The a16z Show
00:11:37 · Connor Love, Erin Price-Wright
[00:11:37] S2: Yeah. I mean, I I think when you look at the physical world, applying software and AI will look different than what has happened in say legal or customer service or any other category where you see AI taking off in crazy way today. But to Erin's point, like many times what you have to do is you have to own the asset in order really to change that asset. Not every time, but a lot of times, you know, kind of showing is believing in the physical world. Mhmm. Whereas, you know, when you're when you're purely a digital AI product, you can kinda just see it, and it's very easily transferable and kind of understanding. What Mariana is doing with this copper mine is really impressive in a lot of ways. It's not just saying, hey, we're gonna be better at extracting copper. It's rethinking everything. [00:12:17] S1: It's It's rethinking entire end to end process of how you plan, operate, mine, and refine materials. And, you know, what's cool is once you once you own that process end to end and are using software at every stage of that journey, you actually capture that data, and then it feeds back Yeah.
00:12:35 · Connor Love
[00:12:35] S2: This is this is an important point. Mean, this is not just Mariana, but if you look at another business like Hadrian, I mean this is a perfect example of how one, you can utilize software, more importantly, it's not one silver bullet that all of a sudden makes a factory or a mine perfect. It's actually 15 or 20 different things, and when you look at what Chris has done at Hadrian, yes, it's kinda upfront what it takes to quote and even take an order of a part. It is the actual physical making of the part. It's how you inspect it and transfer it, and there's all these different interim steps that exist within a factory, and the beautiful thing when you kind of either start from a fresh slate like Chris has in building his factories, or in Mariana's case, when you kind of own the thing and take it over, you can look at every single node from a first principles perspective and say, hey, was this process that's been done the last hundred years the right process? And that's what's happening across every layer of the physical world right now. It's not just mines, it's not just factories. Mean, again, it exists in many different permutations. And and, you know, my personal belief, and we see this in a lot of our portfolio companies is, look, AI will be a phenomenal tool that can be used in the physical world. The reality is AI cannot turn a wrench yet. It will at some point, but we're far off that, again, can maybe have discussion about the robotic models and autonomy that we see in We're this getting closer, I just think the nuance here, and the portfolio companies, many that are on this list, or many non portfolio companies on this list, that they understand that the physical world is not about just putting a silver bullet forward, like I said, that some shiny AI product, It's deep understanding of what a process looks like, physically how to manipulate the process or how to change something that's happening in the world. It's just gonna look different based off.
00:14:26 · Erin Price-Wright
[00:14:26] S1: Well, well, maybe like a specific example or... To talk through is like like low mix high volume versus high mix low volume. Yeah. And how how AI can kind of change change the game there. Historically, like, you're spending a lot and setting up a line with a lot of automation, you really need to produce... Be producing, like, a lot of one single thing. Mhmm. And that's how you get the cost down, and that's how you justify the cost of investing in, you know, a pretty bespoke manufacturing line. In The US, we do a lot of low mix, high volume manufacturing, which is very hard to automate because it's very expensive when you have to change the line all the time. But what we're seeing across our portfolio is that manufacturing technicians themselves are... Have this... Have the intuition and now have the software tools to be able to to to change the economics of what it looks like to produce low volume, high mix on the line. So you can have a manufacturing technician. This is not someone with a college degree. This is not someone... Maybe they are, but not necessarily, who's studied software engineering, is now able to use these tools to reprogram and reconfigure their manufacturing lines on the fly using the fact that they understand the manufacturing process and they know how to do what they're doing in the physical world. They now suddenly have all these tools so they can... You know, you can retool a line instead of having to call a system integrator. You know? And it takes three months and hundreds of thousands of dollars to kind of reconfigure a line for a new setup. You can have a manufacturing engineer use Claude and do that in, you know, twenty minutes. And so that's been a real game changer across lots of companies that we've seen
00:16:02 · Connor Love, Erin Price-Wright
[00:16:02] S2: in And this why that's important is when you look back at how we've articulated AD 50 this year, it's kind of three categories. And a lot of times what we focus on, at least, you know, I have and and many people in the in in the sector are focused on, is just the systems companies. And look, there's a lot of really interesting systems companies. I mean, Anduril, Serrano, Castelion, mean, you know, the list goes on and on of who's building really good end systems, But the problem is, is if you can't iterate on the design and iterate on that prototype, you don't not only not get to an end system that is different, better, but you also just move at a slower pace. So then what happens is by the time you finally fix a production level, missile or plane or whatever, then all of a sudden, then you gotta go back and rejigger and build up all these subsystems in order to meet the scale. But if you have a more transferable line, a more adjustable line, number one, you can also build a prototype faster, but then two, you can then, you know, set up each one of these different components in a way where it actually can get to rate production quicker. [00:17:05] S1: Yeah. You're you're you're designing for you're designing for manufacture from the outset.
00:17:05 · Erin Price-Wright, Connor Love, Elena Burger
[00:17:05] S1: Yeah. You're you're you're designing for you're designing for manufacture from the outset. [00:17:09] S2: Exactly. And and that that is... Again, I I keep going back to warnings from, you know, Tesla and SpaceX. In a lot of ways, they they had to make those choice because they didn't have any other opportunity. But in the end, it's actually just that mantra of design for manufacturing that now exists within all these systems companies. And to some extent, like, we are now in the age of now seeing them produce. I mean, again, like, go look at what Anduril's built in, you know, in in Ohio and now on the East Coast in in in Baltimore. You know, go look at what, you know, Ceramic has built with the shipper. I mean, these are massive. Yeah. Huge facilities that are, at some point, you know, in a very soon soon timeline, gonna be producing thousands, tens of thousands, hopefully hundreds of thousands of systems. We haven't done that in America in a long time. [00:17:52] S1: Yeah. This is no longer an on paper theoretical exercise. This is happening in in in the real world, and and real stuff is coming off the lines. [00:18:01] S3: Something that I think is, like, special about AD Founders is they make any problem an Elon sized problem. And and I'd love to I'd love to hear sort of that applied to the different industries that we're talking about here.
00:18:18 · Connor Love
[00:18:18] S2: Yeah. I mean, I think when you look at defense specifically, this has actually been one of the most historically hard, you know, sectors to just build the thing that you think the government needs, because unfortunately, and again, I experienced this in my past when I served in the army, buyer is not the user and the appropriator of the dollars for the buyer is completely removed. You have all these disjointed decision makers or groups that matter. And so what happens is, companies, back then we didn't really have neo primes or AD companies to serve, but what you had is you had primes would just sit there and listen, tell me what you want, I'll develop the thing after you give me the money. And the age old problem is you get a system that's actually not technology forward, you get a system that was evolved for ten years ago, and then the end soldier gets something that actually is not super valuable to them. What's different now, back to the like, you can think through an Elon problem now is, you can say, okay, what does the military actually need? And the beautiful thing is, venture capital and what we do at American Dynamism, but also the rest of the capital markets is saying, we will actually fund that in advance, and you can build in advance of the government saying, here's the thing for you to build, and the outcome of that is the soldier gets a better product, gets something that actually fundamentally changes the battlefield and helps with deterrence or security or safety, and then more importantly, you get a timeframe of type of system that is actually built, aligned with the requirement of the market and what technology can do today. So it's a fundamentally different way of like instead of sitting and waiting for the government to tell you what they want, the government signals and you see it, and you build in advance by using private capital, and that just has not happened in the history of defense. So you can build, like Castelion, building a hypersonic missile that does a, not a super exquisite thing, but it's really hard to build a missile that goes that fast and is reliable. And the government now has given them contracts, and kind of showed up and bought missiles, but if you go look at the last three years of that business, I mean, was private venture capital.