IndustryWeek's curated collection of semiconductor news
In his State of the Union address on Tuesday, President Joe Biden called Intel’s $20 billion initial investment in semiconductor manufacturing in central Ohio “the ground on which America’s future will be built.”
Construction begins this year on that first phase, and if subsequent phases go forward, eight semiconductor manufacturing facilities total could rise on that 1,000 acre site, with an investment totalling $100 billion.
Mark Granahan, CEO and founder of iDeal Semiconductor in Bethlehem, Pennsylvania, has been involved in semiconductor manufacturing and R&D his entire career, with Bell Labs and Texas Instruments and now at iDEAL Semiconductor, a startup focused on developing energy-efficient semiconductors in the power space.
Granahan talked with IndustryWeek about what the Intel plan and other plans like Taiwan-based TSMC’s expansion in Arizona mean for semiconductor manufacturing in the U.S. long-term.
IndustryWeek: Will these announced expansions solve the problem of the semiconductor shortage, or will there still be holes in the supply chain?
Granahan: [Intel’s expansion plan] takes a pretty major step forward in improving the U.S. supply chain, in terms high-end, high-performance and similar types of circuits. So maybe the next question is, where are they going to be packaging and testing these products? Because if they have to send that overseas, it really doesn't help the supply chain as much as it should.
As technology becomes more advanced, it's less about actual labor cost, and more about the technical expertise to actually manufacture and build the products. These factories, once they’re up and running, will be producing chips for 20 plus years, at least. And so, as we move into the latter part of this decade, next, having that assembly and test capability in the U.S. will be a really strong calling card for U.S. supply chain base.
What sorts of roles and training are needed for these new semiconductor facilities?
Granahan: It will be operators, basically hourly individuals that will definitely need training, but a lot of that training can be on the job. Then there will be technicians, the next level up in terms of education. Most will likely need a two-year degree or some type of associate's degree, in equipment, maintenance, electronic electronics, chemistry, those types of things. And then from an engineering perspective, there will be a range of bachelor's, master's as well as Ph.D.s that will be in in the facility, working on the device structures working on the process technology, working on advanced process flows and equipment.
The good news is this will be an across-the-board opportunity for people who don't necessarily have a college degree all the way up to very advanced Ph.D. types of types of individuals in the areas of chemistry material science, metallurgy, ceramics, polymers, physics, electrical engineering—a pretty wide range of engineering functions will be necessary. And then, of course, maintaining such a facility is going to require lots of mechanical and civil engineers.
It's also going to bring companies that supply equipment to that type of facility, so that's companies like Applied Materials, which is U.S. based, ASM, which is U.S. based; ASML, which is Netherlands-based' Lam Research. U.S.-based; Kla Tencor, U.S. based. Those equipment suppliers will have to set up shop in that area, too, in order to support all the tool sets that will going into those facilities. And then you've got all the chemistry and gases for the facilities, which are companies like Air Products, they do industrial gases. Chemical companies that provide the raw materials for the manufacturing process.
Do you think the CHIPS Act ($52 billion in government funding to grow the US semiconductor industry) must pass before these expansions are fully realized?
The CEO from Intel said that [passing the CHIPS act] wouldn't impact his plans, but it might accelerate his plans. I'm hoping the funding can go toward areas that are cash-starved. To draw a contrast, if you look at Intel's balance sheet, or Nvidia or Qualcomm, or Texas Instruments, or Analog Devices, or any of these large U.S.-based semiconductor companies, they have so much cash on their balance sheets. They have enough cash available, and access to equity markets and access to debt markets, they should be able to do whatever they need to do in terms of capital expansion.
And that's why you see companies like Intel moving forward on their own nickel, basically, because they’ve just got a huge stash of cash. I'm hoping that a lot of that $52 billion goes towards more advanced research so that we can stay ahead. The U.S. is actually behind; TSMC has surpassed Intel in terms of their technology performance. I’d like to see funding go toward advanced research in very high performance areas like silicon and advanced materials. And things like analog circuits, power circuits, optical circuits or sensors—a lot of other ancillary and associated semiconductor components.
The other thing I hope for is some type of public-private matching to support investment in startup activities because they have really dropped off in the U.S. For every startup here in semiconductors, there's probably 100 of them in China. And so ultimately, that Chinese venture investment is going to pay off for China. The fact that we're not investing foreshadows a problem in the future. If we could take some of that $52 million, and plus that up with private money—for instance, create a $2 billion venture fund that is run by the Commerce Department or another department within the federal government, and have private industry match that with $2 billion, and have it focus on real, venture-based startups here in the U.S.—that’s the kind of thing that will get a lot of private dollars flowing into the space.
You support expanding the U.S. semiconductor supply chain throughout Latin America. Why, and how would that work?
Most assembly and testing for the industry is in Asia. And trading and expanding that type of capability in the Americas would be a really good thing. It would a great use of some of that $52 billion, but it's going to take more than that. I believe it's going to take some initiative by the, by the federal government to, to start talking about, and putting money towards expansion of that capability here in the Americas. Clearly, this supply of semiconductors here in the U.S. as it expands, it creates more and more of the conditions for having the the total supply chain in the in the Americas, as opposed to sending stuff in overseas all the time. It’s not taking away from Asia; it’s actually adding more capability and capacity here in the Americas. I really would hope the Intel CEO would start evangelizing such such types of expansion.
