National Security Challenges for Microelectronics

Transcript

thank you so much all for for joining uh today we're going to to just have a couple slides here talking about some of the national security challenges for microelectronics so the topics we hope to cover today are regarding the importance of microelectronics some of the recommended areas of focus within the microelectronics sector that we think of have really not been talked about as much and then how the government can address gaps that we see as we've analyzed these sectors and then finally a summary of conclusions so really the term microelectronics or semiconductors or integrated circuits those are used interchangeably technically an integrated circuit are semiconductors that contain one or more transistor while the term semiconductor itself refers to materials that have unique electrical magnetic and optical properties microelectronics are small electronic components built on semiconductor materials that have multiple transistors connected to form integrated circuits if you really look at microelectronics you see that they are the underpinning for almost every single technology revolution that you can think of from radios to computers to mobile phones displays arvr headsets wearables hearables autonomous systems hypersonic missiles and everything else that's important within our country's department of defense so if we look at the microelectronics kind of how a microelectronic is actually made just like anything else it starts with coming up with an idea designing the product sending that design or blueprint over to a place to manufacture it which is called a foundry and actually making the product packaging it up and sending it back today most of the designing is done in the us so we really own quite a large share in terms of the actual design and ip created of fabulous semiconductor chips but almost all of those designs themselves almost all the blueprints are actually sent over to asia to either taiwan for tsmc or korea for samsung so that's really what's happened this change has really happened over the last 20 years the significant change in the ship and shipment of chip money shift excuse me of chip manufacturing or foundries or fabrication to asia and this happened because of relative labor costs differing environmental sensitivities and government's willingness to help fund the large capital investment needed so the korean taiwanese could have these merchant fabrication facilities now people may say hey well there's intel now intel does have a foundry however the majority of their production is used for their own ships we know anecdotally even from intel capital and actually almost every single one of the startups in which they've invested in do not manufacture within intel's own foundry they manufacture it at tsmc so really the majority of the startups we see if not all are all fabulous and really have to to rely on tsmc most of the time or other facilities in asia for them to actually get their chips manufactured so this slide really talks about how we look at the semiconductor industry itself so on the left here is is a kind of the incutel microelectronics stack if i just go through these briefly at the bottom you have the foundries tools and materials as i mentioned um those are used to make you know if you akin to making a house right it those are used to make memory which might be like the storage or the closets in the house um or and then circuit blocks which might be akin to the wiring etc in the house that you can put together to make processors which might be a specific room in a house and then further put those things together to make a custom asic which might be the house itself and then the packet you know packaging so that you can eventually put this into a larger product and ship off now the areas we really think uh you know that haven't really been covered uh that are are the tools the materials and packaging which we'll cover here so i'm gonna start with the tools so the tools themselves as i mentioned you know most of these designs are sent off to asia and they're fabricated there now unlike a house where you could just use a hammer and nail the tools are really very very specific to this industry and they are really the choke point of any foundry's ability to fabricate a leading edge node so what i mean by leading edge is for microelectronics to get better and better you probably are all familiar with moore's law the transistor needs to get smaller and smaller and the only way you can do that is you need to pattern it with very fine tooling um in this case it's a company's uh call called asml so today there's this the industry for semiconductor duals is highly concentrated the good news is that three of the five top five providers are domestic kind of the bad news is the only company that can make this specialized tool to make those transistors really really small is asml and that's based in holland now what's going on from a national strategy point of view of course the us is dominant here but china is absolutely trying to cut into our lead they're trying to become independent of our tool vendors and have tried to invest in startups that compete directly with the tool vendors themselves if we look at our venture capital activity versus theirs in this space you will find a lot more startups that they're backing here whereas in the us you know there's only been about five investors that keep wanting to invest anywhere in this space um from our point of view what the u.s needs to do is really work with the european suppliers like asml and then continue to support homegrown capability that can help maintain our leadership position in this space if we go to the next slide i'm going to hand over to my colleague yan to go over some of the other two areas the first meeting materials yeah thanks jenny so most discussion so far around the eileen's discussion on tools and these advanced boundaries are focused on making these the smallest tiniest features at the leading edge node but making chips at a leading edge is very difficult and is only getting harder in fact this complexity translates to longer lead times lower yields which all means higher cost the reality though is that not all parts of a chip need to be made at the leading edge in fact not all parts of a chip needs to be even of the same material and we'll get back to those materials a little later so imagine if we can make a chip as easily as putting together lego blocks you know by selecting the most appropriate a component what we call a chiplet snapping them together or even stacking on top of each other now this technique is called advanced packaging and we think it has profound national security implications and should be thoroughly explored for to see what it can do for for the nation and here's some reasons why so for one because the whole chip doesn't have to be made in one foundry designers can then mix and match chiplets for example one can potentially make a unique chip that has commercial state of the art triplets from one foundry such as tsmc and placed next to chiplets made from secure trusted us facilities giving us a blend of you know the best of both worlds and a secure a chip that operates at the at the best of performance and two the us has a number of secure trusted facilities but they're all of older or legacy nodes advanced packaging actually enables better utilization of all those assets and can breathe new life into these existing resources but gave them a platform where they could actually participate in um and three advanced patching is is a lot simpler than advanced fabrication some of the lithography that eileen was mentioning in fact some estimates that this could be it is typically less than one tenth the cost and so it's a a low-cost way for us to get new capability without spending that upfront capital expenditure so advanced packaging really changes the focus from making chips and changes into assembling chips and the government should think about investing in the development of advanced packaging facilities um as an innovative approach to reducing our reliance on foreign foundries while providing a platform to better utilize existing domestic fabrication facilities especially those in the secure and trusted facilities if you go next time please but if we really want to disrupt the semiconductor industry we really need to start from the basics and look at the fundamental material building blocks and we see that in in the 1980s semiconductor production involved around you know eleven elements on the periodic table by 2015 semiconductor products involve over half of all elements on the periodic table now these include not just semiconductor materials like silicon but also a variety of gases metals and even rare earths which has been in the news recently now despite the explosion of all these materials that are being used today over time the industry itself has grown really around one material silicon and the surefire way to disrupt today's semiconductor supply chain is to change the materials the whole industry is based on this is not as outlandish as you might think because as it turns out silicon is a rather mediocre semiconductor there are ones that can operate at higher frequencies than silicon the ones that can handle much more power the ones that actually can emit light our magnetic and various other properties that we can exploit now not only would control these new novel materials exert an outside influence up and down the supply chain but development will also leverage the us dominant position in semiconductor equipment and tools that eileen just mentioned and so to achieve these outcomes really we think that funding a fundamental research in discovery of these next generation materials that can display silicon is is of the most importance but furthermore building an ecosystem that provides researchers access to commercial grade tools to quickly accelerate that material development process is going to be essential as well and finally leveraging these international partnerships to ensure access to material supply chains and processing capabilities will ensure our continued access to the various mechanicals and materials that we need in semiconductor production so you know we think of this process as what we call an innovation pipeline you know that is uh that starts at federal and private r d dollars being put into an idea then picked up by grants and private investors followed by the venture capital and large private equity funds and for many software and internet deals that you know we are all pretty much familiar with this pipeline is pretty robust and there are ample private sector investment investments because of the low initial cost large market in short time to return on investment for these types of deals but in commercializing science-based innovation what we call hard tech like semiconductors there's a widening gap in private sector investment because of the high upfront costs and long time it takes for iterating multiple prototypes in incorporating equipment and for design now this gap has left many government-funded ideas stranded despite their importance to national security incotel sees great opportunities and possibilities and new approaches to promote the commercialization process and address these investment gaps so one and we think that the government really needs his own investment vehicle to fill the gaps in private equity ecosystem where the technologies are of national importance and second secondly we think a microelectronic sandbox which is a facility that can provide researchers and startups hands-on access to production line equipment that mirrors commercial lines can be that can be adapted to explore new tools and new materials and components and processes greatly accelerate the process of getting ideas from lab to market and lastly many researchers who have phenomenal i technology don't pursue commercials commercial as a commercializing those ideas because they lack the entrepreneurial experience professional networks or business acumen needed to turn that technology into a viable commercial company and additionally government-based r d often engages with the commercial sector way too late in the product development process leading to products that are ill-suited for the commercial sector and at a time where it might be too late to pivot and that is why nktel as steve mentioned launched inky telemergy which is a new effort to support the commercialization uh government of government-funded r d by sharing our uh entrepreneurial venture capital and a deep hard tech insight through collaborations with government research organizations like darpa and so really inclusion in conclusion ensuring u.s innovation leadership in this sector will not be solved by onshoring supply chain capabilities alone in fact we think that the government really needs to engage the broader innovation ecosystem you know the same ones that develop the internet same ones that help send humans to the moon and as of yesterday the most advanced rover to mars you know that means investing in innovative and disruptive technologies like tools advanced packaging and materials but it also means giving the government investment tools itself to address the gaps in commercializing technology that is vital to national security and so with that we'll um you know thank thank you guys for your attention