BREAKING: MITRE Engenuity Publishes their Blueprint for a Domestic Semiconductor Industry. Read the Paper.
IP Rights and Innovating Beyond Silicon
Aired: February 23, 2023
Ron Kelly, CEO of Ambature, joins John Cole on this episode of Circuit Talk: Funders and Founders. Kelly is at the forefront of innovation in IP licensing, heading a company working address a transition from silicon to superconductors. In this episode, he speaks about his path from the law to the semiconductor industry, how that has informed his growth, and where he sees the future going in a post-CHIPS Act ecosystem.
0:09 | John Cole
Welcome to Circuit Talk: Funders and Founders. I'm John Cole, Senior Manager on the semiconductor team at MITRE Engenuity. We are a non-profit dedicated to solving problems for a safer world. Our semiconductor team is hard at work meeting the nation's challenges around semiconductor breakthrough technologies and the CHIPS Act. Circuit Talk: Funders and Founders is part of MITRE's Circuit Talk podcast and video series and it elevates the revolutionary disruptive work being done by semiconductor entrepreneurs and investors. This is an exciting time to be working with semiconductor startups. The nation is waking up to just how critical they are to our national and economic security. I'm joined today on Circuit Talk: Funders and Founders by Ron Kelly, who's CEO of Ambature. Ron's a lawyer by training he studied law and business earning MBA and a JD from Dalhousie University. His IP training started at one of Canada's largest law firms and since he's led a number of companies as CEO where managing intellectual property was just a very important role within the company. Today, Ron is CEO of Ambature, an IP licensing company active and super conductive materials and the semiconductor space. So welcome, Ron, thanks for joining us.
1:19 | Ron Kelly
Thank you very much for inviting me. Appreciate it.
1:23 | John Cole
I gave I opened up with a few hints about your story. You've had an interesting journey here to talk about semiconductors today. And you're the first lawyer that we've had on Funders and Founders. So how'd you find your way from law school to where you are now with Ambature?
1:38 | Ron Kelly
Yeah, when I graduated from Dalhousie, I had a choice to make - one was to go into business right away, because I actually have to finance degrees, or to be a lawyer. And I was very interested in banking law. So I actually was recruited by one of the banks in Canada to go in-house. But I was also offered a job from one of the biggest banking firms in the country. So I figured it was just as easy to learn my trade as an outside lawyer and then go in-house as as opposed to the opposite, which is a little bit more difficult. So, I chose law first; I practiced law for seven years. And then, like many lawyers, you actually get recruited by one of your clients. That's exactly what happened to me. The company that was getting ready to go public and it was half of my practice. And they gave me the opportunity to stop actually practicing law on a day to day basis and help build the company, take it public beyond the board. And at the time, it was a perfect time for me actually to switch out of law and I decided to do it. That was about four companies ago. And I've been involved in multiple tech opportunities since then. And once you actually start to run a business, even if you have a legal background, it's very difficult to go back. You enjoy building things, I've always enjoyed building things. I enjoyed building my law practice. So now I've enjoyed building Ambature, which is now my fourth company.
3:04 | John Cole
That's great serial entrepreneurship. So tell us what are you building at Ambature? And what problem are you solving for your customers right now?
3:11 | Ron Kelly
So, so Ambature is a quantum platform, very deep in quantum mechanics, we are trying to address a number of problems that exist in classical computing, typical silicon based materials. So you know, if you go to many conferences in the semiconductor space now, you won't be there very long before they talk about Moore's Law, as you know, coming to its end. We need new architectures and we need new materials. Well, that's exactly what Ambature is, we use superconducting materials which can be used in addition to or a substitute for silicon. And we actually use a very unique architecture, which is along the a axis. In simple English, we build our materials at the atomic level vertically from the substrate, as opposed to where we make most things today, which is flat or the horizontal axis.
4:05 | John Cole
Can you unpack a little bit more about how the superconductors you're making now, how that's going to fit into what we typically think of as semiconductors, right? Thinking superconductors, I think the first thing I think of is, say, high tension transmission lines or other places where you know, you just want to be able to conduct with very little resistance, but how does this material fit into semiconductors?
4:27 | Ron Kelly
So we, we provide solutions in terms of the resistance problem, which most of you would understand as heat, heat in your computer, heat in your iPhone, your battery dying faster, that kind of problem. We also address the sensing issue sensing detecting is very important. That's usually done on the silicon substrate, and actually clean energy. So drilling down a little bit on the computing side. We're very, very fortunate. We can address multiple opportunities we can actually use our materials to make existing semiconductor silicon based chips faster because the semiconductor actually has slow movement of current through the material, superconductors have no resistance. So, you automatically have the ability to look at things like high speed interconnects. And the more chips, we try to jam on a typical substrate, your wires gets thinner and your heat gets bigger because of resistance. So we address from a pure compute perspective, the issue of need more power, more efficient power, more speed, and greater density. So all of these things are combined in a bolt on to silicon, it could be something like a chiplet, for example, or you could just use our materials. From a sensing perspective, many, many sensors that people use today for the IoT, that kind of stuff, much of that is silicon based. So, what you're having challenges with are the typical signal to noise ratio. superconductors just happened to be probably the best sensor in the entire electromagnetic spectrum. When we take pictures of deep space, for example, that is a superconductor, squid enhanced image that you're receiving. So if you can take advantage of things like this technology for Deep Space imaging, imagine what it could do in an MRI application. For example, imagine what you do in detecting signals, even in this small dynamic range of multiple signals coming into the front end of the cell phone base station. These are the kinds of things we focus upon. And the lack of resistance, remember, is also important for what we all care about, which is clean energy. Data centers, you know, need air conditioning, they need cold water coming through the floor, entire architecture just to deal with the heat, which is generated by these server farms. Whether it's cloud based or your typical data center, even a small data center in your company, they all generate heat, they need to be architected to deal with the cooling problem. So advanced computing, advanced sensing, detecting and clean energy, that's really what we're trying to focus on right now.
7:11 | John Cole
Well, for semiconductors, it sounds like you're gonna open up broader computing, faster computing, but also reduce the environmental footprint, maybe of some of the centers that use it, especially like you said, high use centers like data centers, things like that. When you talk to your customers, which which one of those is most important right now? Or where are you finding the most demand?
7:30 | Ron Kelly
Actually, it's actually in our center, center detector, a lot of it is actually military driven. The ability to detect and map different things is quite important to multiple militaries around the world right now. It has been enhanced by certain concerns in different countries, as I'm sure you're aware of. But any kind of superconductive sensor detector is really, really good. For example, if you are in a combat situation, you can make something called a velometer, which actually is an infrared application of a super conductive sensor, if you want it to be super conductive, where you can actually see the enemy before the enemy sees you. You can use it for enhanced radar. So you can see enemy aircraft coming before they can see you, for example, same thing for missiles. So right now, I would say 80% of our effort, and our pipeline, is really actually detection - super conductive chip sensor is actually called a squid, super conductive quantum interference device. That's that's what we spend a lot of our time architecting right now.
8:38 | John Cole
You mentioned chiplets a few times in sort of describing the technology. Can you unpack a little bit more about what they are for those of us that don't know and how they fit into your technology?
8:48 | Ron Kelly
Chiplets are very important. They're actually one of the major drivers in the entire semiconductor industry, right now - it's been around for a while. But under the CHIPS and Science Act under the PCAST report for the president in the White House, you see very specific provisions to enhance a chiplet platform a chiplet strategy, a chiplet ecosystem. In simple English, think of it as a modular approach to making integrated circuits. We use the Lego analogy, for example, where you put a Lego block on something and it gets bounded, but it's actually an excellent analogy. So chiplets can save a tremendous amount of money and from an IP perspective. It is a wonderful opportunity to get into the semiconductor industry, because you can just focus on creating what people call an IP block or a chiplet. Where you don't need to have to worry about making the wafer or making the integrated circuit growing and fabricating all the vias. All you need to do, is to do what's called wafer bonding wire bonding onto an existing chip. So AMD, Nvidia, Intel, all of these companies are doing it now. And great for young companies, particularly Ambature, because we have a very focused, highly functional chip, we don't have to do all of that work, we can just work with the major semi players or the major foundries and have them do, for lack of a better term, much of the heavy lifting, we just made a chiplet. It's actually good from their perspective as well, because our chiplets and most chiplets will be made offline. So that chip will be built somewhere else, it could be tested somewhere else. So if you're going to bond it on to a silicon chip to make a broader, more functional integrated circuit, that's already tested. And the last thing in the world you want is to create a 12 inch wafer, and then find out there's something wrong with it, it affects the yield. And chiplets are one way to do it. And it's a very important part of the future, particularly for IP based companies, because in many years, you may not have anything other than some IP, that one of the semis or many of the semis can't reproduce, because you've got a patent on it, for example, which is very important in our case.
11:04 | John Cole
Well coming back to your analogy about Legos, I like the sense that you can kind of snap different pieces together to come up with a new configuration or something that's not there before implies there's got to be a standard, right? Or something that everybody can kind of play with and build build to to snap it together. How's that coming?
11:23 | Ron Kelly
It's actually good. We have two wafer-bonding chiplet projects underway right now, one is actually started with people that are very experienced in the industry with superconductive chiplet strategies, wafer bonding, there's many words that go with it. But basically, we have mapped it out, we have already started one project, we're looking forward to the second. And it will be, not just for us, but for many small companies away an easy way for semiconductors to adopt what we're trying to accomplish with a very focused specific type of functionality that they don't own. And in our case, you can't own it unless you have a license from us because we have patents and pretty well, the bulk of the world, including almost every major area that makes chips.
12:13 | John Cole
Yeah. Well, so sort of putting on your lawyer hat for a second, thinking about IP and and protecting that right, you sort of mentioned that earlier, most firms in this space, they breathe and die on IP everywhere throughout the semiconductor industry. One of the challenges or the desires of the CHIP Act, if you will, is to sort of keep that IP here in the US. Do you think that's a good idea? It's an international sort of endeavor to put one of these together. But are we actually protecting American interest there?
12:43 | Ron Kelly
Yeah, I mean, I think we are. I mean, I support the the federal government's initiative to protect certain areas of our IP, that, you know, should not necessarily be available to people that may misuse it against us. You know, it's a very difficult question, because we all like trade, trade is good for us. But in the semiconductor space, we have found ourself in a precarious position, particularly on the high-end chips that we need for military purposes. Hence, the onshoring and policy behind the CHIPS and Science Actually. Let's get these foundries here. And let's start making all these chips in the United States.
13:24 | John Cole
Yeah, well, there's movement in other parts of the industry and some other startups we talked to focused on sort of open source and new tools and new products that are coming out that sort of coming out of an open source movement, and more semiconductor tool chains more IP. How do you think from sort of zooming back out away from Ambature for a second, how's that going to work for, for startups and for the semiconductor industry?
13:48 | Ron Kelly
Well, I think it may be difficult because it's a competitive industry, one, one large semiconductors competitive advantage is their access to tools and commercial equipment. And actually having a foundry, whether you have it, and you own it, like TSMC, or you have your chips made by someone like TSMC, if you're AMD, but for all the young companies, what I'm seeing and all the conversations I'm having is there is a willingness from the industry finally, to start opening up their PDKs and broadband access to EDA tools, and some, you know, sense of nurturing and trying to help young companies. Because at the end of the day, the CHIPS and Science Act is about, you know, helping the country, develop our own homegrown, larger semiconductor industry. So while it won't be easy, because of the competitive aspect of our industry, I'm seeing great opportunities to collaborate with larger companies and the CHIPS and Science Act, the fact that they're going to, you know, put up a lot of money for this particularly the chiplet platform, as we've already talked about, I'm actually quite optimistic about it.
14:57 | John Cole
As a younger man, Microsoft really dominated sort of the computing space, right? That was the more or less for a while the only operating system, it was closed source. And almost, not overnight, but pretty quickly over the last 20 years, Linux has come to really dominate the being sort of an open source platform, I wonder if there's a good analogy in there, or maybe there's not because it's, you know, hardware versus software. As these companies sort of open up PDKs, and tool chains, and more folks can come in and inspect them or Morpheus can come in and use them differently. You know, obviously, open source hasn't destroyed the software community, maybe quite the opposite. Right. Any thoughts on where that goes? Or where that lands? Do companies? I think Skywater is sort of leading the charge on open sourcing the PDKs.
15:46 | Ron Kelly
I know, there's a lot of discussion about that in the industry, as some PDKs are public, some are not. I know that there are discussions going on where people feel that these things need to be more open sourced, I mean, if you're a small company, you know, you need to know what you're trying to hit. It's a big deal to try to get TSMC, or Samsung, or Intel, or anybody to change their workflow in a foundry. So if you don't know what that workflow is, and what the requirements are, it's very difficult to hit it. So, that means you need to get engaged with them some level just to get the right person. And these are huge companies, they're are all very busy, you know, so to get the right person who can actually coach you as to what your process needs to look like, in order to integrate with a PDK for a large foundry. These are not little things. As I said, you know, your main company's competitive advantage is a barrier to entry to somebody else.
16:43 | John Cole
Well, switching gears a little bit and sort of putting back back on your startup hat, at Ambature, eventually you're going to have to engage larger fabs maybe for this developing process at scale, sharing the IP tricky things to do especially for startups point of view, not everyone's blessed with a an IP lawyer as CEO, right? So can you talk a little bit about that process as a startup, what it's like to sort of negotiate with a larger entity of fab and try to get in there and get some time on the fab to make your your chips?
17:15 | Ron Kelly
Well, most of these conversations, thankfully start with an NDA. Most companies are willing to sign an NDA. And if they're not, you may want to question whether you really want to do business with them. There are companies in the semiconductor industry around the world that are amenable to these kinds of conversations and signing NDAs, some are not. There's, there's a problem in many industries, where you, you own a foundry, for example, or you have many scientists doing what the person you're speaking to is doing. You don't want those scientists tainted with whatever is being disclosed to you. So, you have to be very careful about those kinds of relationships, you you want to work with them, you need to work with them. But you really want an NDA to protect what you're doing. So then you fall down to the next level, which is you only disclose what you absolutely have to disclose in the absence of an NDA. And hopefully, you have enough money and some decent patent lawyers where you can actually file patents, because you file a patent before you even discuss it with anybody outside of an NDA. That's pretty fundamental for any IP strategy. And if you have a large IP portfolio, the bigger it is, the easier it is to enforce. And it's harder for people to actually try to patent around it. So you know, again, I'm optimistic that these things are all going to be easier to deal with for small companies. Small companies are hard enough to build as it is, I've built many of them. And even as a corporate lawyer, not a lot of them are successful. So having the ability to work with larger players is important. I hope it becomes easier than it is today.
18:59 | John Cole
If I understand correctly, so chiplets kind of allows you to protect it a little bit more easily. Are you like lowering the exposure to the different technology? Different different players at the forger? How does that work?
19:12 | Ron Kelly
If if you have the money and you have the wherewithal, you will actually file a patent or a provisional patent immediately when you come up with the idea, that is the ultimate. And you know, patents are not for the faint of heart because they're very expensive predict if you want to, you know register them around the world. I mean, you care about Taiwan, you care about China, you care about South Korea, you care about Canada, you care about the U.S. and you want those markets and you want to be protected in those markets. But, you know, language translations are expensive, patent lawyers are expensive, maintenance fees are expensive. But so you have to find a balance about what is really important to you and you really need a patent to protect what you're trying to protect. In many respects, an NDA may just get you where you want to get to and then supplemented with trademarks and trade secrets and things like that. It's a mixed bag. You know, if you have the money, you should certainly consider it.
20:08 | John Cole
Is there? Is there anything that we can do? You know, we've got the CHIPS Act funding coming along. NSTC will be doing R&D, is there anything we can do in that space to help startups protect their IP? To sort of do that defensive work you're talking about?
20:24 | Ron Kelly
Absolutely. I mean, IP is a very difficult complex issue. I know a lot of people are discussing it in the context of the CHIPS and Science Act right now. We need to get the money out from the Department of Commerce as fast as possible to these young startups and the universities so we can start to innovate the way the policy initiative of the statute is supposed to be supported. The nurturing parts important as well, you know, don't underscore what it means to have people in the industry that had been around for 200, 30 years help you. Because these entrepreneurs come from different streams of life, university, they're in the industry, they just want to be in the semiconductor industry or the superconductor industry. I mean, you know, it is it is something that if the government can get the money out sooner rather than later that will really help providing access to incubators, for example, universities or otherwise, many of them actually have lawyers, they have people who've been through it before. So I think that's probably what they can do.
21:26 | John Cole
You talk to a bit about the different places just now about where entrepreneurs sort of emerge from and come into the semiconductor industry. Through your experience, you've seen more than a few startups kind of launch out of the lab space into the commercial space, or been there for a good deal of that journey. What can we do in the U.S. to sort of make it a better place, or a more nurturing ecosystem, if you will, to help R&D folks or folks with an idea in a lab, get it out of the lab and get into the commercial space?
21:56 | Ron Kelly
Yeah, yeah, it takes a village, it really does. If you're associated with the university that has the capability, and or an incubator, and we're going to get a lot more of them under the CHIPS Act, we hope, you know, you really, you really need to go and seek out these people who can help you. And and there are actually lots of people who were willing to help you - retired, you know, executives and engineers and physicists from the various semiconductor companies. And it, it's a lot of work. This is not an easy industry to break into, where you try to break into the industry matters very much. Are you on the wafer side? Are you on the foundry side, because, you know, a new foundries $20 billion, right? What I really like about the CHIPS Act, which is probably worth pointing out, is things like a chiplet, strategy chiplet ecosystem, that becoming a common platform, and I believe the date is 2025, that is wonderful for young companies, because you can eliminate all that capital cost, and just focus on new functionality that really matters that solves a real problem. I mean, the the process node innovation in semiconductors is really important. That's why if you can get it patented, and somehow protect it, and nurture it and get it to the point where you can create a chiplet and wafer bonded onto a silicon wafer, you are very, very far advanced.
23:25 | John Cole
Could you explain real quick, one thing I didn't follow there was the the new chiplets will reduce the capital costs to bring something to market, right? And what are we what are we cutting out when we use a chiplet versus a traditional route?
23:40 | Ron Kelly
Oh, I mean, you know, the ability to or the lack of need in order to create the entire full stack from creating the wafer and chip design, EDA tools, commercial equipment, commercial testing, access to a foundry. I mean, there's a reason why many companies in this industry are fabless, because the cost to actually be in this industry is enormous. That's why you have pure foundry police. So the chiplet lets you focus on what you're good at, hopefully patented it, hopefully get a good chip design, and then get someone who's willing to do the wafer bonding process for you. And then you're you're I mean, that's minuscule compared to what it would take to actually do this in the way that TSMC does.
24:29 | John Cole
When we were talking about some of the capabilities Ambature's superconductors provide, once we get it all together, there's the modularity of bringing a process and memory and sensors all together into one place, the lowering of power, the growth of compute, what sort of applications and use cases do you see this coming out to really attack that we may not be seeing right now?
24:52 | Ron Kelly
Sure, a big one will be data centers. I mean, you can imagine what's involved in building these data centers and some of them, they're just getting bigger and bigger. You know, the amount of heat that's generated in the server racks is not insignificant by architecture of these data centers is not necessarily the most efficient way to produce these things. We're very focused on reducing the heat. If you reduce the heat, you don't need all cooling, some cooling in order to cool superconducting chips. That's true. But if you think about the move to the edge, an edge data centers literally underneath a telecom pole, you can have all the advantages right there, you can have a super conductive sensor detector and amplifier and a cell phone base station, and you can have a small data center at the bottom of the pole. And that small data center can be cooled quite easily with liquid nitrogen, and have the tank refilled from time to time, these are the things we're looking at mean, you have huge private equity companies like KKR, that are actually now out buying datacenters. You have American Tower buying data centers. I mean, the American Tower just spent $12 billion on a bunch of data centers.
26:06 | John Cole
I read something just the other day that said that data centers in the U.S. use just as much power, you know, as I think the city of San Francisco does. So if we just, you know, reduce that or wipe that much power off the grid, that would be a huge benefit for for sustainability. When you think about that data center, those sitting at the bottom of the telecom pole. Think about the other sort of secondary effects that has on the market, too, right? You're the infrastructure that goes into backhaul, the data to processing centers, or the infrastructure it takes to get a lot of electricity and data and different places in the world that could have a significant impact.
26:42 | Ron Kelly
Yeah, absolutely. Just think about where does the electricity come from, that actually gets to the datacenter, through the datacenter, it has to be, I think it has to be stepped up and stepped down eight times in terms of voltage before it even hits the server, we're very focused on the actual server problem and the heat and the cooling required to do that, let alone speed. I mean, this is about speed, you want it you want it to be faster, you want it to be more energy efficient. And you know, the greater the density of these chips, the more heat is going to be the wires are thinner, it just gets worse and the need for computing. I saw one statistic from one of the big players, they said by 2035, at the rate we're going, we would need almost every bit of electricity in the world just to run our data centers. That's a problem.
27:29 | John Cole
It's not gonna work.
27:32 | Ron Kelly
Anyway, edge computing is a big opportunity for companies like us, you know, you're able to sense detect at the edge and process the data at the edge to make a decision. For example, for autonomous cars or autonomous trucks. This is very important. Camera feeds, radar feeds, processing that data in an instant is where it needs to be trying to get it through a telecom system through a data center and have it processed and send an answer back. That's a latency problem. You have that decision made in less than a split second and even that's not enough. So you know, all of these opportunities, medical imaging, there's a huge opportunities and superconductors and medical imaging. You know, MRIs are superconductor a lot of people don't even realize that. You know, again, there's some cooling involved, but the ability to detect and sense things in the body. For example, we're working on applications now in life sciences, that can detect proteins, using our materials, looking at making smaller MRI devices, where you can literally have the minute in a doctor's office, that is not just a detecting issue, that's a processing issue, because you want to be able to process the data on the spot in the doctor's office. It's enormous opportunities in the chip and sensing space for us and in the clean energy space. And, you know, as we're just beginning.
28:55 | John Cole
That's great. Well, Ron, thanks for taking the time to share so much about Ambature and the great work you're doing and the awesome difference that it's going to be in the world. We hope to have you back soon.
29:04 | Ron Kelly
It's been a pleasure. Thank you very much.
29:05 | John Cole