-house foundry model, for both external clienteles as well as its internal merchandise lines.
Demler endorsed the company’s novel efforts to collaborate with Arm, Cadence Design Systems, and other firms as a “positive indication” for future prosperity.
Intel is now focusing on the production of chips for third-party customers as it seeks to regain its position in a market currently dominated by TSMC and Samsung. Major manufacturing missteps, stiff competition from Arm, Nvidia and other chip creators, as well as a slump in PC sales, are factors that have led to the dwindling market share of Intel, previously the globe’s leading chipmaker since the latter part of the 2010s.
During an event on Wednesday, Intel’s CEO, Pat Gelsinger, revealed their ambition to revive the chipmaking industry in the West. The tech giant aims to have around half of the world’s semiconductor chips produced within the confines of the United States and Western Europe over the next decade, an increase from the current statistic of approximately 20%, he elaborated.
