Another view of the Defense Department microelectronics shortfall

Tom Temin And this report basically chides DoD or tries to goad it into making more investments in the strategic semiconductor. Isn’t that what the Chips and Science Act is supposed to do in the first place?

Liesl Folks Tom, you’re absolutely right. The Chips and Science Act was intended or is intended to make some big investments in the semiconductor sector, and to onshore and reshore more semiconductor manufacturing. But the Chips and Science Act is primarily focused on commercial and economic security for the nation, which of course, feeds into national security as well. This report pay special attention to the Department of Defense’s needs to our defense ecosystems needs where it comes to semiconductors. And they are a little bit unique and different.

Tom Temin Right. The DoD, for one thing. Is it also fair to say actually is behind in its needs in terms of the latest technology? Some of the platforms it supports, even updated ones, use technology, the problem of which is nobody makes them any longer. And it’s unlikely that a modern new high end submicron chipmaker would bother to make.

Liesl Folks Right. So one of the challenges for the Department of Defense is that it often has platforms that are in service for decades. For example, when we build an aircraft carrier, it’s going to be out there in the ocean for a long time. But the same is true for our air fleet and many other systems. And so as the DoD needs to upgrade the electronics on these platforms, it’s facing a problem where those legacy chipsets aren’t available anymore, and it’s expensive for them to redesign all of the electronics on something as complicated as an aircraft or, heaven forbid, an aircraft carrier. So it has a really unique set of problems.

Tom Temin It’s almost like a barbell problem because it has the legacy systems at one end. But at the front end is they’re talking about swarming drones and autonomous vehicles underwater, and in the air that a company platforms, image processing and all kinds of artificial intelligence processing. They’ll need that also. And that’s where you get into the overseas supply problem.

Liesl Folks Quite so. So it needs legacy chips and it needs absolutely leading edge chips as well for artificial intelligence, for quantum information systems, and for big data and data analytics problems.

Tom Temin So this gets into one of the classic [Defense Industrial Base (DIB)] problems, the defense industrial base, your fact finding team. What are you specifically recommending that they do?

Liesl Folks So our recommendations are fairly broad, but if I can focus in on a couple of things. One is that we ask that the DoD get back to working really closely with the commercial sector so that it is a fast follower of commercial technology, and always the first military use for new technologies as they’re rolling out of the fabrication lines. The goal there is to make sure that our defense systems are always ahead of our adversary systems. And that does require much closer partnerships between the DoD and the commercial sector than we’ve had in recent decades. What it does is bring us back to an earlier period in our history. In the 50s and 60s and 70s, the DoD was really close and really integrated with many of our commercial companies. That divergence was driven by those commercial companies finding they had plenty of market to work with in the commercial sector. They don’t really need the DoD so much anymore. And the DoD, on the other hand, put lots of barriers up to obviously with good reason to protect intellectual property, to protect national security. But it’s not easy doing business with the DoD. So those companies and the defense sector drifted apart. We’re asking them to come back together and become good partners. But to do that, the Department of Defense needs to change the way it does contracting, the way it does procurement so that it can be viewed once again as a good partner to the commercial sector.

Tom Temin And just one question about that, though. If it becomes a good partner because it needs these really high end circuitry to keep up to date and ahead, how does that run into the need for maybe export controls of that same technology or to keep it out of the wrong hands, and therefore the manufacturers don’t get the volumes they need to justify the investments it makes. They need to get to the first silicon of that type of chip.

Liesl Folks We make a very strong recommendation that the Department of Defense learn to accept more risk than it has in recent decades in these relationships. So going exactly to the heart of the question you just posed. It is going to need to relax its risk posture if it is going to be successful in doing this. It can’t control all of these technologies that the commercial sector is eager to deploy, and which, by the way, all of us Americans take great benefit from. We were talking before we got on this line about our cell phones and how amazing those technologies are. It doesn’t help us if we try and impede the progress of the commercial sector to serve the national defense needs. What we need instead is for the Department of Defense, for the DIB to be more flexible and accept more risk in its relationships with the commercial sector.

Tom Temin We were speaking with Dr. Liesl Folks. She’s an electrical engineering professor at the University of Arizona, and a member of the Committee on Global Microelectronics at the National Academies. And then there’s another kind of funny dividing line here, too, is often military specifications because of the environments in which these circuits operate are different from those in commercial. In fact, the only equivalent is under the hood of a car. Well, maybe as cars get electric, then they’re not so hot, and not so much vibration as they have traditionally had. So how do you divide it along that mil-spec versus commercial spec for the same functionality in a circuit?

Liesl Folks Yeah. So we on the committee heard a good amount of narrative from the commercial sector, that because different agencies within the Department of Defense have come up with their own specifications that are overly constraining, that is an area that we asked the Department of Defense to go and look at again. It is possible, indeed plausible, that we have overcomplicated the military specifications, even when we do understand completely that many microelectronics systems for the Department of Defense need to work and work well in extreme environments. Absolutely, that’s true. But we asked the Department of Defense to come up with a unified strategy rather than allowing each of the agency’s enforcers to come up with their own thinking on what the specification should be for these extreme environment chipsets. By doing that, you’re making it much easier for commercial vendors to come to the table and say, I can do that for you, I can solve that problem. I do want to put a particular piano in radiation hard chipsets, so chips that can work in the extreme environment of space in particular are vitally important to us on the commercial side, and on the national security side. We are all benefiting from the proliferation of satellites on orbit today and the explosion of activity from the commercial side. But we still have an enormous problem in this country in that it takes as much as five years to get a radiation hard chip to be certified as such. And that pace of delivery is too slow. So we asked the Department of Defense in fact, we asked the government to think about a national center of excellence that will facilitate and accelerate this process of designing and testing chips to work in space.

Tom Temin And then there’s the other issue of what is it in the environmental, let’s say, of the United States, such that all of the leading edge circuitry manufacturing left in the first place? It’s easy to say, well, there’s a lot of we have high cost labor. We have environmental regulations that are much more difficult than in some of the nations that now produce those chips, and so on. But is it necessarily only that the United States drove it out or was it drawn out in some manner?

Liesl Folks It was absolutely drawn out. And the US has benefited enormously from that process of globalization and specialization dramatically. We have all benefited from low cost microelectronics, low cost semiconductor chips. We see it in the phones we use, flat screen TVs and now incredibly inexpensive to buy washing machines, automobiles with all the sensors on them. So different nations around the world have done a really good job to incentivize manufacturing coming into those nations. And American companies have taken advantage of those offers of support. And that might be as simple as Tax reduction strategies, but also other supports, opportunities to get land assigned to them, regulatory support and so on. So the US has both benefited from that globalization and now finds it in an uncomfortable position, finds itself in an uncomfortable position which was exposed during the pandemic, which, if anyone remembers, really revealed that this whole ecosystem was a little fragile. And then we had supply chain disruptions. All of a sudden, Americans were not able to get the washing machines they wanted or the automobiles they wanted and so on and so on. So we’ve recognized that the U.S. has taken economic advantage of this globalization, and American companies have benefited enormously. But we’ve also created through a focus, a tight focus on globalization and not enough of a focus on resilience. We’ve also come to understand that we’ve put ourselves in a risky position. The CHIPS Act is absolutely doing a fabulous job of addressing that problem by encouraging companies to reassure manufacturing into the U.S..

Tom Temin All right. And besides the chip manufacturing, well, let’s back up a second. The chip design capability is still here in the United States. And we can cite NVIDIA, a U.S. company, but it doesn’t make the chips here, it only designs them here. But the making of the chips also drives the research into the next generation of lithography, and the next generation of gear needed expensive capital equipment needed to make it. So would there also be a spinoff benefit for returning high end manufacturing in that the other industries that it drives as a result of the research needed?

Liesl Folks Absolutely, Tom. You’ve hit the nail on the head. The CHIPS office absolutely understands that if we can reshore high end manufacturing, we also will benefit from synergies in research and development that come from co-location of the human capital, the brains trust that is working on advanced chip design, and manufacturing and all of the inputs to manufacturing as well. And then advanced packaging is a huge topic of interest right now.

Tom Temin And by the way, the rad hard idea and the environmental resistance, that all has to do with packaging.

Liesl Folks No, the chip itself has to be designed with special materials and design functionality in order to survive radiation. Because a lot of the types of radiation in space, but also, of course, near any actual radioactive sources, will go straight through the packaging and through that chip.

Tom Temin OK. I appreciate that correction. So the idea is that we need the manufacturing here to drive the ancillary industries that are necessary to sustain manufacturing here.

Liesl Folks Yes. And I point particularly to a topic of high interest, which is that of critical minerals. China has a very large footprint in the refining and production of just a suite of minerals and materials that are critically important to semiconductor manufacturing. And they’ve demonstrated that they’re willing to use supply of those materials as a lever in their diplomatic efforts. But that, again, creates a vulnerability for the semiconductor sector. And so the U.S. is actively working now to increase resilience across the critical minerals supply for the semiconductor sector as well.

Tom Temin We’re speaking with Dr. Liesl Folks. She’s an electrical engineering professor at the University of Arizona, and a member of the Committee on Global Microelectronics at the National Academies. And I wanted to ask you about something that’s in the report we haven’t gotten to yet, and that is we talked about updating specifications, procurement methodologies, making sure DoD’s demand signals are clear to the industry. But you are also recommending in this National Academies report actual investments by DoD in these capabilities. Talk more about what the committee thinks DoD should do from a spending standpoint.

Liesl Folks We urge the DoD to really focus significant investment on next generation semiconductor technologies, what many would term post [Complementary Metal-Oxide-Semiconductor (CMOS)] technologies, where CMOS is the standard fabrication technology that’s in use today to make transistors. We are urging the DoD to really focus on what comes next. This again comes to this idea that we want DoD to be absolutely engaged and at the forefront of what is coming down the commercial pipeline, so that DoD is watching and learning how those new technologies are going to help DoD in its mission, and so that DoD is ready to be one of the first customers and absolutely the first military customer for any new technologies that are coming through the commercial pipeline. By doing that, the DoD puts itself in a position of priority, and a position to always be in front of its adversaries in terms of the deployment of new technologies.

Tom Temin And my knowledge is a little bit dated, but whatever happened to gallium arsenide.

Liesl Folks Gallium arsenide still a big thing. Plenty of wide bandgap materials are still being used. Gallium arsenide is widely used by the Department of Defense still today and will continue into the future. Although now there are sort of new formulations that are similar to gallium arsenide, but even better. But gallium arsenide, which refers to actually the substrate, the wafer that semiconductors are built on. Gallium arsenide is still hugely important to the Department of Defense.

Tom Temin And as far as the depositions that you mentioned, beyond CMOS, which has been the kind of the standard for probably 30 years now or so, and it was because it uses less power, which is still a consideration as these get more dense. What is next after CMOS for those that are gearheads here?

Liesl Folks So my own view, not necessarily that of my entire committee, is that CMOS will continue in some form or other for a very long time. But it’s going to evolve. We’re going to see, for example, let’s look at one thing. Were likely to see the introduction of two dimensional materials. Think graphene, which is something you’ve probably talked about, which is a two dimensional layer of carbon. But it has quite a lot of structure to it. Materials like that have that two dimensional structure are very likely to become incorporated in the building of transistors, because they offer a much more exquisitely controlled capability in terms of the transistor performance. So we can already see that is likely to come into commercial production. Once we’ve solved problems like how do you do a really thin sheet of two dimensional material over a 12 inch waiver? So we’ve got to solve manufacturability issues and integration issues. But that’s an example of a technology that is likely to be an extension, if you like, of traditional CMOS technology.

Tom Temin And we should point out CMOS is Complementary Metal-Oxide- Semiconductor, that I remember from when I used to cover this stuff. And what about again, in terms of what DOD’s future needs will be and where it needs to channel investments? Nanotechnologies, could that come into the logic area? And what about Quantum, which is a whole different ballgame that nobody can really visualize?

Liesl Folks So almost every semiconductor that’s built today is fundamentally a nanotechnology. So we’re past that point. It’s all nanotechnology. So there’s actually an enormous synergy between the National Nanotechnology Initiative, which has been running now for 20 years in the U.S. and the post CMOS technology direction. Those things are exactly lined up with each other. It is the nanotechnology effort in this nation that is feeding the post CMOS technology pathways for the semiconductor sector. So that’s a checkmark on that one. It’s all nanotechnology at this point, or almost all. But let’s talk about quantum technologies for a moment. We are at the moment right now of being able to really see a pathway to deploying real quantum information systems, quantum logic and quantum cryptography systems in real world situations. And it’s absolutely imperative that DoD is the first user of those technologies as they arrive for use. That to me feels like it’s on the sort of five to 10 year horizon. And so the report urges the Department of Defense to be absolutely investing in those technologies, ensuring that they’re ready to ingest them and adopt them into their platforms as soon as possible.

Tom Temin And this report now has been out on the street, so to speak, for a few weeks. Any response from anyone in DoD saying, yep, we’d like to try this or are they too busy?

Liesl Folks Absolutely. We’ve had multiple conversations now with different committees and offices related to the DoD and within the DoD. And certainly great conversations and a lot of interest in thinking through now, how would you implement some of these recommendations? Some of them are hard to change the culture around risk and the practices around procurement and contracting. Those are big tasks. Nobody thinks this is easy, and we really are delighted that the response has been so warm and positive, and that people are really thinking about what would it take for us to move the U.S. in the Department of Defense and the DIB in these directions so that we can ensure that our defense systems are the best in the world.