Intel (INTC.O), opens new tab said on Wednesday its deal for $7.86 billion in U.S. government subsidies restricts the company’s ability to sell stakes in its chipmaking unit if it becomes an independent entity.
The U.S. Commerce Department announced the subsidy to Intel on Tuesday, part of $39 billion for the sector including Taiwan Semiconductor Manufacturing Co (2330.TW), opens new tab and others in an effort to revitalize chip manufacturing in the United States.
In a securities filing, Intel said on Wednesday the subsidies require it to own at least 50.1% of Intel Foundry if the unit is separated into a new privately held legal entity. If Intel Foundry becomes a public company and Intel itself is not the largest shareholder, the company could sell only 35% of Intel Foundry to any single shareholder before running into change-in-control provisions.
Intel did not immediately respond to a request for comment on the disclosures. A Commerce Department spokesman said the government is negotiating change-in-control provisions with all direct grant recipients.
Intel would need to comply with the restrictions to continue the company’s $90 billion worth of projects in Arizona, New Mexico, Ohio, and Oregon and keep manufacturing cutting-edge chips in the U.S., according to the filing. Any changes in control could require Intel to seek permission from the U.S. Department of Commerce, the filing said.
In the past Gelsinger has been very clear that IFS will remain a wholly owned subsidiary of INTC, with no intention of selling it. All the rumors to the contrary…
Intel is no longer “cutting edge” in chip technology. TSMC is years ahead of Intel in terms of wafer tech and die shrinkage in order to increase yields, increase chip speeds, and reduce power consumption.
That is not clear. TSMC has said they view TSMC N2 as about the same as Intel_18A. Intel_18A is in production (but not commercially available yet) and has several leading edge technology features known as Ribbon FET and backside power delivery. TSMC will introduce ribbonfet on their first version of N2, and have a follow up version of N2 with Backside power delivery.
"Intel recently revealed that it had given up any plan to make at least the CPU die for some Arrow Lake models using its own new 20A process.
Instead, 20A is now essentially dead, and all of the active tiles in Arrow Lake will be made by Taiwanese chip foundry TSMC. Of course, what with the troubles Intel has had with its own silicon [with the 13th and 14th Gen Core processors], you might view a pure-TSMC Intel CPU as a good thing.
You’re not wrong @jerryab2 but Intel is (hoping to) leapfrog 20A to go to 18A because 20A yields weren’t adequate to take to market, for their own product or for potential customers like Broadcom, and it’s more a priority to get 18A to market in some form than to raise 20A yields right now since their customers aren’t adopting it over the yield issues.
So five nodes in four years… if you skip the fourth one and if the first three are basically moderate respins of their older processes to buy time.