Clearing a Path for Startups
|Aired:||April 20, 2023|
Alon Raphael of Femtometrix joins John Cole on Episode 7 of Circuit Talk: Funders and Founders. With unique insight into the difficulties of standing up a new technology within the semiconductor ecosystem, Raphael shares his expertise on both the length of the process as well as the necessary concessions and partnerships required along the way.
00:09 | 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 nonprofit 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 by Alon Raphael, CEO and founder of Femtometrix. Alon has years in the startup space before founding Femtometrix, including DreamWorks a LAM network technology, internet infrastructure startup. Before Femtometrix, he co-founded Tamarisk, which is a biotechnology diagnostics company that was also spun out of the University of California, Santa Barbara. So welcome to the show. Great to have you here. Alon.
01:13 | Alon Raphael
Thanks for having me.
01:15 | John
So we have quite a bit to talk about. But can you start by just telling me a little bit more about Femtometrix, you know, how you got started and what you're working on today and what what the company does?
01:25 | Alon Raphael
Sure, so we make a novel quality control tool for the semiconductor industry. And we were founded in May of 2011, as you mentioned, spun out of the University of California, Santa Barbara, known for surfing, partying and nanotechnology, of which I only got to do one of three. We then came over and, you know, moved into a house together out of the dorm room, very typical Silicon Valley story, kind of like the TV show, where we lived below the poverty line, but in an enjoyable way for five years developing the technology to a state where it could be viewed as viable by the larger entities. Received financing there from Samsung ventures who led our A round iteratively developed new machines with financing from SK Hynix and other large Korean company, leading to a strategic investment by a publicly traded Korean company Avaco, and the further deployment and adoption of our technology in an industrial sense.
02:25 | John Cole
The fundamental thing that Femtometrix does is metrology. And it's, I believe, it's in line in production. Is that right? Can you tell us a little bit more about that technology?
02:35 | Alon Raphael – COME BACK TO
Sure, so we have a unique technology from others. In the world today, it's not a me too in so far as its existing technology is put together, or an add AI to that which we have. It's something that's fundamentally distinct, insofar as a novel optical technique to be able to look into places of semiconductors and semiconductor materials that are otherwise blind spots for the existing technologies and any paired or AI combination.
03:03 | John Cole
So that's very valuable if you're running a fab, right, so that you can run more efficiently and also operate just at a deeper level.
03:11 | Alon Raphael
For fab, the benefits are multitudinous. One of the benefits is the ability to make new things at all development of new technologies. The second is the ability to do process control, or ensure that the yields and expected number of chips at the end come out as high as possible. And lastly, field reliability. How many times have we looked at our phones and said, gosh, this thing is broken again, I must have downloaded a dumb app, when in fact, it's actually a microchip that went wrong because of some pre-process in the factory issues that were not caught until the chip made it out into the field.
03:46 | John Cole
Yeah. So, you're a bit of a unicorn sort of picking equipment that goes into a fab. It's different from any other startup that we've talked to so far in Circuit Talk. Can you give us a sense of what that, what that path is like? And how I know that's quite challenging, right? So, it's hard to sort of innovate in that area. And you've got to essentially be the ant that dances with an elephant in some ways, right?
04:09 | Alon Raphael
Well, as far as I'm aware, at the leading fabs that we work with are the only new, fundamentally new, process control technology under consideration. Part of the challenge is that the process of getting a tool built given the risk profile, the years of difficulty, simple components, procurement and the multidisciplinary team necessary to make principle prototype are disabling but also the understanding of the marketplace knowing what your total addressable market is identifying what's known as a killer application or something that's going to have a stable larger, longer term market that makes those years of investment of life and effort financially and, and personally worthwhile is something that's a shifting landscape. So, ultimately, there's phases under which one has to operate – the first of which is, you know, do we have an idea? And that comes to you typically from a university or research lab? Secondly, can we make this idea into something that shows basic product market fit? And then from there the industrialization process in conjunction with one of the customers or multitudinous of the customers in a way where the first iterations can be productized into something that's industrially viable, and that that process typically takes between 10 to 15 years, and meaningful revenues may take 10 years or more, I can suggest that my company has been around for 12 years without meaningful revenues.
05:39 | John Cole
So a technology like this is fundamental to how fab operates. Why are fabs out there or larger companies out there finding these technologies like Fento’s built on top of and doing this themselves?
05:53 | Alon Raphael
Well, because it's too much risk, both for the larger companies who expect that that risk should be borne by the equipment companies and for the equipment companies who can't get commitment from their factory customers into the development. So, for instance, one could develop a vaccine in an analogy, that they spend a decade getting through FDA, etc, that that has very little market value, though, the people who would have been sick and get the inoculation and benefit from it. So, the fab is saying, Look, we don't really know if we need this thing in five years, or in 10 years. And we can't tell you that we're going to buy X number of tools for y dollars, because we really don't know. So we want you to go and fill the speculative machine for us. And we'll give you the benefit of the information necessary to customize it for our processes. And in the end, perhaps you'll sell some perhaps not you take the risk. Of course, the large equipment companies are saying, well, goodness, that sounds like a lot of risk for us. And as an executive, that's a 10 to 15 year timeline, before I get paid, why wouldn't I just iteratively develop the technologies I already have. Or, if you can identify a very key process area that we know is stable, that's going to have a huge market, probably something someone else is already doing - let's go there, because then we can deploy dollars in a way that have known returns if we're successful with clear targets. So as a result, one has to take an enormous speculative leap over 15 years and somehow get investors to have that continued vision without having key metrics, like total addressable market, in play prior to the development of the technology. So, it's a 15-year leap of faith that’s financed by who? And therein lies the challenge.
07:30 | John Cole
Yeah, that's a that's a that's a long leap of faith. I think most folks have a hard time financing, you know, the materials and the machinery that you're building go inside of a fab. What was it like to, you know, how long did it take you to actually get access to a fab to be able to sort of take, take your prototype, test it out in real life and get some data and show some value to a customer?
07:53 | Alon Raphael
Well, I wish it was that simple. So, the problem is, before one ever gets into a fab, is you have to build a machine. And the machine that's built in prototype format is guaranteed not to be good enough. But it's something that at least you can use proxy samples, hopefully, they're attainable, because sometimes the leading edge processes by the fabs are only manufacturable by the fabs and they don't want to let them out of the factory. So, it's a chicken and an egg problem where you may make a machine that can solve a problem for them, but they don't care enough to risk the secrecy of testing you. So you have to find a way to get a compelling proxy for whatever their problem is typically out of one of your competitors, a large process control tool company or something like that, to give you some baseline samples that are approximate enough to be able to use a kluge machine that was built to gain enough attention to begin a test process to raise the beginnings of capital to build what I would call an alpha tool or a prototype, it's not really fab ready, but at least shows that there's capability sets to be able to further refine and develop the application for deployment. So ultimately, to get a tool into a factory for the beginnings of the development process, one has to complete the development process. And this chicken and egg problem comes about in a way where even with that that kluge first alpha machine that does the basics, one gets into the factory and has insured have no anticipation of sales for an extended term, while they're improving the system, and servicing the customers every need during that term with all expenses borne by the entity that servicing them.
09:26 | John Cole
So, is this a this is a problem that all equipment developers face? Or is this specific to metrology? Or how does it shift when you sort of look up and down the stack?
09:36 | Alon Raphael
if you're a very large company, this is what you do for a living and you have other process lines and other tools and products that are going to come over and give you economies of scale to have revenues to support this development process for an extended term. So, it's less of a challenge for the entities that are already in business in this very oligarchical sense. However, for a new entrant with a novel technology, one has to have all of the support infrastructure behind them to give the fab the customer at the end assurances that the technology is of any validity at all independent of its technical validity. For instance, one needs to be able to show that it's reproducible, one needs to be able to show that they can mass produce, etc. And that's enormous infrastructure. So fundamentally, to be able to move to a step where a fab can consider working with a smaller entity, one has to have a partnership, or collaboration of some sort, in terms of development and manufacturability, with what's fundamentally one of their competitors in this oligarchical environment. So, that in mind, you know, to get to a place where a customer will try your system, you have to beg your competitor for the right to have that attempt.
10:47 | John
Well, a sounds like a problem that when we think about the CHIPS Act, and the Department of Commerce, it sounds like a problem that might be solved by some of the funding that they're looking to invest in different infrastructure, you know, a fab is kind of closed off to start up right now, right? If you were to say, be able to access an annex, or be able to access, I guess, a fab in North America more easily. What does that take off of your development timeline, like how much faster can you sort of get your, your product to market?
11:16 | Alon Raphael
It depends on the caliber of the fab. So, if one is trying to develop tools for leading edge processes, they need a leading edge fab. So if we're talking about a company like Micron or Intel, you know, opening the doors to research facilities where it's easy to place a system to work on novel processes, and being willing to share information to the extent that in a domestic capacity, that exchange of information, can iteratively develop a technique with them, then that's fantastic. But that would require these American entities to wish to collaborate share secrets, and potentially even purchase systems that are not entirely perfect in every way, such that companies can show revenues to be able to raise external capital. Because ultimately, if a small company turns over and says, look, I have no potential customers that I can see on the horizon for 10 to 15 years, I need $40 million. And I may or may not even get the information that I need to be able to serve as one of my customers, we're a constantly shifting market, it's an impossible sell. If you add a bonus point onto that, and we need to do it in Asia, you know, and they need to see a factory and infrastructure behind it to start. Well, of course, that's a non-starter. So that in mind, if there was help from MITRE, hey, this is a place where we make these kinds of tools together. And there's financing to carry you through some of these times. Well, a large company who will open their facilities to you in a way where there's shared information to be able to iteratively develop your process in conjunction with them. And there's, you know, a way to carry you through that development process, not just in terms of the capital, but also in terms of the other support networks necessary. A great optical scientist who has a novel technology that can truly change the game also may not be a mechanical engineer or know how to manage them. So having those sorts of generic skill sets in a centralized place alongside open communications with a domestic fab that shares their problems and seeks to solve them in conjunction with smaller entities in perhaps a limited development, pilot fab. But nonetheless, advanced production would certainly make a huge dent in this chicken and egg challenge.
13:24 | John Cole
So, access to facilities access to capital, and then access to specialized SMEs, that may be harder for startups to actually be able to interact with or to bring on right now is that...
13:36 | Alon Raphael
I might augment that with the ability to get parts. Very simply put, because ultimately, as a smaller company, when I go over to large entities that supply huge quantities to the oligarchical companies that exist, I have not yet third priority, they don't want my money. Because fundamentally, I'm a small customer who needs lots of customizations may never materialize into a larger customer and requires a lot of babysitting, frankly, to be able to do all these small things that we may not understand. And as a result, we're a last priority if they'll service at all. So getting some basic componentry, that's absolutely critical to the development of these tools is a sticking point because we don't have the economies of scale to be perceived of as a worthwhile investment for some of our suppliers.
14:23 | John Cole
Yeah, that makes sense. So, you mentioned you mentioned Asia a few times, I know that you've done a lot of work in Asia, and then the number of fabs you work with over there. Give us a sense of what they're doing right, what they're doing wrong in their startup ecosystem, and what we have to change maybe about ours to be effective.
14:42 | Along Raphael
I don't think they necessarily have a startup ecosystem. So I wouldn't say that there's a compare and contrast. I would say that there really aren't novel technologies. I'm not only the only new process control technology that's fundamentally different from others in the United States. I'm the only new one in the world. So. with that in mind, it's not as though in Taiwan, or in Korea, they have some thriving ecosystem where there's clear step function process to be able to support the development of novel technologies. They simply look to the equipment manufacturers, the American ones and say, well, you should know my problems and you guys make these systems, make the systems years in advance, collaborate with us, and fine tune your tools to our next steps and process needs. So that's a relationship that's held between mainly the large American but some global equipment companies, and the fabs in Asia and otherwise. And that locks out the prospect of novel technology unless there's room for entrepreneurs to be able to get the economies of scale necessary to be able to have a shot at it. Or alternatively, these large entities have their risk profiles reduced to be able to either a, finance these types of startups in a way that makes sense, or alternatively do the development on their own in a way where they're not bearing all of the risk.
15:57 | John Cole
So, how do you change? I mean, how do you change that? Is there a way to sort of facilitate trust, more trust, between larger entities within the ecosystem and smaller upstarts? Or is there something that we can put the government or say, a public private partnership, as the NSTC has kind of envisioned can change about that dynamic?
16:19 | Alon Raphael
Well, practically speaking, yes. But the change has to come from the fabs. So they have to say, look, we want this risk, we want these new developments, we acknowledge that the existing tool sets are not enough. And we want to try something new, as opposed to copy exactly, because if we copy exactly, and there's changes in that, which we need to do, fundamentally new architectures, new nodes, new materials, we can't necessarily use the existing tool sets to continue to manage those changes in an iterative process. And we can't rely on our existing, large scale, Titanic partners to come over and innovate fundamentally new things 10 years ago, that will allow for those changes to happen in a seamless or less complex fashion. So here, we see a prime example of that in gate all around, whereby the technologies necessary to really ramp and deploy, are aware. And of course, everybody in the industry is saying, well, I have this thing, and I have that thing. We've been using it, you know what it is? Why don't we just add it in with something else or throw some machine learning it? And I bet you we can solve the problems. And that may be true for 80% of them. But, what about the 20 that remain? And that's where it is that the introduction of novel technologies a decade ago, ones like mine, can help solve problems today. So, with that said, executives at companies have to be incentivized in a way to take long term risk on projects that, frankly, won't benefit them personally. And that's where the government can step in, hey, guys, at this big company, you're not going to get a quarterly bonus, this isn't going to help your yields in the short term or your reliability. But if you don't make this 15 year-long investment alongside these people, there's going to be something you need tomorrow, that isn't going to be there. And that's where the government and MITRE can help step in and bridge that gap because of the large development cycle for these technologies, and ensure that the tools of tomorrow are being built today. How can we expect an orchard to grow overnight? Yet we all want fruit?
18:21 | John Cole
Well, I think when you read what the Department of Commerce is putting out right now, the NSTC, one of the fundamental things is to build that sort of long-range roadmap. Hopefully, not fully decided or not fully determined by some of the larger entities. But everybody gets a say in what that roadmap looks like. So that we've sort of, you know, looking further on the horizon.
18:43 | Alon Raphael
It's a splintered ecosystem, and so far is there used to be things like the ITRS, where International Technology Roadmap for Semiconductors, groups would come together and talk about common problems. That's long since ended. Now every one of the major entities what all five of them, really six, maybe worldwide, have their own roadmaps, their own secrets and their own needs. So fundamentally, there are six, five ITRS. Having one singular ITRS will not necessarily solve the problem, because there are nonetheless common problems that are well understood in the industry. Amongst all of these entities. I give you a specific example in the case of my company, I don't know what one wants to call their specific variant of gate all around. They can call it MVC fit, they can call it ribbon bed, they can call it what they like. But fundamentally, some of the basic process steps are going to be the same. And they're finding that there are lack of control provisions over some of those basic process steps. So, putting that into a document and having everybody agree that it's a problem isn't necessarily going to solve the problem. What will solve it is having companies like mine present with the right capitalization, the ability to get through the basics of building that hardware to be present for the test when those things need are are needed, and having access to samples and fab environments to be able to give proof concepts to all of the companies worldwide, that these types of problems are solvable with novel ways. Now there is a file is there could be 100, companies like mine with a technology that never manifests, that's years of development and money went into. But the key is that one of them will. And after 15 years, if there's something you really need break glass, you know, and pull out that fire extinguisher, you've got to be able to have it there. And even if 15 of them, 100 of them are duds, and it was 15 years and a lot of money and a lot already, you don't have that one, then you don't get the ramped yields and the next generation technologies, etc. But who should bear that risk? Should it be the equipment companies to come over and make these things that have uncertain markets? Should it be the fabs to come over and take their already sometimes very lean margins in many cases, and pay for the development of technologies that may never manifest or yield any value. And that's where it is that having a third party come in and say, you know, we're going to work on this together to be able to produce things that our value later can come about. And that's frankly, I think how some of these entities that are now titans in the industry, especially in the equipment side, managed to get where they are because they went through that process in the early phases, when there was still more problems to solve, they weren't as complicated to solve, and fabs were more open to collaboration.
21:19 | John Cole
One thing we've been talking a little bit about a lot this week, over at MITRE Engenuity is about data sharing, and how there's an opportunity maybe for the NSTC to sort of be an aggregate for data, but especially think of like production tools and process information. Do you see any hope in that, like if there were a say, a, if there could be a black box system or a way to sort of share data around different processes, or at least allow fabs to sandbox some of that data so that either engineers can come work on it, startups might even possibility startups could have access to it to work on those problems? Do you see that ever existing within the world? Are we just a problem that we just can't overcome?
21:59 | Alon Raphael
It’s 100% solvable. In fact, there's already models worldwide of this working and producing novel technologies that power the industry today, let's take ASML coming out of imec as an example, an excellent research institution that embodies that which we're trying to achieve here. And with that in mind, there used to be Sematech, some years ago, which was, you know, an attempt at it. But I suspect that if if we were doing something like IMEC, we could produce our own ASMLs. And that's something that would take time, but nonetheless, you know, assuming that in particular our partner nations were willing to contribute to, and support, as they did with ASML and imec, we could replicate and excel. So with that in mind, you know, if we if we particularly focused it on the things that no one wants to do, because imec does a good job at what they do. What if we focused it on the places that are the least sexy, the least money and the most challenging, like process control? That's where it is that we could not only succeed, but also have competitive advantage in a in a parallel way where we both added unique value.
23:04 | John Cole
Did that result in sort of the Department of Commerce wants to see sort of revolutionary technology, not evolutionary technology, I think is their their byline, right. When you think of the developments in process control, is that truly going to leap frog us over the current paradigm of the current technology that sort of Asia dominates right now? And could that recenter industry here in North America? Or is it incremental change?
23:28 | Alon Raphael
I'm right here for that reason. And I have not given up over 12 years, because I haven't lost faith in this country's ability to do just what you said. So when times get hard, and I got pushed out of this country, quite literally to Asia to go develop something that I knew was needed for this nation's future and security, but did what needed doing in that regard. So I remained hopeful every day, that the opportunity to fix these problems here in the heartland is present. And I live my life that way. So with that in mind, I can suggest that my technology in particular is a shining example, in my eyes, and hopefully in the eyes of others, about what's possible. And with the right resources, we could have done it faster with the right information, we could have done it better. And problems that are facing American companies today, about the next generation of semiconductors that will enable them to remain competitive, would not be present in some instances had we have had the right pathways forward together. So with that said, I am 100% confident that process control is one of if not the keys, because we can as a nation spend a lot of time and money being really innovative and really smart and doing that which we do best designing things. But if we design all kinds of wonderful, unique and novel things, and we can make a prototype and hold it up and we can claim victory, that's wonderful. But how are we going to get those things practically into the economy in the hands of people who can use them to generate benefit because I love to see, you know the latest research from one, three letter acronym company in New York and they do things that are mind blowing amazing. I don't I mean, I'm constantly impressed yet, how are they going to take that one thing that they've done that sets the standard for what's possible and bring it into the hands of the consumer. And that's where it is process control is the missing link, because we can come up with all kinds of really cool, crazy stuff that we can't build. Yeah. And that's where it is in Asia. And we just tried to copy report. Well, China's trying to do that, too, not going so well.
25:28 | John Cole
Well, zooming out a little bit outside of metrology, just looking at the full stack for a second, that vision you kind of outlined of, hey, we can build a pathway for the later stage manufacturing, let's hold for a second assume that we can do that, right. And we kind of look earlier stages of venture launching, right? You all launched out of UCSB on a technology and you were able to innovate based on university of technology, we need to do more to encourage sort of technical entrepreneurs, as a before that the US still really dominates in terms of early stage R&D. Is that fair?
26:06 | Alon Raphael
It's accurate in every way, I would not put anybody in the world ahead in terms of the creativity, the heart, and simply the caliber of output of what the American university system in particular.
26:21 | John Cole
So what can we change in the ecosystem to get more of those sort of potential technical founders that are still in the lab out in launching commercial?
26:33 | Alon Raphael
They have to have hope. So reality is, is that when I started this, I had no idea what I was getting into, I just saw this, I saw wow, you know, there's changes in an industry that are going to require new technologies, you know, they're going three dimensional, there's novel materials. And there's no new tools at all for process control, no new technologies coming out. That doesn't make sense. So I went and investigated within the ecosystem of the universities for what technologies might be able to support the manufacture of these new types of devices. And I found one working thing. Yet, to take it from I have an idea and a little bit of a prototype and a paper published into this is something that really helps. Well, that's where it is that MITRE and the government can step in bridging between the fab wants 100 of them and, you know, cool story, bro.
27:28 | John Cole
I think the where we've identified through the valley of deaths are that first launch off the lab bench, right to to commercialize, but then the second ramp up to production, where you just got to have either fab time or you got gotta have iterative shots on goal with the technology to, to gather data and to modify and to bring it to scale sort of goes across the entire stack in terms of developments that are tools that you need to bring a product to market. But you know, you mentioned a few times along the way, you need a lot of money to sort of to build these companies, your venture funded, you told us a little bit about an early investment from Samsung, how did you find VCs in this space to fund your startup?
28:12 | Alon Raphael
I didn't, I am entirely strategically funded and angels, there are no financial investors that I'm aware of at this time, who would even co invest with those entities. So, the reality is, is that this is entirely a fab dominated and strategic dominated financing play, from the best of my knowledge. Now that said, if you're already a profitable company, well, like most financial institutions, they love handing umbrellas when it's sunny. But to get through the development process, you're reliant, as far as I know, on the fabs and on your competitors to some extent. So that's, you know, a bleak way to go about it. Nonetheless, you know, my, my solution to that was run lean, why do I need to make more than $25,000 a year? And those types of questions came about? So how are you going to incentivize somebody to leave the university to leave a cushy job, you know, where they make, you know, six figures at Google and have a clear trajectory towards the cafeteria that gives them nice food every day to eat ramen for five years and make $25,000 to live in a house with a bunch of guys, you know, into their 30s? I don't know. I mean, in my case, I did it because I went and saw what I thought was a pressing need a clear working solution, and add that confidence in the technology to a to a naive, at least do and what I ended up with was, you know, one working product five years later. So how would I convince somebody to do what I did? Well, I'd hide the truth from them and say, it's going to be easy. Alternatively, what I would do is I would give them the resources necessary a clear pathway to success with milestones along the way that enabled them financing and some balance in their life, such that they can see progress from their efforts and have confidence that even technical successes will be rewarded in a way that they can continue development. Because it could totally be that along the way, the economic benefit that's necessary for the fabs and the supporters of this is not had in the short term, and the vision necessary for the longer term value is simply eroded as a result, so giving them confidence, you're not going to be searching for the basics of what you need to live, let alone operate the company every month, it would be a great start. But until then, I don't know that I could tell them, please do leave your job, because what they're going to get into is something that they'll probably quit and should.
30:34 | John Cole
So definitely have a lot of convincing to do to get entrepreneurs off the bench and, and starting companies in terms of like changing the entire the game within the ecosystem and, and making some changes there. How do you get more venture capital into this?
30:48 | Alon Raphael
Well, somebody has to win. There's a graveyard of companies over decades, with billions invested in nothing but losers. So the reality is that, if there are no winners, why would you go it's like a casino where, you know, if you go in and you lose all the games, and if you've somehow managed to find a way to win a slot machine, they mug you on the way out, it's just simply not a viable strategy to compel people to come to your property. So, with that in mind, somebody's got to win. And that's where there has to be, you know, a pathway towards getting the shot at all, because just getting into that, because you know, you can get beaten on the way in, that's got to be removed, there's got to be some way of playing that slot machine and having it once in a while come up some sevens doesn't have to be a big winner, but enough that somebody says that was worth it. And on the way out, they can't get mugged. Those three things, a poster child would probably compel others, especially if the pathway was replicatable. and sustainable.
31:48 | John Cole
Yeah, no it's nothing begets success, like success, right? So a couple of couple of winners. But what it takes to get a few of those winners out is, is years off, and billions of dollars of investment, right?
31:57 | Alon Raphael
Not necessarily, I'm right here, I can do something that can make these people operate in a better way I can help the American microchip companies gain a competitive advantage over entities that they are seeking, seeking to compete with. And it really just takes a little bit of you know what, let's try once. And that, combined with you know, MITRE and the government giving a clear support pathway and a partnership with domestic fabs to enable the next generation of this type of entity would be I think enough, because then I could tell a very different story. Instead of kids don't leave your job, I could say here's how we do it. And let's do it again and again.
32:43 | John Cole
Let me understand one thing about product just coming back to what you're making right now. I've heard this story before that if you're a startup and to get time and a fab, say you're a fabulous company, for example, you almost have to convince, you got to show up at the fab with a customer that wants what you're making. Ideally, what you're doing is saying this is the change that we're going to make in the world. And here's how you know, here's who's going to pay for it. You think about that change that Femto is going to make in the world, who does that look like? Like who does that eventually benefit other than the fabs?
33:18 | Alon Raphael
Well, let's take that guy that's going to the fab. And he says, fab, I want you to make this, I have a customer for it. And the fab looks at that design and says, I don't care if you have enough customers to warrant an entire fab, I can't cost effectively build it, tell you what, I'll sell it to you per wafer at this very high price, because I know that I can't make this. See that that entity who's developing that new chip and getting those customers, they should be talking to me right at the beginning, so that when they go to that fab, they say, you know, I know this is a really difficult thing to make. I know it requires new materials and changes, I have a little bit of a pilot plant that I've set up. And I've made a couple of these with a process that actually works using novel technologies and things that are really deployable today. So, I'm not asking you to take a risk. I'm not asking you to have such economies of scale behind what I'm doing that you have to put a development budget out to see if it even works, or charge me so much money that it doesn't matter to you. I'm saying here it is in a way where you don't just have a new chip and some customers you have profitable new customer here, by virtue of what I can show you about your own processes. Yeah, that's where it is that having MITRE involved with companies like mine, at the early phases of the development of these novel technologies can help and I can name you know offhand key materials that would immediately benefit from process control at the early phases. For instance, silicon carbide and gallium nitride. There are critical problems in the basic processing of these materials that would novel process control could be resolved during the development of these chips themselves. So, with that said, it's not an either or changing the world with new materials and microchips requires the ability to build them practically, and actually started day one in conjunction with new technologies to do so.
35:06 | John Cole
CHIPS Act, it's got $50 billion overall and then $11 billion just for the NSDC. How do we make sure that money is used usefully and help startups?
35:15 | Alon Raphael
Well, one of the critical points is collaboration with the fabs. There's two, in particular, American companies that are larger scale and they're the leaders in their respective spaces. In many ways, were those companies to say, you know what, we're going to take this money, and we're going to build new factories, but we're also going to do is portion a part of it for a development facility that's not just about the money. Heck, MITRE, you guys pay for it. But what we're going to do is we're actually going to care. And we're going to develop things there in an open format way that enable companies that are connected to your entity to get access to the types of informations, types of samples, and types of future problems that would allow for us to have benefit later. So you guys bear the risk. And startups, you guys bear the risk. But you know, what, each time that you need something from us, that's not cash, we will give it to you. I can tell you, in particular, in my company, what I'm struggling to do is get information to calibrate my system in some respects to modern novel processes, that's held cloak and dagger, I mean, under lock by these large entities. So you know, were I able to position my system in a way where I could get a larger data set from one of the failure analysis labs, at one of these entities, I could quickly build models that would show that it does something have such unique and merit a value and merit that they would immediately adopt, but who's going to take that risk, who at that company is going to say, you know what, I'm a salaried person here. And I want to bring that system in, tie myself to that project and take responsibility for it. Because if it doesn't work, you know, in a reasonable timeframe, even I've really gotten some challenge, divorcing the individuals from the risk in these entities to allow them to have the flexibility to commercialize over the longer term is really what it is that I would see as the key component that could be delivered by the fabs in conjunction with the government to enable.
37:13 | John Cole
Yeah, they protect all that information. We're talking about data, they protect that for a reason, right, because their lifeblood does kind of depend on it. So whoever they handed over to, we've got to develop a process that respects that URL to let any data go will be a nonstarter, right?
37:29 | Alon Raphael
It's not just the data security, it's also the individual risk. So, some executive at a company has to come over I mean, let's put it on the fab hat for a second and say, you know, I want to come over and take part of my budget, or authorize something that may or may not work. And if it does work, it may not give liver value until I've retired. And I want to do that right now. And I understand that that's going to be impactful to my incentives in the short term. So now that we have our executives at entities that are completely disincentivized, from long term speculative projects, so even if one puts a lot of money into the startup company, your guy on the other side of the fab has to want to take political capital and other risks into it independently of that. So finding a way to divorce the executives at entities from the risk profile necessary to support this longer term development is going to be one of the keys because sure information, secrecy is important. But I can suggest that there's proxies that could be handed over that would be good enough for them to decide, do we want to advance technology A or B based on the results of the proxy. But that would not necessarily be one sample, it might be 100, or 1000. So having a pilot facility that uses a proxy for advanced technologies, that's easy to get a tool positioned into with executives there who are not incentivized by the immediate profits of the organization, but rather a long term targets that are frankly nebulous, take the risk with us as individuals because you're not punished for failure. And now, what we have is everyone sharing the risk financed by a MITRE or the government in a way where the individuals and the organizations are not punished for trying to.
39:08 | John Cole
Yeah, is there are quite a few sort of organizational challenges that tile into this too. Yeah.
39:12 | Alon Raphael
That's where MITRE can step in, and having a pilot fab that we could position those tools into, that the executives from those companies could come to and say I authorize that project, because it's not my budget. And we'll review the results. And the ones that are really great that are doing proxy work for us in a way that we think we should take it into the real factory. We'll talk about it. But that joint development work is currently done with the fabs through an enormously expensive and challenging process of great risk to the individuals at the fabs could be offloaded onto a third party entity that would allow them to select the working products that show validity for further development. And that pilot factory with proxy advanced processes that's easy to get a tool into would be a great benefit to the entire ecosystem.
39:56 | John Cole
I agree. That's a really great point. Yeah. I will really appreciate the change that Femtometrix is trying to make in the world and appreciate you coming on to talk about some of the challenges that you know startups face in the industry and that the ecosystem faces and that the Department of Commerce and the NSTC are going to face as we try to implement the CHIPS Act and reassure and bring back these the next generation of technology here to America. So thanks for, thanks for being the change that we all want to see. And thanks for coming on the show and talking to us about it.
40:27 | Alon Raphael
I appreciate your having me.