Intel’s Bold New Chip Manufacturing Plan

Intel is undertaking a multi-prong, multi-billion dollar initiative under new CEO Pat Gelsinger to reestablish the chip maker’s manufacturing preeminence after several years of challenges that damaged its reputation and saw competitors pass it by.

Gelsinger this week unveiled a plan, dubbed IDM 2.0 – an acceleration of its integrated device manufacturing model – that calls for spending $20 billion to build two new chip manufacturing plants in Arizona, making chip factories in the United States and Europe available to make other vendors’ processors in an expansion of its foundry business, and leveraging other third-party foundries to build some Intel products, including chipsets as well as graphics, communications and connectivity chips.

Intel also plans to partner with IBM to create next-generation logic and packaging technologies that will help drive processor capabilities, accelerate the manufacturing of semiconductors, and reverse the United States’ shrinking presence in the worldwide semiconductor industry, a key concern among government officials and vendors alike at a time when chip shortages have roiled a range of industries such as automotive and electronics, particularly since the onset of the COVID-19 pandemic.

Manufacturing stumbles and product delays over the past several years have enabled rivals like AMD to surpass Intel in such areas as 7-nanometer chips and Asian foundries like Taiwan Semiconductor Manufacturing Corp. (TSMC) that can make more advanced processors.

Sticking with chip making

Gelsinger dispelled any thoughts that some industry observers had about Intel possibly outsourcing more of its manufacturing to third-party chip factories, or “fabs.” The CEO said Intel would continue to manufacture the bulk of its processors and suggested that rival chip makers like Qualcomm and cloud providers like Amazon, Google and Microsoft could become customers of its foundry division, Intel Foundry Services (related article: Google Turns to Intel Exec to Supercharge Server Chip Development).

“Intel is and will remain a leading developer of process technology, a major manufacturer of semiconductors, and the leading provider of silicon globally,” said Gelsinger, who spent three decades at Intel before becoming VMware’s CEO in 2012. He returned to Intel in February.

Patrick Moorhead, principal analyst with Moor Insights and Strategy, noted that while the past few years have been hard on Intel’s reputation, the company has done well financially and still holds commanding shares of the server market (90 percent) and PC space (80 percent) and is expanding its reach in other markets like automotive and the Internet of Things (IoT).

While I am certain many of the new gears disclosed today were in motion under prior CEO Bob Swan, Gelsinger’s mark is clearly on the plans, and he exudes senior technical and execution confidence and openness that I haven’t seen at Intel for years, Moorhead wrote in a Forbes column. Gelsinger said all the right words and gave me all the reasons to believe that Intel is ‘back,’ but it now needs to execute.”

Intel’s challenges

Intel’s delays in making the shift from 10nm to 7nm chips has highlighted its manufacturing struggles in recent years, which have come at a time when the chip industry has become more competitive. A resurgent AMD is making inroads into the server and PC markets with its Epyc and Ryzen chips based on its Zen microarchitecture. AMD this month unveiled its 7nm 3rd Generation Epyc “Milan” processors and ARM has been making advances in its server chip designs.

GPU maker Nvidia also is bidding to buy ARM for $40 billion in a move that would make it a stronger rival to Intel in the data center. Meanwhile, Amazon Web Services (AWS) has developed its ARM-based Graviton chips to use in its hardware, and Google this week announced it has hired longtime Intel engineer Uri Frank to oversee its upcoming efforts to design systems-on-a-chip (SoCs) that it can use in its data centers. Apple also is replacing Intel processors in its laptops with its own ARM-based SoCs.

7nm development ‘progressing well’

Gelsinger countered that Intel’s current 7nm development is progressing well, due in large part to the use of extreme ultraviolet lithography (EUV) in its re-architected process flow. He expects Intel to tape in the compute file for its 7nm client chip, codenamed Meteor Lake, in the second quarter, with high-volume manufacturing in 2023.

The increased use of third-party foundries like TSMC, Samsung and Globalfoundries will help Intel get more product into the market. That will include manufacturing products that Gelsinger said are at the core of Intel’s portfolio for both data center and client segments beginning in 2023. It will help the chip maker reach the flexibility and scale needed for Intel’s ambitious roadmap in terms of costs, performance and supply.

That combined with its new standalone Intel Foundry Services business – which will be run by Randhir Thakur, who has been with the company since 2017 as chief supply chain offer – and its new fabs in Arizona will help address the skyrocketing global demand for semiconductors that has led to the current shortage. It also will offer chip makers foundry services not only in the United States but Europe as well, which is something that TSMC can’t offer.

Foundry business

Gelsinger said the foundry business will manufacture not only x86 chips – like those from Intel and AMD – but also those based on RM and RISC-V architectures. It will compete in a market that will hit $100 billion by 2025, he said.

Moorhead said the $20 billion investment in the new Arizona fabs likely won’t be the last money Intel will put into manufacturing facilities.

I don’t believe this is the extent of Intel’s investment, particularly given the need of U.S. and European governments who are asking for more on-shore, leading-edge manufacturing, the analyst wrote. Obviously, if Gelsinger and the board didn’t have confidence in its future capability and reinforced by its end customers, it would not be making these investments. This isn’t Intel hedging its manufacturing bets – it looks to me as the company is all-in. I was pleased to hear that the $20B investment is not determinant of any government funding, but I believe Intel should get funded by both the U.S. and the EU to build more capacity more quickly.

Intel currently has manufacturing sites in Arizona, New Mexico, Oregon and Massachusetts, as well as Ireland, Israel and China.

Intel was one of a number of semiconductor companies and organizations signing onto a letter in February to President Biden urging for money to help expand the United States’ dwindling presence in the global chip market. The letter noted the pervasive presence of semiconductors in modern society and the benefit to the U.S. economy and competitiveness by increasing manufacturing capabilities in the country.

Putting the U.S. back on the semiconductor map

Biden later in the month stressed the importance of the semiconductor industry to the United States and signed an executive order authorizing a 100-day review of supply chains in the space. In response, Jeff Rittener, chief government affairs officer for Intel, wrote in a blog post that the combination of the executive order and full funding of the Creating Helpful Incentive to Produce Semiconductors (CHIPS) Act in Congress “help level the playing field in the global competition for semiconductor manufacturing leadership, enabling American companies to compete on equal footing with foreign companies heavily subsidized by their governments.

The construction of the new fabs on Intel’s Ocotillo campus in Arizona also could be a boon for the local economy, creating what Intel said will be 3,000 permanent high-tech jobs, more than 3,000 construction jobs and about 15,000 local long-term jobs.

In addition, Gelsinger said Intel is bringing a new company event to San Francisco in October called Intel On, which will echo the vendor’s popular Intel Developer Forum, which was last held in 2016. 

News Around the Web