The companies have announced that Intel, a US chip manufacturer, will produce high-end semiconductors exclusively for Microsoft. Intel is competing with TSMC, Samsung and other suppliers to provide the next generation silicon for artificial intelligence.
Pat Gelsinger, Intel’s chief executive, said on Wednesday at an event that the company is “rebuilding western manufacturing at large scale”, boosted by geopolitical worries in Washington regarding the need to bring cutting-edge manufacturing to the US.
Since Gelsinger became CEO three years ago Intel has tried to reinvent itself by becoming a foundry, producing chips designed and manufactured by other companies. This is a way for Intel to regain its edge in manufacturing the most advanced semiconductors. The demand for these chips is increasing, fueled by the growth of generative artificial intelligence.
Intelneeds, however, to convince manufacturers that the chips it produces will be treated the same way as those it designs.
Intel’s goal is to ensure that 50% of all semiconductors in the world are made in the US and Europe within a decade. This compares with 20% today. The majority of global production takes place in Asia, primarily in Taiwan.
Gelsinger stated that Intel has become “maniacally customer-focused”. Intel is restructuring into two separate units to reflect this. Intel Foundry manufactures semiconductors and Intel Products are involved in the design.
Gelsinger stated that Intel is “establishing two vibrant, new organisations”. He added that the company had already signed foundry agreements with a total lifetime value of 15bn.
Microsoft’s chief executive Satya Nadella said that the company supported Intel in its bid to become the world’s leading chip manufacturer. Intel will produce Microsft’s chip design using Intel’s node 18A.
Intel’s 18A is a manufacturing method that reduces the size of semiconductors and makes them more energy-efficient. Taiwan’s TSMC, and South Korea Samsung are deploying to produce the next-generation of cutting edge chips that will power smartphones and data centres.
Microsoft, along with Nvidia and AMD are among the companies that design the high-end processors that provide the immense computing power needed to roll out AI.
Gelsinger stated, “I want to manufacture all AI chips in the industry through our foundry.”
These include semiconductors built on the architecture designed by UK chip designer Arm. Rene Haas, Arm’s chief executive, appeared at the event to announce the new partnership between Intel and Arm. He joked that the two companies were “strange bedfellows” given their past. Intel’s designs rely on the X86 Architecture, which is in direct competition with Arm’s.
The federal Chips Act of 2022 promises $52.7bn for US research and development as well as manufacturing.
The Biden administration announced earlier this week that California-based GlobalFoundries will receive $1.5bn to build a new facility, expand and modernise manufacturing at sites in New York, Vermont and for the production of automotive, communications, and defence semiconductors.
Gelsinger stated that Intel would announce its own federal grant, under the Chips Act “very soon”.
OpenAI, Google, and others developing and implementing large language models for generative AI will require “mind-boggling volumes” of semiconductors in the coming years. The demand for the latest chips is “just going to explode”, said US Commerce Secretary Gina Raimondo, who appeared on screen at the event.
She said that Intel is the chip leader in the United States.
Intel shares closed 2.4% lower on Wednesday.