Securing Our Future
Securing Our Future
SOF 018: Navigating the Semiconductor Revolution with Mike Wishart
Host Jeremy Hitchcock sits down with Michael Wishart, Chairman and CEO of eFabless.
Mike brings a wealth of experience to his role, having retired from a distinguished thirty-year career at Goldman Sachs, where he specialized in covering the technology industry as an investment banker. Currently serving on the board of Cypress Semiconductor, his extensive knowledge and strategic insights have also been evident during his tenure on the Spansion board.
Mike's expertise extends into the heart of the semiconductor industry, and his contributions have been instrumental in shaping its landscape. In this conversation, Mike will delve into the exciting Efabless Design Challenge, offering his unique perspective on the future of chip design and innovation.
0:00
Introduction
Welcome
The Efabless Vision
Manufacturing Methodology
Community Designers
The CHIPS Act
Advice
Outro
Transcript
This editable transcript was computer generated and might contain errors. People can also change the text after it was created.
Michael Wishart: Motivation right there the idea of two people in the garage was a real thing. Not just a metaphor. But by the time you got into the 90s companies were spending more time on all sorts of things that were other than Innovation, right? It was building factories, maybe or service and support or just maintaining the technology that they had it was getting incredibly expensive to innovate and at the same time the IPO Market had changed dramatically. So companies have to be larger and larger day go public. So what I observed and started to think about how to fix was this problem where the incentives to innovate we're no longer what they used to be. It costs way too much money to innovate right the cost of creating something was going up dramatically and then ultimately you had to actually build a whole company and that was very
Michael Wishart: Expensive returns for low so not surprisingly, I think chips I sees were sort of the canary and the coal mine you stop seeing companies getting formed and they weren't getting formed because Venture Capital wasn't investing in them for very good reason. So my approach this is I said, what's going on here?
Michael Wishart: How could you provide something that was different than that model of having to create a company? And so I didn't quite know what to call it. I sort of euphemistically called it sarcastically, I guess the Hollywood studio model because in the Hollywood business, right if you were an actor back then this is pretty YouTube right? You had to go convince some fat cat and some studio that you were the perfect person they were looking forward. If you couldn't do that you had zero right zero. Hope of bringing your renovation your creativity to Market and I felt the same thing is happening to chip. So my initial idea was how do you band a bunch of people together provide a little balance against the power of the very large companies into which you're trying to Market your ideas because they were the channel, to the end markets and so I called up my Hollywood studio model at time. I was running leaving Brothers technology business. I went to their Venture arms and boy, this is a great idea. Why don't we start
Michael Wishart: Business around this they said, once you go back to work Michael doing what we're paying you to do I moved over to Goldman. I was a partner at Goldman ultimately I chairman of their Tech group there. I brought the same idea to Goldman said hey Mike. what the heck is this? Why don't you go back to what we're paying you to do so, I kept thinking about this. And in the meantime, software was kind of getting it right, they were moving to these models like crowd.
Michael Wishart: Say I said, that's exactly what I'm talking about crowdsourcing in the context of posing a problem to a community of people who would then go solve that problem for you to more efficient way than ever been done before companies top coder were formed and in that model and then Along Came the flip side of that these Creator platforms, I think Instagram thank YouTube where people didn't have to respond to somebody else's request but had the barriers so obliterated that they could create their own content and then all of a sudden, you're seeing literally hundreds of thousands of over a thousand at 3,000 new apps per day being created into the app store because people could get them out there you were seeing, literally millions of YouTube's content that would be created, the speed of light going through there and in the meantime hard
00:05:00
Michael Wishart: Just sort of languishing, right, yeah you had maker fairs and that was fun. But in the chip industry you were going from literally hundreds of companies to 50 companies. So all this problem that I felt I had sort of seen was playing out in space and I said the guy this had to change so I was blessed. I got to know Jack Hughes who had started popcorner was sort of on the content side.
Michael Wishart: You sort of the innovator in these Community models NFL name Lucha alonza who was looking at it from the electronic design automation standpoint and they pattern matched. They said Mike we agree with your idea here, but you don't really have any technical capability. We've got this fellow over here Muhammad Kasam about a TI who's got the exact same vision you have as far as we can see and he has that technical capability. So why do we put the two of you together and in effect take Creator models and crowds. So let's say Creator models that were at a hundred percent level the area into non-physical product content because you can literally create and distribute and let's go solve the rest of the problem that relates to
Michael Wishart: Products so how once you've marshalled the whole community of people to innovate and innovate it and exponential level, how can you get them to go create and so a physical product and so out of that efabulous was born and that was about nine years ago. So I don't know probably more than you were looking for, but that's my journey. That's kind of how we got here.
Jeremy Hitchcock: I don't know. That's great and thinking about the open source Community allowing and democratizing a lot of innovation a lot of creativity and I think back to early Unix days capital people I think forget that that was all super close and it wasn't until a lot of those things got open source that there's a lot more Innovation and it's kind of interesting to watch almost like a retro film and Hardware about this Resurgence and anchoring with small electronics and the barrier between hardware and software. and from that I'm curious. How do you think about the vision free fabulous and the world in which companies like e fabulous this open Hardware? Movement, that's along the way now,…
Michael Wishart: right
Jeremy Hitchcock: how would you describe what he fabulous hopes to accomplish and 10 years from now. Where do you hope the microelectronics spaces?
Michael Wishart: Yeah, and thanks you touch on a very important thing that I didn't mention in this context of sort of Creator models and crowdsourcing which is open source, And so if you didn't have open source and software, it's hard to imagine being to the level of people who shared and accelerated like they did and all these other things. So what Muhammad started to do before we met was actually build the first open source ecosystem around ships, And it was a very creative novel radical thought and so Eddie fabulous. He'd actually started a company before we
Michael Wishart: Reformed it and got going and we built on what he did and we actually created the very first open source Flow Design flow. So people think of proprietary flows like Cadence and synopsis a mentor, we created a similar flow to design chips based entirely on open source, and then we didn't really stop there because that had been kind of a thing in terms of tools for a long time, but we actually then took elevated that we took that up to creating open source designs literally designs.
Michael Wishart: That people could start using all the same things you saw in open source software to accelerate design activity accelerate sharing and we put all that inside of a platform and then ultimately with Google we created that final layer which is to open source that whole interaction between your design and Manufacturing. So if you ask me sort of where I see things going I say a world where an open source increases by orders of magnitude the number of people who can actually design so rather than having 200,000 or so designers across 60 companies who can design chips. We imagine a world of literally millions of designers who can design chips with software like methodology and instead of 60 companies that sort of make chips. We imagine a world where anybody who wants to have a product that needs,
00:10:00
Michael Wishart: To we just specific envelope of power and performance think machine learning out at the edge can actually vary cost effectively create chips. And so we see this as a market grower not a market cannibalizer, there's going to continue to be markets for standard product chips or advancing very quickly and lots of terrific ways. But on top of that we see this explosion of custom silicon the sort of thing that people like Apple and Google and Facebook and Amazon are doing for their own purposes, but they do it with lots and lots of money by just flashing the cost of getting into that and the complexity of that we see bringing that to thousands and thousands of companies. So we see a rebirth of innovation. We see a rebirth of innovation that would not have been possible if you hadn't put the open source to the task and that's what we feel. We were put on Earth to go do
Jeremy Hitchcock: That's cool. Yeah, because a lot of people talk about the fabrication chipsack certainly is as one of the areas that's been most talked about has been on the capital side of building Fabs. I'm curious about the workforce side. We'll get that in a second.
Michael Wishart: Right sure, and that actually was something as I gotten sort of the last decade of my Investment Banking career. I spent a lot of time on one of the things I took a look at and it actually became influential in the various stacks of the chip industry was the implications of scale on manufacturing. So what happened over time in the chips, and this gets to exactly your question.
Michael Wishart: Is two things are happening at once one people went to more and more advanced nodes IE. They come the density packing. It's called Moore's Law effectively, right the density back in Dublin transistors, every two years this led to people who create faster for a long time less power intensive chips more effective chips. Second thing they did is they increased these way for sizes that the wafer chips are created in batches on a circular silicon desk called away for and they were doubling the way for size every whatever five or ten years. The reason they did that is you could manufacture twice as many chips for the same amount of reactor time. And so therefore you could have the cost. This is all rough Justice. I have to cost of making the chips.
Michael Wishart: Once you went to 300 millimeters the scale f of a size of manufacturing plant became so significant in order to manufacture those Wafers and the billions of dollars. That it really only was effective for the most high volume parts, So it really cemented this view to creating very high Standard products that were driven around high volume standard platforms first the mobile phone various networking platforms that drove everything and so Moore's Law continued to pay so there's extremely important but what drives more as law because of the cost of the manufacturing and ultimately the cost of the design is to amortize that design and that manufacturing content over the highest possible volumes.
Michael Wishart: So as a country, the US is understandably concerned as other countries are and staying on the Leading Edge of that advanced process Design Technology, right? Because that's what's driving so much of things. It's driving. a lot of the AI stuff at the core that's built on very advanced technology because of the economy of scale. I talked to before because these plants have to get bigger and bigger to do that. The economy is a scale drove the manufacturing plants to go to the most efficient places turns out tsmc became by Far and Away the global leader manufacturing started to move offshore. The country was concerned about it. And so in the form of chips act and other things the US United States is understandably doing what they can to sort of on Shore that manufacturing capability now from our perspective. We're not disputing the
00:15:00
Michael Wishart: importance of that in other countries will look at the same sorts of things. That's why you see China poor and billions of dollars into the same thing. But the reality is manufacturing people don't buy manufacturing, right you don't buy the phone or the TV because of the manufacturing plant you buy it because of does it fit your needs and so that to me comes back to chip design if you're not designing the chips to make the applications that
Michael Wishart: People want manufacturing doesn't matter. And so we're covering that design side and that's a significant portion of the chipset. It's not the largest dollar number because of the cost of the manufacturing plants, but it is a big part of it. And in that design thing you want to make design more efficient. And so there's a significant amount of money going to that think of the microelectronics commons lab to Fab. How do we take ideas in the lab move them to the Fab but it also means creating more designers, right? It's started to make chip design actually interesting again, right, if you're a software developer golly you could make an app in high school, So you get immediate gratification for your work. If you're a chip designer, if you're lucky you're making your first chip as a PhD it only the largest universities all that stuff. I was talking about before using open source and Community to dramatically reduce the design cost and all that.
Michael Wishart: Deploy that at education and so our portion of this problem is that education and Workforce Development side. So we go everywhere from Purdue with their Stars program is based on our platform UCLA and Stanford and other universes Berkeley beginning to build educational content the university level around our platform. But we also got a project in the community college space where people were through Funding they're looking to expand Community College engagement get people interested earlier to work down at the high school level. We've got a group in New York state working with, people from marginalized communities to introduce them to chip design in high school.
Michael Wishart: That wouldn't be possible. If you didn't have a platform like ours where we can literally Drive The Experience down through the hundreds of dollars a couple hundred dollars a design per student to design. So that's where we focus manufacturing is there to deliver the designs to make the applications happening. We're focusing on that front and making the designs happening and…
Jeremy Hitchcock: It's almost a completely different funnel.
Michael Wishart: getting enough people to do the designs.
Jeremy Hitchcock: And what's the size scope or complexity? what if using your App Store analogy where people are making thousands of apps per day if you're a high schooler. what sort of designs are you seeing people developing on? The platform today, what's the extent of where people can go when you talk about growing the community? Where's that growth coming from where we're the people going to show up from?
Michael Wishart: we think they're going to come from everywhere. So we have 10,000 over 10,000 registered members on our miniscule compared to Facebook, but in a world where it's sort of, only a couple hundred thousand chip designers in the world, begin to be a meaningful number most people come 50 countries now the UN and they range from, 20% of the market in the US 20% in India all the way down to countries with one person designing and who knows where those people even are but they're doing incredibly interesting stuff. So in the US we had a project that came out of an RPI team Why are they They were designing the chip because they had a specific idea for a new type of a watch and they needed the chip to do that. So they actually created the chip to make the application.
Michael Wishart: We have a team. From University of Missouri a woman there who created a chip for sleep apnea detection, So her perspective her observation is things like sudden infant death syndrome. we're connected to sleep apnea. So we need to have a more effective way to do that. So that's what she went and did people are doing influenza detectors.
00:20:00
Michael Wishart: We have people doing water quality management. We have people creating satellite radios epileptic seizure detection. So you begin to see a medical chain that's going through here a lot of design for good. There's a group that created a and soc for Ellie controller for LED lighting off the grid, I think in Bangladesh specifically think of places that are not on the grid they need to be very power efficient and then all the way down to somebody did it Donkey Kong keyboard thing, right? So it's whatever turns you on and when you can get the design cost down near nothing and let people who had never designed to chip before do it. Stuff happens and it's really really cool.
Jeremy Hitchcock: That is neat while Switching gears more towards the industry and wish you well on the journey fab list, but when you're thinking about chips act and one of the first major industrial policy bills, the United States has ever done, certainly a lot of conversation on Was it done in the right areas it, talk about chipsack specifically and from our prior conversations that understand that you're following along and along the way but What was right in that effort? What wasn't right future things along chips Aqua what should go in that? and where do you think the impacts are going to be?
Michael Wishart: So I think what would I say? So things like the chips act are very controversial in the semiconductor Community, So it's Government funding. If you go back if you read chips Wars, it's absolutely terrific fascinating book. But Chris Miller, talks about things where we're big government programs had created the applications which birth to some I conductor industry to a large extent is as we know it today as we move more into the 80s and 90s there were various other programs. It became extremely controversial right with many companies thinking these are great things others saying, Government funding leads to inefficient deployment of resources, all of that largely stopped, but I think as people were confronted with some of the geopolitical issues in the world today and the offshoring of the manufacturing a genuine concern came up about what
Michael Wishart: To what does that mean for I say us but in America as both a strategic in a commercial implication and out of that the chips Act was born. I'm sure there was a tremendous amount of compromising and the usual constructive destructive political engagement around creating big pools of capital but at the heart of it was to create Capital that was to go from the manufacturing side all the way to that the design and the Workforce Development as we talked about before I think that is a worthy Endeavor. I think it is initiative that is also being
Michael Wishart: Deployed in other countries around the world obviously China is very prominent people have seen that in Europe. There are also doing very very similar work in India. They're doing very similar work. So it's happening, around the world. I think this injection of capital.
Michael Wishart: Will possibly lead to probably a lot of manufacturing capacity. I think it will lead to an excitement level around designing and chips again, which I think is unequivocally a good thing. And so I think a lot of good will come out of it. I think time will tell as to how much good per dollar spent right, we'll see but I think they're on the right track with that. what I think is less visible in all this is how does that sustain? So we now have got a political will to go do this sort of thing and you've got a lot of countries and companies and individuals joining forces around it, but it is a finite pulse of activity at this point in time and the future will be dependent on that and the similar budget and success and the future one of the things that we're very conscious of any fabulous and
Michael Wishart: we started our company before there was any such thing as this chips act. The whole point was what you really need to do is change that whole ecosystem and development process so that the economics of it look more like software and digital content and things like that so when those dollars stop flowing to force feed this if you will and you're left with the new world that the propels itself, so it's that latter part that actually think is almost unaddressful other than as the government's governments and the US government in particular the chips, I think about it. We need to come out the other side with a different view a different economic Financial structure to these things that encourage Innovation going forward and don't require continuing votes.
00:25:00
Michael Wishart: In political context where the winds may change to keep them sustaining because if we do the latter you'll have a pulsive activity, but I think we'll be disappointed with the long term.
Jeremy Hitchcock: Yeah, it's interesting in chip Wars. I think Chris Miller writes about how and I think it was the late 80s that the then CEO of intel was spending a lot of his time actually doing I think there was some tongue and cheek references that today's chip DAC as chips act number two, and there was a lot of this conversation that was taking place is especially some of the more advanced nodes and methodologies around lithography especially became prevalent, and At that point the same type of thing was a lot of political polls. That led to kind of what we have now asml. Tsmc. We have a lot of us-based Eda and design-based companies,…
Michael Wishart: Yeah.
Jeremy Hitchcock: but that methodology is really a kind of a product of eighties nineties and then a couple decades of just ignoring it for lack of better word.
Michael Wishart: Yeah, we need to have the creation of semiconductors and they're deployment needs to be more effective. The fundamental issues that are happening are that we've gone from Little hundreds of Chip companies down now to 50 or 60 of any magnitude that is a trend that's being driven by the economics of that business. If we don't change the economics of that business, we're going to continue to move to fewer and fewer Innovation entities. And there is a direct analogy between Innovation entities and the network and not to get to wonky about this whole thing. But this is thing called netcast law that says the power of a network is exponentially related to the number of notes that are connected.
Michael Wishart: If you decrease the number of nodes of innovation, you will have a reverse exponential effect on Innovation. And that is a big problem. So we need to emerge out the other side with more people innovating more entities of innovating and if that doesn't happen then I just think this will not the exercise will not have been worth what it could have been.
Jeremy Hitchcock: Yeah, no interest in I guess just in wrapping up would love to hear your advice or counsel for somebody who's again listening to chipsack thinking about semiconductors microelectronics worse. Where's the white space to innovate in this area? And how do you look for signals of where people can go again looking at not a lot of big school chip designers or tool makers or manufacturers all super big. We're working. Somebody go in and make a difference on day one.
Michael Wishart: So that's a really cool question. So somebody's an interesting thing, So if you want to be part of that super leading Ed stuff, then a lot of that, go work for the asmls of the world doing lithography or the big companies drive those nodes and very Advanced ways. It's exciting. It's
Michael Wishart: it's all that but if you want to be any of the other of the somebody's who want to make a difference, then you need to look into the sort of than more space and more than more was a word that was couched to say, there's Innovation that can be done on older technologies that are really interesting think sensors thinks things like that. Those don't need to be made on advanced technology. They can be made on Old technology and you could do really cool things of census because that's where real life intersects with the digital world as we now move to things like machine learning. I think everybody should Look around everything that they interact with every day and decide where could putting some intelligence in this help me with what I do today as they identify that they should be like that woman at the University of Missouri and say in her case, how could I make sleep apnea detection a thing? I can do it making chips and I think when they do that
Michael Wishart: They'll work their way back to embracing and joining this open source Community. They should come back and embrace and join what we're doing in the ulous. But join this community of people who are discreet pieces of light all over the world who are coming together around this open source and Community infrastructure to share their learning because if they do that, they can innovate they can make a difference and that difference will literally be a product and they have maybe a product that only they consume it may be Product that a thousand people consume but it will be their own thing. And that's something that has not been possible until quite literally right now.
00:30:00
Jeremy Hitchcock: Very cool. appreciate the time Thank you for your insights and wisdom and look forward to seeing where you fabulous goes.
Michael Wishart: Right and Jeremy thank you for being a great partner. you joined us when it was very important to join us and your wisdom and guidance and support has been and spent terrific. So, thanks a lot. Yep. Bye.
Jeremy Hitchcock: Thank cool That's good.
Michael Wishart: Okay. …
Jeremy Hitchcock: Yeah, perfect.
Michael Wishart: let me know there's a bunch of my messages beeping in there. I couldn't shut them off.
Jeremy Hitchcock: Don't think they bled through. I didn't hear anything. So it's perfect.
Michael Wishart: Okay.
Jeremy Hitchcock: Yeah, so we do a scrub on it kind of just touch up the audio levels and do a little production stuff. Probably the next four or five days. We'll send you a link or maybe even sooner we'll send you a link and then we'll kind of publish it in our next cycle. So happy sent it to you and if you want to take a review and if there's anything you want to change your modified, let us know and go from there.
Michael Wishart: Okay, sounds good. Hey, and also I want to catch you and Lucio and start talking about people I'm a little slightly anxious…
Jeremy Hitchcock: Yep.
Michael Wishart: because you…
Jeremy Hitchcock: Okay.
Michael Wishart: I've got three Creek key people two two who got, some Financial constraints just to get it out there because Financial constraints in their life. One is…
Jeremy Hitchcock: Yeah.
Michael Wishart: but one is Andy and we need them I mean the future of this country. If we really talk dirty to the company's gonna center around those two right and…
Jeremy Hitchcock: yeah, it's
Michael Wishart: how they work together in their engagement.
Jeremy Hitchcock: yeah.
Michael Wishart: And so I want to be on my front foot and out of my back foot with them. So earlier you say early in the weeks are generally better for you than later.
Jeremy Hitchcock: Okay.
Michael Wishart: Is that okay?
Jeremy Hitchcock: Yeah, yeah.
Michael Wishart: Let me Circle up with Lucio come back with some dates and then let's see if we can do.
Jeremy Hitchcock: All…
Michael Wishart: Okay. Super Jeremy. Thanks.
Jeremy Hitchcock: Thanks. Have a good old hands. Yeah.
Michael Wishart: Yeah, thanks, bye-bye.
Meeting ended after 00:32:55 👋