THE WASHINGTON FOREIGN PRESS CENTER, WASHINGTON, D.C. (Virtual)
MODERATOR: Hello and welcome to the Foreign Press Center’s briefing on Semiconductors: From Lab to Fab and the Future of Semiconductors. My name is Wes Robertson, and I’m the moderator for today’s briefing. Our briefer today is Professor H.S. Philip Wong. Professor Wong is the Willard R. and Inez Kerr Bell professor in the School of Engineering at Stanford University. He joined Stanford as a professor of Electrical Engineering in 2004. From 1998 to 2004 he was with the IBM T.J. Watson Research Center. Professor Wong is a fellow of the IEEE and received the IEEE’s Electron Devices Society J.J. Ebers Award, the society’s highest honor to recognize outstanding technical contributions to the field of electron devices that have made a lasting impact.
He’s the founding faculty codirector of the Stanford SystemX Alliance, an industrial affiliate program focused on building systems; the faculty director of the Stanford Non-volatile Memory Technology Research Initiative; and the faculty director of the Stanford Nanofabrication Facility, a shared facility for the device – for device fabrication on the Stanford campus that serves academic, industrial, and governmental researchers across the U.S. and around the globe, sponsored in part by the National Science Foundation. So as you can see, Professor Wong is a busy guy, and we are pleased he made time to be with us here today.
And now for the ground rules. This briefing is on the record. The views expressed by the briefer are his own and do not necessarily reflect those of the Department of State or the U.S. Government. We will post a transcript and video of this briefing later today on our website, which is fpc.state.gov. Please make sure that your Zoom profile has your full name and the media outlet you represent.
Professor Wong will now give opening remarks and then we’ll open it up for questions. Over to you, sir.
MR WONG: Thank you, Wes, for the introduction. First of all, thank you for attending the briefing. It’s really wonderful to see so many people interested in semiconductors.
First, a little bit about my background. I’ve been a professor of electrical engineering at Stanford for 18 years. My area of research is semiconductor technology. Prior to coming to Stanford, I was with the IBM Research Center in New York for 16 years as a researcher and research manager on semiconductor technology and nanotechnology. Some of the research I worked on that – products that are in use today, such as the mobile phones and the computers that you are using today to join this Zoom call.
Today, I’d like to focus on the research and development or R&D aspects of semiconductors. I mentioned semiconductors several times by now, and some of you may wonder: what is semiconductor? That’s a fair question. The simple technical explanation of semiconductor is a material that can turn – can be turned into something that conducts or does not conduct electricity by applying an external input, something like turning on a –and off a light switch. But providing this technical explanation does not help much, since it does not tell us why one should care about semiconductors.
So let’s take a step back. I said earlier that – today that we are connecting through Zoom using mobile phones and computers. These devices are in turn connected through the internet. The internet is not just a piece of software. It is also a vast global network of computers, often called servers, that serve information to the users, connected by communication gears such as Wi-Fi routers, cell phone towers, your local internet service providers, and long-distance fiber optic links that go across oceans. These technologies run on so-called computer chips or just chips in short. Computer chips are made of semiconductors. So chips and semiconductor technology are often used synonymously.
What I just said is an illustration of the fact that the digital society is only possible with a comprehensive physical infrastructure at its foundation. To give an example, my mother once asked me where her photos go when she clicks the button to save her photos on the cloud. Of course, the photos do not go into the sky as some water vapor. They go into a network of computers and data storage devices in warehouse-sized facilities. And what makes up these computers, networking, and data storage devices? The computer chips, the semiconductors. Twelve out of the 17 of the United Nations Sustainable Development Goals rely on continued advances of information and communication technologies that are based on semiconductors. Every day, more than half of the world’s population use the internet, powered by semiconductors.
Beyond direct impacts on the economy and the national security, semiconductor technology plays a major role in enabling solutions toward many societal challenges: the digital transformation of healthcare, combating climate change, protecting the environment, broadening access to education and economic opportunity, and developing vaccines for COVID-19. Semiconductor technology is a foundational technology that has transformed the way we work, the way we interact with one another, and the way we enjoy our lives. And the demand for semiconductors going forward is insatiable.
So know you get the idea. Semiconductor technologies are foundational to all modern digital economies. This observation is prompting governments around the world to consider how to rejuvenate or build up their domestic ecosystems surrounding semiconductor technologies.
Proposals for supporting domestic semiconductor manufacturing and R&D are arriving from multiple corridors in response to initiatives such as the U.S. House passed America Competes Act – one of the many versions of what is commonly known as the CHIPS Act – and the European Chips Act. Of course, other governments such as Japan, South Korea, Taiwan, and China have their own initiatives.
The main focus of many of these proposals is on manufacturing and how to catch up with the competition. Most of the investments being talked about are exploiting existing technologies and knowledge, liking building more fabrication plants for semiconductors. These have short time horizons and lower risks.
However, as I’ve (inaudible) with my colleagues at Innovene and (inaudible) in an article in a publication, The Hill, in December last year, what we need today is groundbreaking innovations that will bring us to next generational chip technologies with exceptional performance that fulfills societal needs. This means embracing risk, fostering competition of ideas, and enabling the best new ones to get (inaudible).
Stanford professor James March called out this dilemma four decades ago in one of the seminal papers on innovation when he compared exploration and exploitation. Exploration is about new possibilities, new ideas, and new ways of doing things, while exploitation is about established certainties with finding existing products and processes. Because the returns from exploration tend to be long-term and have a high amount of risk, while the returns from exploitation are shorter term and likely more predictable, exploitation tends to crowd out exploration.
In semiconductors, we should not let catching up distract us from our real goal: leading the way. Building onshore fabrication plants is necessary, but not sufficient. The learning that comes from operating the plants is essential to ensure that new ideas are not unmanufacturable pipe dreams. But a new fabrication plant will become obsolete in four years if there is no R&D to sustain its relevance. Manufacturing and R&D go hand in hand. In the same way that doing R&D without manufacturing is akin to building a bridge to nowhere, exploitation of established certainties is not enough.
What is needed today are ways to accelerate global semiconductor technology advancement, which can be summarized as four simple ideas that we should keep in mind as we formulate policies and plans. Number one, force the competition and marketplace of ideas. Number two, seek leaders who are willing to take risks and ensure that accountability doesn’t lead to risk aversion. Number three, underwrite some of the risks of technology development. And finally, reduce the formidable barriers to commercializing new process technologies.
On the last topic, namely reducing the barriers to commercializing new process technologies, one of the bottle necks is the translation of laboratory innovations into manufacturable products, often called the lab-to-fab translation, where lab refers to small-scale laboratories such as those in universities, and fab refers to semiconductor manufacturing plants, commonly knowns as fabs, or fabrication plants.
As outlined in the white paper I published recently with colleagues at Stanford, UC Berkeley, Harvard, and Lam Research, we must adopt a mission-driven approach to build national or even global infrastructures to meet the lab-to-fab challenge.
The next-generation semiconductor R&D infrastructure must consist of multiple facilities designed in coordination around a set of technological moonshots, each requiring innovation across a substantial portion of the technology landscape. This mission-driven approach drives co-design and co-optimization across system, component, and technology dimensions to reach the ambitious metrics set by the moonshot.
Let me give two examples of moonshots. The first example is a realistic, interactive, virtual model of the physical world, such as Metaverse or Omniverse, but with 100,000-times improvements in computation or energy efficiencies needed. This is roughly equivalent to operating at energy efficiency levels comparable to the human brain.
A second moonshot is an augmented physical portal to the virtual world that operates at equal or better functionality, sensitivity, efficiency, size, and security comparable to its biological equivalents at a reasonable cost. In other words, it should weigh 10 grams or less, have 16 hours of battery life, provides data protection and security, and is priced in the same range as today’s mobile phones.
These moonshots share in common the need for accelerated pace of innovation in hardware not achievable via business as usual. The moonshot model has proven to be an effective means of fostering broad innovation and continuous renewal by the U.S. Government’s agencies such as DARPA and NASA, as well as private entities.
Similar to Stanford, universities around the world has a unique role to play. With the help of government, universities can serve as the mouth of the funnel that feeds research innovations into a technology pipeline by building up research facilities that can demonstrate product potentials at scale. By “at scale,” I mean at a scale that is relevant to practical applications.
To give an example, a typical computer chip today can have several billion transistors, with – billion with a “B” – while a university laboratory can at best produce a chip with a thousand transistors. In order to convince industry to further develop the university innovation, it would be necessary to do demonstrations at a scale that is far larger than what can be done in university laboratories today. A national or global R&D infrastructure that aims to facilitate such at-scale demonstrations can make a big difference. Research and development are done by people; universities are the training grounds for people who do the research and development. Filling the talent pipeline for semiconductors is a unique role universities around the world can play. Government support can go a long way toward that goal.
With this background, I’m happy to answer the questions you may have.
MODERATOR: Thank you very much, Professor Wong. We will now move to the Q&A portion of our briefing. If you have a question, please go to the participant field and virtually raise your hand. We will call on you, and you can unmute yourself and ask your question. You can also submit questions via the chat box.
I did have one question that was submitted in advance. This is from Stefan Beutelschacher from Welt, Germany: “How important is the U.S. as a location for global semiconductor production? Can you also talk about U.S.-China chip rivalry?”
MR WONG: Yes, thank you. This is a wonderful question. This is a topic that’s been discussed much today. And the – you asked about how important the U.S. is as a location for semiconductor manufacturing. Manufacturing needs to go, as I mentioned before, hand in hand with R&D. The U.S. has in the past and still is a powerhouse for the R&D aspects of semiconductors. So as I mentioned before, manufacturing and R&D go hand in hand. So in order for the U.S. to contribute to the world body of knowledge in terms of R&D for semiconductors as effectively as it was in the past, it has to have manufacturing on its shore very close to where the R&D is being done.
So just because of these aspects, that manufacturing and R&D go hand in hand, it is very important that U.S. maintain a very vibrant manufacturing infrastructure for semiconductors, because having that manufacturing infrastructure allow us to carry out effectively R&D, effectively carry out R&D, and also have – able to train the students that are required to further the R&D in the future, and also for the manufacturing of the chips itself.
MODERATOR: I don’t see any hands raised yet at this point. I was wondering if briefly you could also discuss the types of cooperation you have among other universities in U.S. and then also internationally. How is that cooperation?
MR WONG: Oh, you mean the cooperation of the – among the entities?
MR WONG: Well, the – semiconductor is really a global industry. No one country or region can have the entire ecosystem self-contained within their particular country or within their particular region. Take an example: Today the chip design is – a lot of the chip design happens in the U.S. A lot of chip manufacturing itself happens in Asia in countries such as South Korea, in Taiwan. And a lot of the packaging activities happens in other parts of Asia such as mainland China and Singapore, Malaysia, and so on.
And a lot of the materials, semiconductor-grade materials, come from Japan. And the tools and electronic design tools for designing chips mostly are – come from – innovated and developed in the U.S. And the fabrication tools that the semiconductor manufacturing plants are using to fabricate the chips comes from – mainly from the U.S., and also a lot come from Japan. And the – one of the important tools, the lithography tools, tools comes from the Netherlands in Europe. And the European companies design a lot and make a lot – design a lot of chips that goes into industrial use and also automotive uses and so on.
So each of the region has developed an ecosystem that has a – what you would call a specialty around that specialty area. And in order to have a robust ecosystem, you really need the entire world to collaborate and operate in synchronous with each other.
MODERATOR: Thank you so much. I actually don’t see any hands raised, so do you have any final – like sort of a closing statement you would like to make, sir, or any other things that you’d like to mention?
MR WONG: Yeah. First of all that the points I would like to really come – to make across is really the importance of semiconductor technology to the entire world’s not only economy but also solving some of the societal challenges that we have, we are facing. Things like combating climate change, you need powerful computers to model the climate. You also need a lot of sensors and a network of sensors and collecting the data. And you – and those are all made of computer chips.
So to solve many of the societal problems that we have going forward, we really need to have not only today’s semiconductor technology but also the continuous advancement of these technologies, because the needs from the applications that we are envisioning for semiconductors far exceeds the capability of today’s semiconductor. So that’s number one.
Number two really is how the semiconductor technology can continue to advance year after year. Many of us, as I read in the media, like to compare semiconductor technology to – as the new oil. But there is a very different attribute or quality that’s a huge difference between oil and semiconductors. First of all, oil is a commodity that is just in the ground and you need to dig it up, and it will stay there forever if you don’t dig it up. Semiconductor technology, on the other hand, is a constantly evolving technology, constantly advancing technology. A simple thing – way to think about is you cannot run today’s internet or your applications, your Zoom call, using cell phones that were made 20 years ago. So technology advancement is needed in order to capture the value of semiconductor technology.
So in other words, a constant – there is a constant need for research and development and translating those research and development into actual manufacturable products. So semiconductors is not oil. It is a technology that needs to be constantly renewed, and the constant renewal part is an absolutely important aspect of thinking about semiconductor economy and the ecosystem.
MODERATOR: Thank you. We do have a hand raised, so we do have a question that came in from one of our journalists, from Jacob Fromer, South China Morning Post, if you’d like to unmute yourself and ask your question.
QUESTION: Hi, thanks very much. Can you hear me okay?
MODERATOR: Yes, we can.
QUESTION: Okay, great. Thank you so much. I’m just wondering if you could talk about the CHIPS Act here in the U.S. and I guess how important is that for the kind of – the U.S.’s ability to compete and stay ahead of the game in this field, and also if – on the other side, if the CHIPS Act does not pass and that funding doesn’t become law, what would that mean sort of for the U.S.’s ability to stay competitive and ahead of other competitors in this industry? Thank you.
MR WONG: Thank you, Jacob. This is a wonderful question. I think the U.S. CHIPS Act and many other similar legislations around the world are absolutely important, and U.S. being one of the bigger economies in the world does have a leader effect in these aspects – namely, if the U.S. CHIPS Act is passed, then other countries would likely follow with similar or related legislation that would bolster their local or regional capability in semiconductors. So that’s – being a leader in this field, having passed this CHIPS Act is absolutely important in that sense.
Now, as far as how important it is into the U.S., it is quite important in many ways. First of all, it would – part of it, a good part of it, will bring in or manufacturing in the U.S. through these subsidies and encouragements for countries to build semiconductor manufacturing plants on U.S. soil. As I mentioned earlier, manufacturing and R&D goes hand in hand. If the U.S. has a very strong and robust semiconductor manufacturing on U.S. soil, that will ensure that the U.S. continue to have very strong R&D activities, and that would ensure that U.S. will continue to be able to provide the lead in innovating and coming up with the next newer generation of technology. And as I mentioned before, comparing chips with oil, that chips needs to be constantly renewed and needs to be constantly innovated on. So that’s very important aspects.
Now, a small portion of the CHIPS Act funding goes into R&D. That’s absolutely important in the R&D world, because as I mentioned earlier in the opening statement, today one of the biggest bottlenecks in semiconductor research and development is the lab-to-fab translation aspects of it. Today university laboratories are ill-equipped to do this lab-to-fab translation because the tools and the facilities they have do not – are not able to do these technology demonstration at scale, as I mentioned earlier. And to have a national facilities that – of course, we should also welcome other countries to participate in these national facilities to join into the R&D, because R&D is a competition of ideas and we should let the best ideas win out.
And the best ideas does not necessarily come in from the U.S.; it could come from other countries as well. So having a national infrastructure or a infrastructure that is – that would enable the researchers to be able to demonstrate the innovations at scale would be – would go a long way towards accelerating the pace of innovation.
So in summary, to answer your question, it is quite important on multiple levels. One is – and important from the level of being – leading the rest of the world in providing adequate investment into semiconductor technologies, that’s number one; number two, the funding would allow the U.S. to continue to have or build up this manufacturing capability on U.S. soil so that R&D and manufacturing can go hand in hand and have a strong impact in R&D. And finally, having a national infrastructure that is funded by these CHIPS Act would go a long way towards lab-to-fab translation and accelerating innovation.
MODERATOR: Thank you. We had another question that was submitted in the chat function. I’ll go ahead and read this. It’s a little bit long. It’s from Eva Schram, Het Financieele Daglad, the Netherlands. She says: “I was talking to an analyst who commented that the downside to bringing more manufacturing to the U.S. and Europe, basically decentralizing manufacturing, is that the process becomes less efficient for the manufacturer. There are upsides, of course, like national security, but downsides remain. Do you agree there’s tradeoffs there?”
MR WONG: That’s a fantastic question. Yes, there is a tradeoff, and you’re – the analyst that you talked to is absolutely right that in semiconductor manufacturing the larger scale that you have, the more efficiency that you have. And so a lot – concentrating a lot of the manufacturing ecosystem within a small region that you can access locally and very effectively is a very effective way to increase manufacturing efficiency and increase the yield and also increase profitability for the companies that does that. So that’s clearly has been demonstrated in the recent past.
Now, of course, there are downsides to all this. The downside is – several – and obviously life is always a tradeoff, a balance of everything that is part – on the balance. So the downside are several. One, as I mentioned before, R&D and manufacturing go hand in hand. If the entire world’s R&D, manufacturing is focused on one geographic location – just take it a very extreme example, right, so it’s one geographic location – that means that the rest of the world would have a hard time learning from that manufacturing experience to ensure that R&D does have relevancies to eventual manufacture of products.
So that’s very bad, because advancing R&D requires human capital, requires brains, requires people. And so people live everywhere around the world, so we need to be able to enable innovation around the world. And the innovation has to learn from the manufacturing aspects. So if all of the world’s manufacturing is focused or concentrated in one geographic location, to take an extreme example, that would not be very conducive to accelerating innovation going forward. That’s number one.
Of course, the other obvious downside is from a supply chain issue, that a single point of failure would cause tremendous disruption, in case there’s – if there’s a problem, and having decentralized manufacturing around the world would be good for – in a – from a supply chain stability point of view and a supply chain resiliency point of view.
Now of course, then you need to balance all this, right. And if you spread out the manufacturing, then your efficiency would be less, but then there are other things that you would gain by doing so. So there is always a balance, and where the right balance is is something that policymakers can make a contribution on.
MODERATOR: Thank you very much, professor. So I don’t see any more hands raised at this point or more questions, so this concludes our briefing. I want to give a special thanks to Professor Wong for sharing his time with us today and to those of you who participated. Thank you very much and have a good day.
MR WONG: Thank you.