The Hard Thing about Hard Tech – Greg Smithies on the TRL 4-5 Challenge

How do you move from novel idea to working prototype in hard tech? And how can you bend the rules of the game to give you a better chance of winning?

Greg Smithies is a finance and operations leader who has worked with some of the world's most ambitious companies. He led finance and operations at Elon Musk's Neuralink and The Boring Company. He's also been an investor with Battery Ventures, BMW iVentures and Fifth Wall. He's helped raise over $2bn for startups, and over $950m for funds.

This episode focusses on TRL 4-5 – the prototype stage where many Deep Tech companies stall.

Greg shares how companies like Twelve (e-SAF / carbon transformation) signed a 14-year offtake agreement with British Airways parent company (IAG) while they were still at TRL 4.

You'll learn:

- Why storytelling starts earlier than most founders think

- The three types of hard tech – and why each needs a different capital strategy

- How Neuralink built teams before the company formally existed

- Why off-the-shelf components can be a major advantage

- The biggest hiring mistake – hiring industry insiders too early

Greg outlines three types of hard tech: fundamental new science (Neuralink), re-engineered systems (Boring Company), and economies of scale (solar, batteries, electrolysers). Each faces a different lab-to-market journey.

Essential listening for anyone navigating TRL 4-6

Chapter Notes:

00:00 Runway Experiments Milestones

01:13 Podcast Mission And Stages

03:12 Meet Greg Smithies

05:22 Operator Vs Investor Tradeoffs

08:06 Storytelling And North Star

11:52 Prototype Stage Communication

13:08 Vision First Neuralink Example

15:22 Building Diverse Technical Teams

18:19 Elon Speed And Timelines

22:51 Three Deep Tech Company Types

27:47 Capital Tools For Each Bucket

33:26 Selling Offtakes Early

36:25 Investor Evidence Vs Story

38:05 Defining Best Bets

38:55 Milestones for Next Round

41:23 Runway Experiments Tradeoffs

42:45 Showmanship and Tangibility

45:29 Hard Tech CEO Duality

49:35 TRL1 vs TRL2 Shift

55:14 TRL as Iterative Loop

57:00 Finding the Real Buyer

01:01:24 Who Should Sell It

01:03:39 When to Replace the CEO

01:06:38 Avoiding Industry Insider Trap

01:10:02 Closing Thoughts

Learn more about Lab to Market Leadership: https://www.deeptechleaders.com

Follow us on LinkedIn: https://www.linkedin.com/company/deeptechleaders

Transcript

Copy

0:00

And, and this actually goes to the, how, how the company has to operate internally, right? Because you say you're the CEO, or you're the head product manager, or you're the CFO, it's figuring out I have X amount of dollars. Um, I can, with those dollars, run Y number of experiments, and just call it an experiment, you know, just figuring out a certain path, like on my critical path. Um, how many of those things can I knock down that get me closer to a milestone that I think I can raise money around, right? Um, and that is a sort of three-party dance that you're trying to do, is dollars of runway, numbers of experiments, and what milestones they get me to, uh, and therefore, what am I actually going to raise money around? And sometimes you might say, "I am going to actually, instead of going down the critical path, I'm going to do a slight veer off course. Because with the money I have, I can hit a milestone that is maybe a little bit closer, but off to the left slightly, that I think I can raise money around, and I'm gonna spend my money doing that because I can raise money on it, even if it's slightly off-kilter from the true most, most obvious critical path to do." And that's the trade-off you have to be doing internally. Welcome to the Lab to Market Leadership Podcast. Too many advanced science-and-engineering companies fail to deliver their innovations from the lab to the market. We're on a mission to change that. My name is Chris Reichhelm, and I'm the founder and CEO of Deep Tech Leaders. Each week, we speak with some of the world's leading entrepreneurs, investors, corporates, and policymakers about what it takes to succeed on the lab-to-market journey. Join us There are five stages in the lab to market journey. There is the research stage, there is the prototype stage, there's the pilot stage, there is the first of a kind stage, and finally, there is the market stage, the stage everyone's trying to get to in the end. Today, we're gonna be talking about the second stage, the prototype stage. The prototype stage sits between TRL four and five. It is a notoriously difficult stage. It ends with something called the valley of death, which many of you, I'm sure, have heard about. In TRL four, we're taking a prototype of our technology and creating it in a lab environment. But in TRL five, we're taking a prototype of our technology and creating it in a relevant domain environment. This is our first exposure to the ultimate domain environment that we wanna be a part of. From TRL five to TRL six is the actual valley of death. And why is it called that? One of the reasons, it's a big step up in terms of technology progress, in terms of scale-up and engineering. Early manufacturing readiness levels kind of come into play here, and we're already advancing our thinking on commercialization pathways. It's also expensive to move up to that stage. Helping me make sense of what is one of the most difficult transition points in the whole lab to market journey today is Greg Smithies. Now, Greg is an operator and he's an investor, and he's operated at the highest level of both. He served as finance and operations manager for two of Elon Musk's companies, and he's been an investor in three funds, including Battery Ventures, BMW i Ventures, and Fifth Wall, where he raised nine hundred and fifty million dollars for their climate fund. Greg is one of the most experienced hard tech and deep tech investors, and so from the investment side, he understands this part of the journey extremely well. But he's also lived it and breathed it with someone like Elon Musk, so his insight into what makes a successful stage an unsuccessful stage is gonna be profound. I'm excited to get into this with him. Let's do it. Greg Smithies, thank you so much for joining me. Thanks for having me. Very excited for this conversation Now, at the outset of my introduction, which you wouldn't have seen, by the way, I did before we've had this conversation. Hopefully, hopefully you said some nice things. One or two. But, you know, I described you briefly as operator and investor, and I think that's probably the most basic kind of intro I could give for someone like you, 'cause you're not an ordinary operator, and you're not an ordinary investor, to be fair, either. You've done extremely well in both. An operator, you worked for Elon Musk for two of his companies, Bor- uh, at The Boring Company and Neuralink. Uh, as a, as an investor, you are Battery Ventures, BMW i Ventures, uh, Fifth, uh, the Fifth Wall F- uh, the Fifth Wall, uh, where you raised the fund there, too. So, I mean, you've had a lot of really incredible deep tech and hard tech experience. In that operator and kind of investor, uh, space, do you find that you gravitate towards one more naturally than the other? No, I, I think honestly this is, this is the, the conundrum why I keep bouncing, bouncing between the two because the… And I often get asked, uh, by people who are kind of making this decision. You know, like, "Oh, I, I have an engineering background, but I'm kind of interested in this investing thing," or,"I'm a finance guy, but I'm interested in, like, the building thing, you know. Which, which way should I go?" And, um, it's almost like the decision between going into industry versus doing your PhD, which is, um, you know, if you join a startup, you are going to go and spend the next three, five, seven, 10 years working on one problem and getting really deep into it, right? Like, the same as if you're doing your PhD. Versus being an investor, you're going to spend the next three, five, seven, 10 years, it's probably the, the same sort of amount of time because a venture fund is, like, a 10-year thing. Um, but you're going to spend that time doing lots and lots and lots of different stuff, right? So I think that there are pros and cons, and they scratch different parts of, of your brain and amygdala in terms of, uh, um, how excited you can be about it on a, on a typical day-to-day basis. I'll give you the drawbacks. Uh, investing, yes, every single day you're talking to someone really interesting and, uh, learning about something new and all of that stuff, but you're never getting that deep, right? And at the end of the day, frankly, uh, your job is coffee meetings and, uh, writing checks. Right? Maybe not that inspiring, right? But then the flip side is, uh, you could be stuck inside a very, very tough problem at a startup if you're on the operating side, uh, and your company fully hits a wall, right? That thing you thought you could build, maybe you can't build it or maybe you build it and nobody comes, right? So I think there, there are pros and cons to either side of it, and that's why I've never been able to make up my mind about which one to do. Yeah. So when you can't, just do both, right? Yeah. But the, you know, the high-low problem where I think as an investor you have to be able to do both. You have to stay high level, you have to be able to see the market, and then you have to be able to deep dive too. Look, in management you have to do that as well. You have to- Yeah … be comfortable taking that journey, the high-low journey. But the, the s- I guess it's the staying in one problem piece, staying and, and, and working through the different phases of that lab to market journey with one group of people, you know, one brand, one opportunity- Yeah ignoring most of the others, just concentrating on this. And that feels in some ways, it may not be harder, but it feels like it would be harder. Well, I think it's actually got a lot to do with storytelling here and, and this is the storytelling you are telling the people external to that company or that, that- Mm-hmm that thing you're working on, and also the storytelling you're telling yourself, right? Yeah. Um, you know, which is… And, and just to use an example here from, from The Boring Company, uh, I don't think you would get a whole bunch of people and investors and, uh, news coverage Excited about just taking a boring company and digging what is in effect a stormwater drain, you know, about a mile, um, and doing it on electricity, right? And if that was the internal story you were telling yourself as an employee there, you probably wouldn't last at that company, right? So a lot of the storytelling instead is, is about, well, actually, if we solve, uh, we solve this problem, we can clean up the world because, you know, uh, all transportation can be electric. Um, and potentially if you really want to go and dial this to 11, it's, you know, this is how we're going to, uh, merge the lava tunnels on Mars that all of humanity's gonna live inside, right? That's something that people can get excited about working on for 10 years, right? Or 10 or 20 years. Um, so a lot of that is just the storytelling that we do for ourselves and, and externally. And some people, obviously the example here being Elon, are incredibly good at making those sort of somewhat crazily big opportunities seem tangible. Um, and other people, which, uh, uh, I bump into a lot of these technical folks, uh, coming out and spinning out of universities, um, have been so focused on the thing they've been working on in the lab, right, that they've actually lost touch with, but why is this big and why is this important? And then those are the folks who ultimately after a couple of years tend to lose steam on it because the, they kind of miss the forest for the trees, right? Yeah. Do you have to… When you're in that company, in the same company, you're continuing along that journey, do you think you have to… Is part of this refreshing that story as you go? Finding a new twist- Uh, yeah … to it, if for nothing else than just to keep the inspiration and motivation, but making sure you don't let it get too far ahead of yourself because e- you know, equally you've gotta deliver on that as well at some point. Yeah. Yeah. It's the, um- So keeping, uh, just think of it like a, like a, I think if you, if you deal with people in branding, they'll, uh, say that you, you need what's called a brand house, right? Where, where you have like there is a guiding North Star of your overall company, company vision, and that ultimately probably shouldn't change. Um, and then day to day is how am I actually getting closer to that North Star? Like, what is the critical path today? What is the milestone we need to hit? I think ultimately, or, or in general, especially engineering type folks get really focused on the like what is the next step, and they lose track of the North Star, right? And so if anything, what we in business tend to do badly is reiterating the North Star enough and focusing far too much on like this morning's Kanban or the, this month's KPI or whatever it happens to be, right? And then we lose sight of the North Star. So in the best organizations that I've seen and been involved with, they're actually talking about the North Star a hell of a lot more than you would almost feel comfortable with, right? It's like it's literally that is the mantra every morning, the same way some people might get up and do their meditation, right? The CEOs of those companies are getting up, and even though they said it yesterday to all the employees, it's reminding everyone like, "This is actually why we're here. This is our North Star." Be it going to Mars, be it, uh, cleaning up the planet, whatever it happens to be. And because of that, what are we doing today, right? Um, but the, the natural tendency of most people with an engineering or a, or a, a scientific bent is to instead just focus on the what are we doing today portion, and, and you lose the forest for the trees again. Yeah. Yeah, yeah, yeah. We're gonna be talking about the prototype stage, TRL four five. We're talking about this valley of death. We're gonna touch on other things as well, but this is, uh, uh, recently we've been talking about the research stage. Now we're moving out of the lab. We're, we're, you know, in, you know, we've built our first prototype in the lab, then we build our first prototype in a domain relevant environment, and we're getting ready to kind of move into that pilot stage, maybe with an industrial sponsor. That's gonna require more capital, more management expertise. We're kind of early manufacturing readiness levels if we're in hard tech. Uh, we're, we're on the commercial… You, you know, we're actively considering the commercialization pathways as well. We're starting to think about application area and so on. So we're gonna come into this, but getting back to the communication piece, when do you start, when do you- See the best people start to really hone that communication piece, you know, the message. When do they start bringing that to bear inside their, inside their companies, and who is doing it? Um, it has to happen much, much earlier than, than you think, right? Which is e- exactly to my, to my previous point, which is, um, let, let's just say in, in the case of, um, of say, say Neuralink, right? Um, here the overall North Star vision is we ultimately want to make it so that you have massive high bandwidth input and output into the brain, because if you can do that, you can do everything from brain-to-brain communication through to, you know, you're in the matrix and you can, you know, jack in and, "I know kung fu." You know, that sort of thing, right? Like, it unlocks incredible things in, in the vision. Um, and in that case, we're actually trying to solve that North Star, but we are not sure what the technical path to get there is, right? And so you are starting out even from day one in a company like that with the vision rules, and we'll figure out the technical path to get there. Which is very different from the way that many, uh, especially universities spin out hard tech companies, uh, start, where they're, they're, you know, a technology looking for a problem, right? Like, we figured out this one really interesting thing in the lab. What can we do with it, right? Some of the best, uh, university professors who run their labs in the other way, where, you know, their lab discovery processes instead, um, they are looking for some bigger thing. You know, let's, let's unlock, unlock the boundless opportunities of the human brain and then see what we are gonna bump into whilst we do that. Those are the ones that tend to ultimately be, uh, more successful in the long run, um, because they are not necessarily tied to a particular, uh, solution looking for a problem, right? So there… So, so even in that research stage, if we think about it, if we're staying along the journey for a second, the principal investigators, the lead academics, they're often thinking about the more inspirational long-term goal. What could we create here? Yep. And then not getting too wedded to the technology at that point. Let's just think about- E- exactly… where we might go, and that's the way often that startups set up. So startups often see that, which is why there have been way more successful startups than spin-outs Or that could be a contributing factor to that stuff- Yeah, 100%. You know, you know, to the reason for that. Yeah. So, so if you look back to the founding of Neuralink, for example, what they tried to do is pull together, um, a bunch of people with different backgrounds and different types of technologies that might actually be the correct solution. But no one, uh, was married to what the correct solution might be. So for example, some, uh, some members of the team, uh, were looking at putting, um, electrodes in through your, your, uh, brain arteries, and then up in, in through the vascular system. Other people were looking at something which at the time I think they were calling neural lace, which was, is essentially open up the entire skull and put like a spider web of stuff on top of it, right? Other people were looking at what they've ended up doing now, which was these very, very thin hairs that go down into the brain. Um, but at the founding of the company, what they said was, "Let's just get together a whole bunch of people with disparate backgrounds and who are interested in different ways of sort of skinning this cat," for want of a, of a better term, and then, uh, sort of let the best pa- uh, the best path win, right? Because each one of them had trade-offs. And when you're trying to solve something very, very fundamentally difficult like this, uh, I think it behooves you to not be married to a certain path because chances are any path you pick is going to be wrong. Yeah. That, yeah, that's fundamental. That's very interesting. How, how diverse was this crew that they brought together? And it, it was everyone from, uh, uh, the specific, uh, um, sort of neurosurgeon and, and brain computer interface people all the way through to, "Well, we were pretty sure that whatever device is involved is probably going to be small. Uh, and therefore you're going to need a lot of people who are good at MEMS." So, uh, if you, uh, know about basically making very, very small, um, mechanical systems, right? Um, so a whole bunch of people who are good at MEMS, a whole bunch of people who are good at very like microfabrication around say, uh, 3D printing at the nano scale or, uh, laser ablating stuff so that you can take a very, uh, a, a block of metal and then cut it down to like a nanometer thin, but with a laser. So there were a couple of like fundamental things where, where, you know, the team was saying, "We're pretty sure it needs to be very small." What are technologies for very small stuff? MEMS, laser ablation, 3D printing, stuff like that. Get a bunch of people who are good at those, right? Um, what are the ways you can talk to the brain? You can be on the surface of it, you can be inside it, you can be in the vascular system. Let's get people who are good at each one of those, right? Um, you probably in this case are going to need someone who's good at chopping open heads, right? So you don't necessarily need to have 20 different ways of chopping ope- open heads. I'm pretty sure that the medical, uh, folks have figured that one out, right? So maybe you just need one person who's good at chopping open the heads, right? Um, but the point is, is you're, you're trying to get, you have hypotheses about the general fundamental things you're going to need for this, and then you back into what are the types of disciplines that you're going to need to do those things. How, how long did it take them to bring this, uh, this group of people together? Uh, that was, uh, I mean look, most things in Elon Musk world, uh, happen far faster than I think normal human beings can do it. So that entire, that entire process, and, and I will caveat this with I, I was brought in after this, after this fact. There's sort of like the preamble of doing the research, finding all the, uh, this team of people, talking them all into, into coming together. It was probably on like the nine months to a year sort of, sort of timeline before Neuralink actually even became a company. And a bunch of this is, is, is spoken about in, in, in some, some articles and things. But the point was, I think if, if you were going to do this, say as a spin out out of a university or something like that, uh, that process would probably take not nine months to a year. It might take four to five years, right? Yeah. Uh, so I think th- things did go a lot faster just simply because of Elon world. Um, but there was, uh, for Elon world nine months to a year is an, a infinite time. Yeah. That's a long time. That's a very, very long time. But he brought together this group of people before there was a company. So it was at this point just a, "Hey, we're thinking of doing this, so we should bring people together here. Let's gather up this group of people." And, and do you have a sense as to how long that kind of early research period lasted? Uh, that, that I actually don't know. Okay. Because the thing with, the thing with Elon, and, and again, this is written about like in the, in the Walter Isaacson thing, uh, he might pick up like a dog with a bone on something that he's really interested in and then stay awake for, you know, two weeks straight reading everything he can about it and ingesting all of it. Um, and again, doing, doing the sort of research in two weeks that might take someone else six months to do. Yeah. But then very rapidly moving from that to, "Okay, we should do it." And, uh, again, just to go to the, to the other example, uh, where I worked at The Boring Company, that effectively went from Elon is very, very pissed off with traffic on the 405 to, um, "Let's dig a hole in the SpaceX parking lots," on the order of under a month, right? Um, now he had already done a bunch of the sort of pre-work there because he had recently released that, um, uh, uh, his paper on the vacuum tunnels that they were gonna put, put, um, trains through. What was that called? Um, can't remember. But, but either way, the point being that, that he was sort of a- already had something similar in mind that they'd spent, you know, six months working on. But to go from, "I hate traffic," to, "We're gonna figure this out," to, "Let's dig a hole in the ground," yeah, under a month, right? Was that the Hyperloop? That was the Hyperloop. That the Hyperloop. Yeah. It was the Hyperloop. Yeah, it was the Hyperloop. Yeah, he was talking about that. Yeah. Um, okay. Wow, that's under a month. So God, that's extraordinary. That he's… Yeah. You know, and we know he moves very quickly. I mean, I think he built one of the world's largest supercomputers in something like three months. Yeah. And, and it can take year- it can take three years to build, uh, uh, uh, uh, something that large, uh, at that scale. So he moves incredibly quickly. The… Do… If we think about lab to market journeys for most deep tech or hard tech companies- And there's a s- you know, there's a similar rhythm, and often there are similar timescales. In Elon's world, uh, you know, time doesn't have the same impact, uh, that it does for, for us mere mortals. You know, when you were going through, you know, you know, when the companies were, you know, we, you know, were emerging from those early stages, and I appreciate they're gonna be different because Neuralink, you've got kind of an FDA qualification, regulatory- Yep you know, journey to go through. That's gonna be very structured. There's no way of hacking your way through that. If you wanna put something in people's brains, you're gonna have to take, you know, you're gonna have to, uh, you know, follow the regulatory pathway that's been set out. Um, with Boring, uh, uh, a company that's a little different, it's infrastructure, it requires a dif- you know, it, it brings with it, uh, different challenges. But was there ever a period where they felt like, you know, where it felt like, you know, you know,"Christ, we're gonna hit this valley of death soon. We're gonna be-" Yep."… scaling up. Now we've gotta get to kind of a pilot stage. How the heck are we gonna get to this next level? It's gonna require," even if the capitalization challenge isn't there because it's Elon, "but we're still gonna have to involve a much greater number of stakeholders. It's gonna get a lot more complex." Were those pressures there? Yeah, they, they were and, um, maybe I'll sort of expand the examples here to, to a bunch- Mm-hmm … of the other companies I've worked with and, and invested in. You know, like the, the power to liquids folks at Twelve trying to put- Yeah … CO2 into jet fuel and electric hydrogen and, and a number of these. I actually generally around the space I would think of almost kind of like three types of hard tech or deep tech companies. One would be like Neuralym- Neuralink is the, is the prototypical one, which is you are doing fundamental new science. There is very little you can pull off the shelf for it, um, and there is a massive scientific sort of mountain that you need to climb in order to do that thing, whatever it happens to be, right? That's the Neuralink-type thing. The, um, the Boring Company thing would be we're actually going to take a whole bunch of off-the-shelf stuff and arrange it in a system, uh, in a way that's going to be cheaper, better, faster, whatever it happens to be. But there isn't really fundamental science going on there. This is much more of a, an at-scale engineering type problem for that company. And then the third one might be just, uh, an economies of scale, so you wanna take a technology down the learning curve in terms of cost. This would have been like the solar companies or, um, in the case of like electric hydrogen, this is, this is taking your electrolyzer, uh, down that cost curve so that your cost per, you know, kilogram of hydrogen coming out the other side is, is… You're iterating, iterating, iterating, iterating, iterating. Those tend to be the sort of three, three buckets of all of these different types of companies, right? Um, and each one of them have different issues. Um, probably the easiest one to solve is the Boring Company type one where you are rearranging off-the-shelf components because you actually technically, whilst the system might be a four or five TRL, each underlying component is like a TRL a million. Yeah. Right? Yeah, yeah, yeah. And, and you're really just, just re- rearranging things. Yeah. You can take that attitude and go to a Neuralink type company and take pieces of the problem and make sure they aren't getting in the way because you've got TRL a million type things. So for example- Yeah … at Neuralink, um, we went through… There, there were a bunch of iterations on how does the microchip inside the head, uh, A, get charged, and B, talk to the outside world? Well, it turns out if you've got, you know, a, uh, um, an electric toothbrush, we've solved the charging problem in a wet environment, right? You're just gonna put, put a magnetic coil inside it and, and we're going to, uh, you know, punch, punch magnets through someone's brain in order to, to charge a thing. Very simple. TRL a million. How does it talk to the outside world? Well- Every device around you right now has Bluetooth. Why can't it just be Bluetooth, right? Now, of course, there are constraints to both of these things. You can only put milliwatts of power through, through the induction coil. Uh, you can only, uh, put, you know, a couple megabits per second through Bluetooth. So obviously there are constraints, but now you've taken, uh, things that other, you know, traditional, um, uh, say medical device companies might have spent a billion dollars inventing a new communications protocol to get, uh, stuff through the skull. Um, but you can just eliminate that because you're in startup mode, you're just trying to move quickly. Let's just use Bluetooth, right? Um, those sorts of things. You can take the attitude from the one type of company, which is the we're gonna rearrange things that are off the shelf type company, that's the boring com- uh, boring company type company, and apply them to these other fundamental science type companies and ensure that the actual difficult thing you're doing, the scope of the difficult thing you're doing, is as narrow as possible, and anything around that that is not critical to that critical path, you're gonna make as simple and as sort of like bone knuckle, knuckleheaded easy as possible by using things even if there are technical trade-offs to using Bluetooth in that case. Um, and then that third category, which are, you know, the things coming down the learning curve where you need to iterate, iterate, iterate. Um, the issue here is you t-tend to actually run into a project finance problem because to iterate as many times as you need, you very, uh, quickly need to probably parallelize and go like many, many, um, uh, manufacturing lines and each of those manufacturing lines going like roll to roll manufacturing of whatever it is. Uh, and typically that means each of your lines is gonna be like a couple hundred million bucks. Um, and so you have a very different set of problems in each of these three kinds of, kinds of companies as you are getting from the TRL to the TR- uh, TRL four to five and, and then up to six. Yeah. The, uh, uh- Most people when they think of that valley of death, and I really, by the way, I really like the way you kind of outline that and the way you distinguish between The different types of platforms that can come down the line, and in a way I think they're representative of a lot of the different types of hard tech or deep tech platforms that we see emerging. Yep. You've got the super scientific ones. Um, uh, you've got the ones that are engineering concerns, and then you've got- Yep … the ones that are gonna kind of, you know, w- where the economics and the whole process is gonna improve over time through the various iterations. But you need time, and you need, and you need capital in each. Are, are there different capital requirements at that kind of valley of death stage for each of these? Do you, you know, do you use different capitalization tools depending on what type of platform you are trying to develop? Regardless- Yes, so- … of the scale-up piece. Yeah. And I, and I appreciate the scale-up piece is going to impact that obviously, but, you know, do you see different tools that you can leverage depending on the platform? Ab- absolutely. So, um, let- let's start with, uh, the, with the Boring Company example, you know, the we're re-engineering a system, um, with off-the-shelf type stuff. This is the least capital intensive because the great thing is, you know, over the last 20 years we've built up a massive ecosystem of contract manufacturers, right? And very often these folks will be willing to spin up a small sort of test line for you just as long as you can guarantee a certain volume through it, and you don't even have to pay very much upfront for it, if anything, because many of these large companies, you know, like the Foxconns of the world and, and there are a bajillion of them now, um, uh, it's competitive enough, especially if you're in a, a, an economic sort of situation we are right now where, like, China has excess capacity. These folks will frankly fall over themselves to actually have something keeping one of their lines busy, right? Mm. Um, so that actually if, if you, i- if hypothetically you are an entrepreneur looking for different types of hard tech companies to start, this is probably the best from a capital intensity point of view because you have this whole ecosystem of contract manufacturers who can e- you're effectively outsourcing a lot of that cost to them, right? And they have massive balance sheets to be able to do it, um, and they're pretty damn good at doing this. Now, obviously you are not going to do-- or, or if you can get, even if you're doing one of these fundamental science companies, right? If you can say, "Well, I'm gonna do fundamental science for just this one piece of it, and the rest of it I can outsource to a contract manufacturer," wonderful, because then you kind of get, like, a best of both worlds type, uh, type thing. But if you're in one of those re-engineer existing technology type companies, the contract manufacturing ecosystem is wonderful these days. Um, the fundamental science ones, um, tend to be a total amount of money you have to set on fire before revenue is insane type problem. So think here, this is fusion, this is SMR nuclear, this is the, the portion of the actual brain computer interface inside the Neurolink chip that's going into, into the head. Um- Uh, the problem with these sorts of companies is y- yeah, you need like a billion dollars just to get to the start line, right? Like, just to say in fusion you have hit ignition, like a billion would be nothing, right? Actually, it's like 10 billion, right? Um, so in this case, your capital requirements are heavily tied to your storytelling ability, right? Because the true amount of money you need to do most of those fundamental science things is way more than venture capital, it's way more than any grant will give you. It's way more … Well, there's no project finance people who are gonna give you any money there. So it really comes down to, I actually like to call these businesses yacht racing for billionaires, which is basically you are you need to talk someone who is a bottomless pit of capital into believing your long-term vision because it is almost like a non- uh, non-logical thing for them to give you money. They actually have to believe fundamentally in the mission that you are doing because the fundamental science is going to take so much money just for you to get to, like, the start line, right? And the start line here being TRL6 and actually scaling up manufacturing and everything or whatever it happens to be, right? Yeah, yeah, yeah, yeah, yeah. Those are those two. The, the third bucket which is, you know, we wanna bring a technology down the cost curve. This is solar, this is heat pumps, this is batteries. Um, you can get it up to a TRL5 on r- very often relatively little money. You know, it's, it's you know, single millions to tens of millions to prove out your new perovskite coating or to prove out your new anode or whatever it happens to be inside the battery, and then you hit this massive infrastructure capital wall, which is now I need 500 million, a billion, 10 billion to build my gigafactory, right? And then there the, the, um, the entire story relies on I need to sell to project finance people, sell the story to them, and they actually care a lot less about your technology and a lot more about who's gonna buy it, right? Mm-hmm. So then your entire story becomes who's the offtaker, right? How hard is this contract that you've signed with them for, you know, a 10-year, $50 billion offtake? Because they actually, uh, they are not qualified to, uh, to figure out whether or not your technology is good. They are qualified-- Typically, these are much more kind of like finance Excel jockey type investors. They are very qualified to figure out if your offtaker is an S&P 500 company who is triple A rated by Moody's, you know, who has this amount of balance sheet. Is this likely to follow through on it? And then if they're giving you a $50 billion offtake, all they do is they do like a very massive haircut, uh, given the risk of your technology. So like, uh, the rough numbers I was thinking is probably for $5 billion of offtake, you probably get somewhere between $100 to $500 million of project finance money, right? That's like the rough, the rough ratio, and then that-- I know that's a very wide range, 100 to 500. But that, that wide range is entirely to do with like the quality of the offtake contract. Like is this a take or pay agreement? Is it like really solid? The quality of the, the size of that offtaker, you know, is this an S&P 500 company, triple A rated, blah, blah, blah, right? So it's actually a lot less about you and your technology, and it's much more a story about how hard that offtake is. But is that a TRL six, or is that when we're getting up to, hey, now we've got our, our, our commercial grade first of a kind? No, it's when you hit that TRL, TRL, uh, the, the best companies are thinking- Yeah about this at TRL four, which is basically you're saying- Wow … you're starting to sell your offtake agreements at TRL four to these people because you want to have them already lined up by the time you hit six. So I'll, I'll give you an example and, and a lot of this- Sweet Jesus… is, is public. Yeah. Right? So, so Twelve who are taking CO2 and turning it into jet fuel and plastics and a whole bunch of, a whole bunch of other things, they were announcing when they were at TRL four roughly, they were announcing the partnerships with PG&E, with, um… They did Mercedes full plastic in the A, A pillars. Now, these were very tight, small scoped projects. I think for PG&E it was a certain ingredient going into Tide, you know, the washing, washing detergent. Yeah, yeah. Um, for Mercedes it was an A frame pillar, like inside a specific model. So it was like one piece of plastic inside one model, which ends up being, you know, call it 100 tons of product. It's, it's not very much in the greater, greater scheme of things. But at TRL four saying, "We're going to announce these in partnership," because then by the time you get to your TRL five, then you're doing instead of the 100 ton one, you're doing the 1,000 ton one, right? And then by the time you get to TRL six, that's when you get the British Airways announcement, and I think it was close on$14 billion of offtake on the jet fuel. Right? But if they hadn't started those conversations at TRL four, um, they never would've gotten there, right? That's so- So- That seems so insanely early. I mean- Exactly … do you have enough… I mean, most companies don't have the stability in their technology at that point. Well, but this goes to you are selling the vision. So what you are selling in that case to, to, uh, say a Mercedes, um, or the PG&E is, "Look, today we, we can deliver a very small amount of whatever it happens to be. But ultimately in 20 years' time you are going to have no option other than us," right? For, in, in, in the case of, in the case of Twelve because I think you could, you can probably see that like in the case of plastics- plastics are directly coming out of the oil and gas industry's like cracking process, right? That is probably going away. We can argue about the timeline, but it's going away at some point. Is it 15 years? Probably not. Is it 30 years? Probably, like right? And so you've got entire industries who literally have no option, and they are the sorts of industries that think in decade terms, right? So you can start having that conversation with them, even if you're saying to them,"Look, I'm first gonna roll this commodity off my pipeline," or whatever it happens to be to you, um, in a decade's time. But we need to announce something today because otherwise you're screwed in 30 years, right? Um, so it's entirely storytelling- Yeah um, and getting, getting these long, long-thinking companies on board with it because it, it behooves them to. When you're wearing your investor hat and you're looking at some of these companies, then, you know, uh, at what stage are you-- how much of your decision, and I appreciate it's gonna change a little bit depending on who you're talking to, but how much of your decision is really, uh, caught up in where they, i- in what they can actually prove versus the power of the storytelling? You know, how much- Yeah … confidence do you need to have in the ability of the team if they've never built anything like this before? I mean- Yeah … they may be super smart. You, you know, let's say they're super smart. You bring that diverse team, but they haven't built anything like this, 'cause quite frankly, no one's built anything like this. And in hard tech- Yeah … and deep tech you see that situation a lot. So, you know, l- what is your decision-making process for, you know, you know, you know, for saying yes to a group that have never done this before in a space where it's never been done before? They're talking a g- great game. Are you looking for evidence, or is it just a be- power of the storytelling, the backgrounds of the people, I have a good feeling about this? No, it's, it's gotta be evidence-based because in this case, and I'll just, like reiterate on the, on the Twelve thing, they're not the only game in town, right? Like at the, at the same time that they're raising a round and call it their B round, you know, you've also got Air Company, you know, out there raising their A. Uh, there were a whole bunch of other sort of power to liquids and power to chemicals type, type companies out at the same time. So you do actually want to say like,"I'm, I'm picking the best out of- Mm … out of these." So there, there has to be some evidence base. Um- But what is the best there? I'll say the, the- But what is the best there- Yeah Greg? Is that- Ex- exactly … in terms of technology performance? Is that where they are? Is that in the, you know, the time, you know, in the progress they've made in the time they've had? Now how are you determining- Yeah, exactly … what the best is? The, the word best is doing a lot of work there right? So best has to be an amalgamation of likelihood to get offtake agreements, um, likelihood to hit the technical milestones, right? And probably likelihood to be able to like hire the team to actually do this as, as you scale up. Because right now, like the sci- core science team they have if you're looking at like an A round in a company where they're at like TRL three or four, is not the team who's going to take this to scale and actually deliver the stuff, right? And they probably don't have anybody in that team who's ever built, you know, a gigafactory or whatever it happens to be. Those people are… So you need to believe that this, that this CEO and this founding team can hire those sorts of people. You need to believe that they can sell, like sell ahead even at TRL four, sell these offtake agreements. Um, and then, and this is the third, uh, third piece where you can actually get some, some, uh, more hard data is do we think they're gonna hit whatever the next, uh, critical technical milestone is? And this is where I find many investors get lost again on what do we actually need them to do for the next round to come in? Basically to make sure that there is someone else down the line who will finance this company. It's not what do they need to do to ultimately deliver a $10 billion company, you know, pumping out millions of, of gallons of jet fuel. It is what technical milestone do they need to hit for the next tranche of money to come from someone else that's not me, right? Mm-hmm. And there, that's where you're looking for analogies. A lot of this is really just, um, and, and entrepreneurs should be thinking of this too. How is my technology actually very similar to something else on the market? So for example, in the case of, of any of these electrolyzer companies, the, the hydrogen guys, the power to liquids folks, um, it turns out that we have spent Probably close to 75 years building very large electrolyzers for things like chlorine manufacture, like chlorine electrolyzers, um, for the Haber-Bosch process, the things coming off the back of that, right? So you can actually go to industry experts who have built very similar things, maybe they've just got, uh, different, you know, chemist- chemistry involved in certain steps of it. But like, can you build an electrolyzer that pumps, you know, two megawatts through something, for example, right? And you can actually find people by analogy who can opine on whether or not you think they're going to get to whatever the next milestone is. So, you know, maybe it's, uh, do they need to put 1,000 layers in the stack of this electrolyzer? Has anybody ever done 1,000 layers in the stack of any electrolyzer at the time? No. But a lot of people had put 1,000 layers into a fuel cell stack, and it turns out that a fuel cell stack is just an electrolyzer running in reverse. Or better to say an electrolyzer is a fuel cell running in, running in reverse, right? So you go and find some people out of the fuel cell industry to tell you how crazy your power to liquid people are, right? Um, but you're really just focusing on what do I think the next critical technical milestone is for the next, uh, investor schmuck to come along and give these people the next tranche of money, right? But I think that question is such a, is such a strong one. Like, what is the- Yeah … what is the next technical milestone they need to meet for the next investor to come along and back this? I think that's absolute- And, and this actually goes to the, how- Yeah … how the company has to operate internally, right? Yeah. Because you say you're the CEO or you're the head- Yeah … product manager or you're the CFO, it's figuring out I have X amount of dollars, um, I can with those dollars run Y number of experiments, and just call an experiment, you know, just figuring out a certain path, like on my critical path. Um, how many of those things can I knock down that get me closer to a milestone that I think I can raise money around, right? Um, and that is a sort of three-party dance that you're trying to do, is dollars of runway, numbers of experiments, and what milestones they get me to, uh, and therefore, what am I actually going to raise money around. Yeah. And sometimes you might say, "I am going to actually, instead of going down the critical path, I'm going to do a slight veer off course. Because with the money I have, I can hit a milestone that is maybe a little bit closer, but off to the left slightly, that I think I can raise money around. And I'm gonna spend my money doing that because I can raise money on it, even if it's slightly off-kilter from the true most, most obvious critical path to do." But then hope- And that's the trade-off you have to be doing internally. Yeah, but then hope that that other external investor is gonna weigh that milestone the same way- Yeah … they would've weighed l- let's say a more orthodox milestone. Yeah, exactly. And I mean, this is, this is think of like Tesla with their original Roadster when they're getting it out there. Like- If you're saying, "I'm going to build a mass brand with the world's cheapest electric cars, and I'm gonna make electric cars, you know, that anybody, anybody can buy," doing a Lotus Elise-based two-seater that sells for, like, they were like a quarter million dollars, you know, inflation adjusted is probably close to like $300,000 yeah, these days, um, that, uh… Oh, and by the way, it's only gonna have like a 45-mile range and all of these things. Like, technically, that product is off of the critical path. But you know what it did? It, it accelerated like fucking lightning. It did, yeah. Right? Excuse the French. Yeah. It did. Right? It really did. And you put an investor in that car, and you do a, you know, quarter mile in it, they get out, and they write you a check, right? Like, that, that's it. Um, so, so I think you've gotta be pragmatic, and some of these founders are… They have that sort of natural showmanship, you know, the, the, the tangibility. It's when Nicholas from Twelve walks into an investor meeting, and he has little vials of, "Oh, and here's jet fuel, and here's plastic made out of our stuff," you know- Yeah to make it all feel tangible. Um, you need some of that, like, pizzazz even if you are selling something that is not quite on your critical path, but it makes it feel more tangible or more exciting or whatever it happens to be. Yeah. Well, these- So much then, based on everything you're saying, so much then really relies on that storytelling ability, the charisma, that showmanship, as you've just mentioned, of your founder CEO to be able to not only kind of tell that story, but to also define that story, to think in terms that are that ridiculously bold. What you've just kind of gone through the last 20 minutes is ridiculously bold. Yeah. Starting at TRL five to start thinking about offtake agreements by TRL six, which are worth the kind of money we're talking about with the kind of clients we're talking… I mean, that's huge. I… You know, th- that's very, very rare. Those kind of offtake agreements generally don't come if you're lucky, if you're lucky by TRL eight. Um- Yeah. It's, it turns out it's called hard tech for a reason. Yeah. Uh, yeah. It is. But, but that, but that whole, you know, what you've just described so wonderfully is like a massively jumped up, super accelerated, ambitious version of all of that, and then so much pressure is then gonna be on the back and shoulders of that founder CEO to think in those terms, to get their team inspired in those terms, and then to go off- Yeah and be able to deliver a message like that, that's going to excite everyone and is going to get the market looking. Yeah, uh, 100%. And the, uh, look, the best people… I, I do not want to, to play down the technical acumen, right? Like you still… This is the, uh, why truly good hard tech CEOs are so difficult to come by, because you need this duality of big picture crazy storytelling, plus the incredible technical acumen to, you know, figure out what the critical path is. And as you grow, yes, you can, uh, build out the team and you'll have product and program managers who can figure out a lot more of that critical path. But that first, you know, TRL one through five t- tends to just be the CEO is doing both sides of this coin, right? Yeah. And then you've got a, um, another nuance here that many of these sort of founders are coming out of a PhD environment where if anything, you are told to not oversell, right? You're en- actively encouraged to not do this. Um, and this is more just like a, an anecdote I'll say is, um, you almost need these sorts of founders, um, and these sorts of teams to have gone through one failed company in a spinout first- Before they've gotten almost like the PhD beaten out of them, right? And then they can come back in their second hard tech company, um, with this extra sort of nuance of being able to, you know, be the storyteller, but also good at, uh, good at the technical stuff, right? Um, tho- those first time spinout technical founders tend to be overly focused on, "I have a small solution that can… is looking for a problem, and if, you know, I build this, it's worth a couple million bucks. Please give me, you know, 500K." Um, and that just like, that, that story just, just doesn't sell. It- that dog doesn't hunt, right? Um, or, or, or, or they, they raise some capital from, you know, the grants and things and they, they, you know, beg, borrow and steal it together. Um, but then the company doesn't go anywhere because they're missing the pizzazz and the showmanship, right? But second time around, some of them figure it out And so what is happening that in that failure? Is it that, "Shit, I should have been way bolder. What the heck was I doing?" Yeah. Is it that? It, it's that, and then it just goes back if we wanna like truly dig in on- Yeah you know, the, the theme of this, this podcast today, which is, you know, um, how do we get from TRL 4 to 5, and then obviously after that it is number of experiments you can run times by, you know, dollars per experiments, you know, divided by how mu- how many dollars I have in the bank. And if you're doing something truly difficult, you need many experiments, and so you need lots of dollars, and it's always like 5x more dol- more dollars than you actually think you need, right?'Cause the, that's also the other thing is, is most people coming out of PhD programs have been writing grant applications, and they're trying to do things like down to the cent. Um, whereas when you're actually trying to get a product out there, you're going to hit so many dead ends. Um, and it's just random things, right? Like, um, uh, it turns out that, that, you know, a plate inside your electrolyzer works really well if it is tungsten, and turns out… But it turns out that it's gonna be, you know, a bajillion dollars to make 1,000 layers of it if it's tungsten. It has to be aluminum, um, if you actually wanna mass manufacture it, but you just spent, you know, the past four years doing all your science with tungsten, right? Like so- something like that, that you just don't think about those sorts of things when you're doing lab scale and then trying to s- trying to scale it up. And so therefore it turns out, "Oh, shit, we've actually gotta spend a million dollars on redoing a whole bunch of experiments we thought we'd passed already 'cause we've gotta rethink- Sure you know, something to do with aluminum." So it's always gonna take more money, um, than you expect. So it's basically like you think of how many experiments do I need to run to hit whatever the milestone is, and then probably multiply that number by like five. And that's actually the amount you, you need to be asking for. Maybe it's multiply it by 10 because you only get half of that money, right? So it's like if you think you need $1 million, you probably need to actually get five, which means you probably actually need to ask for 10, right? Yeah. That's the, like the, the- … the rough back of the envelope. That's the calculation. Yeah. Yeah. I wonder if in, in these situations that we're talking about, you know, the discovery phase, the, you know, TRL 1 to 3, where it's, it's often seen as a scientific discovery phase. We need to discover. There's no real commercialization pressure at the time. There's no kind of scale up manufacturing pressure, you know, you, you know, none of that. It's, you know, we're observing phenomena. Let's see if we can, you know, figure out how we might be able to manipulate this phenomena in such a way that it matters to someone in the outside world. That's kind of where we're trying to get to. Yeah. And I wonder if in the situations that you're describing that p- that people start kind of guiding that thinking towards a problem. They start directing it towards a problem earlier You know, in universities- Yeah … they're, especially in the UK and Europe, they're doing the science to do the science. They're- Yes … you know, and then I think it's only later in that research stage, in that discoveries stage where they start to think about, "Hmm, where might we use this?" I'm- Yep … I'm kind of roughly summarizing, but I think that's where a lot of it goes. I'm wondering if in these situations you're describing, people start, and we mentioned this at the top of the show, what's this problem that we could solve? How might we do it? Let's get some people around. Here are some experiments. How many experiments do you think we can do? And so on, and let's see how many experiments it takes us to get to this point. But we kind of know the direction we're heading in. Yeah. We know roughly where we wanna go. And I think with that, you can create that sense of urgency, pressure, the message, and so on. I think if you do it the more classic, almost academic way, it's much harder. Yep. Oh, 100%. I, I very much agree with it. If anything, I would say they're entirely two different disciplines, which is TRL1, which is basically fundamental science. Let's, you know, throw some chemicals in a, in a, in a boiler and see what happens, right? And this is, this is how, like GLP-1s come about, you know, is like someone in the 19, what is it, the 1990s goes and gets some lizard, uh, some lizard saliva and is like, "Oh, there's some weird stuff in here. Um, I'm just gonna put that on a shelf for, you know, 20 years. Maybe something will come of it." Right? Um, that sort of fundamental science is very different from then applying that technology. It's the same reason why in this GLP1, uh, 1 analysis, the people who got the, who got the lizard saliva are not the people who invented GLP-1s, right? Like- Yeah … they're very v- and turned that, and turned that into a product. Um, a- and, and going back to the, the Neuralink example as well, right, is we had a whole bunch of different paths to do, to do brain computer interfaces, which was effectively like this TRL1 thing. Which was, you know, are we gonna go through the vasculature? Are we gonna do the neural lace? You know, all of those things. Oh, there was also something called like neural dust. I forgot about that one. Um, but that's kind of like your TRL1. Like we are throwing shit against the wall and just like seeing what happens. So it's like, but really here we need to do this. But did they have the direction there? Sorry. Did they have the direction there? No, there was no direction. Okay. There was- There was no direction. There was no target. Exactly. There's no target. It's just, you know, ca- uh, so one person might be doing, might be doing neural dust, and I'm making this up obviously, might be doing a neural dust because it, it might potentially help epilepsy. And another person might be, might be doing it just because they want to talk to a neuron better, you know. And another person might be doing it for another reason, right? Um, so if anything, the, the value at that fundamental science level is that there is no target, right? And then it's almost a completely different discipline, and this is where I think we, we merge them when they shouldn't be merged, to take an overview of the world, of all of th- this fundamental science, and take a little piece from over there and a little piece from over there and apply them to some actual mission. Right? In the case of Neuralink, high bandwidth brain computer interfaces. In the case of, you know, Tesla- cheap electric cars, right? Um, and, and similarly, if you, if you look at Tesla, there it was the case that, well, you've got the cellphone and the laptop industry which is bringing the cost of batteries down very massively. You've got the solar industry who are making inverters, um, very, very cheap solid state inverters which had never, uh, never invented before. And it took someone not at that fundamental level, but a step up to go, "Oh, we're gonna take a piece from there, the cellphone industry, and we're gonna take a piece from there, the solar industry, and we're gonna put them together and we can make this third thing, uh, with a, with a, you know, completely different mission." I think we, we in sort of academia but then also in starting hard tech companies of, uh, think that if you're good at one, the fundamental science, you are going to also be good at TRL 2 which is let's find a mission and apply it. And I think they're actually fundamentally different skill sets That's really, really, really good. That's really good, and I think you're absolutely right. I'll take it. We had, uh, we had John Mankins, who is the co-author and, uh, uh, sorry, who's the co-inventor and author of the TRL framework on the show recently, and he was… I asked him, "What do you think the most difficult s- transition is between, you know, the different TRLs?" Expecting him to say kind of, "Oh, the valley of death," or, "This valley of death or that valley of death." Yeah. And he said exactly what you just said. He said between TRL1 and TRL2, they're completely different things. Yeah, yeah. He said it's so hard from observing basic phenomena to actually this is how we're gonna apply it. Yeah. This is where we're gonna try to make it matter. That he said it sounds quite simple. We tend to bundle them together in kind of this research phase, but there is such nuance, such distinction between those two things. We don't give it enough, we don't give it enough, uh, uh, attention or credit. Yeah, 100%, and by the way, I actually more think of many of these companies. The, the TRL framework is very useful, but it, it, it also makes you think these things are linear, where actually it's more like almost like a, a waterfall model when you're developing software, right? Yeah. Which is there's a whole bunch of TRL1 fundamental science that should be being done. Grants should be paying for that. Um, you know, America should frankly reverse a whole bunch of the issues. We can go down with them shutting this down. But basically, there should be a lot of TRL1 stuff happening, and then a whole bunch of different people doing this TRL through TR- TRL2 through five almost cycle, which is take a bunch of these TRL1 ideas, band them together into let's solve a problem with it, and try and take it as far up until you hit some fundamental wall, and then almost like iterate back. Like, ah, that wasn't quite, quite right. And, and the best startups do that cycle incredibly well and incredibly tightly, right? It's going two to three, and then two to three to four, and then two to three to four, four to five, and you, you hit a, a, a, you hit a, hit a wall on one piece of the thing we're doing, and maybe that, that work stream goes all the way back down and looks again at the TRL1 world to see if there's a different solution for that one particular work stream. Maybe it's battery chemistry for this thing that you're building, right? But other pieces of the system you're building are progressing very nicely up the TRL chain, and they aren't hitting walls, right? And it's that sort of iterative process that the best startups are very good at doing as each part of, like, the… Because ultimately, you're building systems, right? And each system has its own work streams and its own, and its own critical paths, but one of those critical paths might hit a wall, and you need to go all the way back to TRL1 on it. Yeah. Um, and so I just think of it as an iterative loop. Yeah. I've, I've, I've, I love the way you frame that. Um, the- Another question on the commercialization piece. Who are, you know, in the-- As an investor and, and perhaps as an operator, when you've, uh, seen companies or been involved with companies that are engaging at TRL4 and they start thinking about commercialization pathway at TRL4, they start line up b- you know, you know, they start to line up these discussions with, with big corporates or big industrial players in their domain Who are they typically contacting, uh, or engaging, uh, uh, in, in, in ongoing discussions? Who are they having to win over in order to get the kind of commitment that you're talking about? Because, I mean, those are pretty big commitments. Those are 30-year discussions. You mentioned the one earlier with, uh, you know, on plastic and Mercedes. Hey, you know, y- you know, the plastic market is gonna change. In 30 years' time it's gonna be very different, and you may be absolutely right there. Um, but who… You know, young companies, not much substance, not very well known, getting into some really big, very conservative corporates at, I'm guessing, a fairly high level in order to sign these kind of agreements. How is that happening? Yeah. This is, um, uh, I'll put it on the spectrum between art and science. This is, this is heavily over onto the art side. The art is figuring out who has a problem you are actually solving, right? So let's, let's go back to the, to the Twelve with the plastics and, and Mercedes example, right? And I know Nicholas has, has spoken about a bunch of this publicly, so, um, so I'm not sort of giving, giving up anything. But if you just said, "I have a machine that can take CO2 and turn it into plastic," who do I sell that to? Or how do I, how do I scale, scale that up? A lot of people would say, "Well, maybe I should go and talk to the plastic industry," right? Like the Covestros of the world, right? Um, and, uh, when we were looking at investing in Twelve, there were many people who, who had this sort of technology and said, "Well, we've partnered with some plastic manufacturer because we're like a drop-in replacement into their supply chain. You know, we're just gonna give them pellets, whatever, whatever it happens to be. But it turns out you aren't solving a problem for the plastic guys, you're a competitor to the plastic guys. Um, and this is like that, that scene in The Social Network where he's like, "You know what's cooler than a million dollars? A billion dollars," right? Um, instead you go to the person who ultimately actually has the problem, and the problem is Mercedes. Mercedes has the problem that in 20 years' time they're not going to be able to buy plastic. Now, Covestro, yes, you could argue they're not gonna be able to manufacture plastic, but it turns out all of these big industrials, they're massive conglomerates. They have, like, 17 million different, different revenue line items. Um, yes, they're making this type of, type of plastic, but they make a whole bunch of other things. So, like, it, it, they care less about that. But what the end consumer brand has, and you could have done the same with an example, say, selling to Patagonia or anything is they care about just their supply chain. Like, they cannot make the products that make them money without, without your thing, and B, they care about something else, and in this case was, um, we were in a massive upswell of people wanting to be clean and green. Sadly, that, that has diminished recently, right? But for these companies to be able to say w- uh, to their investors, like,"We're, we're protecting our supply chain," but B, to their consumers,"We are gonna be a clean and green bl- uh, brand, good, good for the planet." Yeah. You solve, like, a dual problem for them. So in this case, uh, and I know Nicholas actually went down this, this path, like, should I go and just try and sell to Covestro? Or should I instead go to an end-end consumer, like the, the end product, the Mercedes of the world, the Tides, you know, the PG&Es, all of those people, and instead kinda do pull through the supply chain? Because at the end of the day, it's kind of like follow the money. So if you go to the end and the money guy is saying, "I want this product," the rest of the supply chain is gonna stand up and say, you know, "How high do you want us to jump?" Yeah. Um, so that's the, uh, uh, the art, is figuring out who in your particular industry actually has the most, uh, the most pressing problem. And in this case, it was actually more of a marketing problem that many of these companies had around being able to say they were clean and green. And because you're selling something that is, you know, 20 years in the future, you can't just say, "Hey, I'm gonna deliver 100 tons of plastic," and that, that can't be the product. You're selling something else. You're selling greenness in this case, right? Yeah. And from what you've seen, who is best to have that discussion? You know, is it- Yeah … is it a founder CEO? Is it, let's say, a commercial leader who you've brought in who understands the supply chain quite well, who may be a bit of a domain expert? Um, h- you know, who is ideally suited to have that kind of conversation and to get some of that stuff over the line? So on the startup side, I think this is very much like a founder CEO should be having the actual conversation with whoever we've identified as the target, right? Yep. But you can still have your internal expert, being that supply chain expert who knows everything about the plastic industry in, in this example, who can explain that whole chain to your founder CEO and say, like, "These are the places where I think we have the most leverage to insert," right? And this is very much like, um, uh, enterprise sales and software, which people are pretty, pretty, uh, used to, where you've got, like, an internal champion, and then there's a decision-maker, and then there's the budget holder, right? You've got, you've got different people inside an organization you're selling to, but in this case it's across the supply chain. There's probably some way, like the Covestro is where you are going to insert your product into the supply chain, but they are not the people who want it. The person who wants it, you know, is gonna be down the line. It's, it's the Mercedes. Um, so I think the supply chain, your, uh, startup's internal supply chain guru who knows everything about this can map that for you, but ultimately the discussion of getting Mercedes to say, "Hey, we really want to put out a press release and be able to tell the world that we are, you know, clean and green in our plastic," that is a CEO founder on the startup side typically talking very high level to someone on… Well, depending on the problem, right? In this case it would be like head of sustainability, and then ultimately the CEO at, um, at Mercedes. But if they had gone in through, like, procurement or supply chain with,"Hey, we've got a better product. We've got a, we've got a, a different plastic," or in this case actually it's the same plastic, but, "By the way, it's the same plastic, but it costs 100 times more," right? You're not making it through procurement, right? Yeah. You've got to go and figure out the person you're solving the problem for, which in this case was, you know, the head of sustainability and, and then ultimately the board and the CEO. Yeah. And at this early stage, so it, you know, it's a, it's a common thing that we see. Companies approach us, they want to improve their commercialization effort. They often think about bringing in a chief commercial officer. Sometimes they think about bringing in a new chief executive officer, because we've got a founder who may not just be that experienced at telling that story. In, in these situations, would you, you know… Is having that kind of founder CEO capable of d- of, of, of taking that message to these sorts of customers and industrial partners? You know, Covestro would be the partner, Mercedes would be the client, or Nike or whoever else. Yeah. Um, is that almost a pre- r- a prerequisite? This one is very tough because I Some people are, are, have the capability to learn this. So let's, let's say you, you have a very good technical founder. Um, can they learn to be the, uh, you know, exciting big picture storyteller-type person? Some people can do it, but I don't wanna make a bla- a blanket statement that all people can do it. It is incredibly difficult. It is very intangible. Um, and some people, even if they've never done it, can pick it up, but other people, frankly, it's just a, it's a non-starter. They are never going to get there. Um, if you, uh, are in the unfortunate situation of having a person who is just never going to get there, it is very valuable obviously then to bring in someone who can. Um, and I would feel quite strongly as an investor at that point that that person actually has to be the CEO. What a lot of people try, try and do is, say, bring that person in alongside the existing CEO, um, or as a COO or something like that, right? They try to do it, like, either they're below or they're, or they're next to, and I think just tactically that ends up failing because you have for the r- for the, the proletariat inside the company, like the, the rank and file, you end up, uh, getting split messaging because your technical founder is talking, you know, the, the weeds, and this new big picture person is talking big picture and, and never the twain shall meet. Um, so my strong preference is if you are in the unfortunate position of having a, a technical founder who cannot learn this, that whoever you bring in who can do the big picture storytelling is, is actually put in as the CEO. Yeah. However, if there is even the slightest chance that your, that your technical founder can learn how to do this, absolutely go down that path, right? Um, and I don't think anyone's going to, to fault me for saying that if you see interviews with Elon from back in his, like, uh, PayPal days, not the most inspirational person, not the big picture change the world kind of guy, right? So very clearly it is possible to learn this stuff. Um, so if you have that as an option, absolutely take that option. If you're hiring the external CEO, are you looking for someone, you know, if you consider professional CEOs or people who are fairly familiar with, let's say, a particular value chain, who maybe come from that value chain The challenge is they're gonna be approaching it in a conservative, very value chain-like way. Yep. And that will be great for building trust, but you lose that boldness. Yep. You lose the boldness and the sense of adventure, and this is where we're going. So in that situation, would you advise getting another early st- very comfortable early stage, technically proficient, but bold entrepreneurial type? W- which is uncommon. You don't see that happen very often. Often, what happens is companies go for the example I just made. Uh, you know, the more conservative, uh, trust-building type of CEO- Yeah. Yeah, yeah … who's going to understand what good looks like in the industry, who's not gonna say dumb things, who's not gonna get a- ahead of him or herself, et cetera. Yeah. I, strong, I, I, I can see you're almost, like, leading the witness here, but I will say I, I 100% agree with, with where you're going with this, which is, uh, a number one mistake you, you have in this situation is people go and look at a new CEO candidate and goes, "Well, this person is an insider to the industry. They know everything about the industry. They must be really great." And it's like, no, they probably are not inspirational, and you're ultimately looking for inspiration. And I'll take this, take this, uh, back to when we were looking at the, uh, the clean concrete, um, sort of world, and this is, this is your supplementary cementitious materials and, and the concrete and, and kind of everything, uh, pozzolanic cements. Um, and out of, you know, call it the 15 startups that we looked at, um, I'm making this number up, but it was like 12 of them had picked that guy to be their scale-up CEO. You know, pick someone up out of an existing cement, uh, company or out of the existing cement supply chain who, you know, quote-unquote, "knew, knew everything and knew everybody." Um, and you know, pretty much, like, out of those 12 companies that did that, like, maybe only two of them are still around, and everybody… Because, like, you are just lacking the pizzazz, uh, as an industry insider. Um, I mean, if you can find an industry insider who's that excited and that, like, you know, bull in a china shop, they're probably not gonna be, uh, in that industry, right? Then they, they'll probably- Exactly… have quit a long, a long time ago. It's almost by definition a self-fulfilling prophecy. Um, so no, I, I will always take the doesn't know much about the industry but is, is, has the brains to figure it out, but has the balls, you know, for want of a better term- To kinda go for it to, to just go for it, right? To kinda go for it. Exactly. I think that's right. Yeah. I think that's exactly right. I've seen that happen- Yeah … so many times. I've been on-'Cause ultim- ultimately you're asking people to do something that on the surface will probably sound a little bit crazy, right? Yeah. Bringing, uh, going back to the, to the Twelve example, you know, you're asking Mercedes to, to buy something which at the time was 100x more expensive, right? It doesn't make rational sense. It has to be, uh, a little bit of a cheeky ask, right? And your industry insider is not gonna do the cheeky ask. No. No, no, no, no, no, no. I think that's, I think that's exactly right. Um- Greg, we're out of time. I can't believe it. This has flown, this has flown by. Time flies. I could have another hour. Yeah. I can't thank you enough. This has been, this has been great. Yeah, absolutely wonderful. Uh, this has been great. Thank you so much. Uh, God, I really felt we had a deep dive here, and, uh, I really appreciate all the examples. Holy cow, you gave so many wonderful examples here. So, uh, you know, this is wonderful. Yeah. Thank you, and, uh, I'm very excited to hear what's happened in the rest of this podcast, uh, series because, um, all of this sounds like very much needed information out there. The sort of dark art of building deep tech companies and getting through this TRL, you know, cycle. Yeah. Um, it is a dark art. There, there are not a lot of places that people can turn to find this stuff, so very excited to see what else you folks come up with. Oh, that's what we're trying to do little by little, step by step, and you've just contributed in a massive way to it today, so thank you. Thanks, Chris. You've been listening to the Lab to Market Leadership Podcast, brought to you by Deep Tech Leaders. This podcast has been produced by Beauxhaus. You can find out more about us on LinkedIn, Spotify, Apple, or wherever you get your podcasts.