A16z Sees Private Capital Fueling US Factory and Arms Rebuild

A16z认为私人资本正推动美国制造业与军工重建 国防需求充当了早期的资本支出补贴，吸收了首创生产成本，并为更广泛的商业机器人和制造业繁荣播下了零部件基础。 对话概述了多年的去工业化如何造成严重的国家安全风险，进而推动了涵盖原材料投入、制造工厂和终端系统的现代复兴。发言者探讨了初期国防需求如何作为资本支出补贴，惠及机器人等更广泛的军民两用商业应用，而新一代创始人正将SpaceX、特斯拉和Anduril的垂直整合模式应用于铜矿开采和输电等传统行业。他们强调了人工智能辅助工装正在改变多品种小批量工厂的经济模型，并认为美国去中心化、高流动性的私人资本市场相比自上而下的国家补贴具备不对称的地缘政治优势。最后，他们总结称，维持这一势头需要国家层面的监管意愿，并与实体基础设施所在的中美洲本土社区实现真正的经济利益协同。 私人资本在政府采购合同敲定之前，便通过资助实体基础设施、工装设备和新型工厂，从根本上改变着美国硬件制造业。 (17:54) 2026年上半年流入国防工业基地的风险资本已超过2025年全年的投资总额，这得益于风险投资基金、私募股权和华尔街机构的广泛参与。 (42:10) 国防需求为执行器、无刷电机和印刷电路板等军民两用零部件提供了早期资本支出补贴，从而降低了更广泛的商业机器人和消费电子产品的成本。 (6:51) 新的“制造中间层”正在通过与下一代主要承包商直接联合设计现代化的软件赋能子组件，逐步取代传统的交易型梯级供应商。 (33:50)

[0:51] Erin Price-Wright: You don't have a defense industrial base without an industrial base. [13:32] Connor Love: The reality is AI cannot turn a wrench yet. [5:09] Erin Price-Wright: I think our world's greatest gift in the United States is the discernment of the American consumer. We are going to push things to be more capable and cheaper than the government ever will.

[0:00] You don't have a defense industrial base [0:02] without an industrial base. It's been a [0:03] real wakeup call that we need to get [0:05] building again. [0:06] >> AI will be a phenomenal tool that can be [0:08] used in the physical world. The reality [0:10] is AI cannot turn [music] a wrench yet. [0:12] >> The top 10 most valuable companies in [0:14] the world, like the majority of them are [0:15] hardware companies. [music] What's it [0:17] like to visit one of these sites? You're [0:18] building a physical thing. You're not [0:20] sitting at a desk on your computer. [0:22] You're [music] out in the world. It [0:24] sounds kind of sexy and romantic, but [0:25] when you're actually doing it, it's [0:27] rough. When you show up at Mariana, when [0:29] you show up at Hadrien, we [music] show [0:30] up at Castillion, you literally get [0:32] people who are so passionate about what [0:34] they're doing, that produces really good [0:36] results. [0:36] >> There's [music] that very famous chart [0:38] of like US power production versus China [0:41] new energy production. What would it [0:44] take to [music] close that gap? [0:45] >> I don't think that that is a technology [0:47] problem. I think that that's a [0:51] Connor and Aaron, thank you so much for [0:54] joining the podcast today. We're here to [0:57] talk about 8050 and this year's list and [1:00] let's just get straight into things. Um, [1:03] I know a big theme of this year was the [1:07] American industrial base. Can you talk [1:09] about what we've learned this year and [1:11] why that's such a big important theme [1:13] right now? We've been saying for a [1:16] while, you don't have a defense [1:18] industrial base without an industrial [1:19] base. And I think this year really put [1:22] that into perspective. um you know with [1:24] conflicts going on in the world and and [1:26] realizing that um you know the hollowing [1:28] out of our industrial base that's [1:30] happened over the last several decades [1:31] actually has real consequences on the [1:33] battlefield for soldiers um both you [1:36] know American soldiers and our allies [1:38] and I think it's been a real wakeup call [1:40] in the United States um that uh we need [1:42] to get building again. [1:44] >> Yeah. I mean I think it's also something [1:45] that is not just one factor but it's [1:47] multiple factors that are all kind of [1:49] happening at the same time. And when you [1:50] take a step back and you say, okay, [1:52] world conflict, you have kind of a [1:54] continued evolution of the capital [1:56] markets, which I'm sure we'll talk about [1:57] a bit later about capital flowing into [1:58] this ecosystem. But I think one of the [2:00] bigger things is you actually have an [2:02] entrepreneurial base right now who is [2:04] effectively raising their hand and [2:06] saying, I'm going to build in this [2:07] ecosystem that's incredibly important, [2:08] not just for our nation, but also you [2:10] can build really valuable companies in. [2:12] So I think it's just this conglomeration [2:14] of a of geopolitical factors of capital [2:17] markets and then entrepreneurs that are [2:18] kind of all just coming together and you [2:20] know I think it's a really interesting [2:22] time to be building. It's a good time [2:23] for us to be investing. When we looked [2:25] at 8050 this year, we thought about uh [2:27] you know kind of where we're at in the [2:29] history of American dynamism and we're [2:31] still early but I think one theme that [2:33] has evolved is like you really need to [2:34] start looking at American dynamism [2:36] companies as a totality from end to end [2:39] from the input that comes out of the [2:41] ground to the factory that makes the [2:43] thing to the end system. And so that's [2:45] how we've set up 8050 this year to think [2:47] through those three lenses. Each one of [2:49] those pillars is uniquely important to [2:51] you know the end product or the end [2:53] system but in the end like I I I don't [2:56] think people think that way and a lot of [2:57] the companies that are addressing you [2:59] know uh uh things in our portfolio have [3:01] thought you know from a inputs [3:03] perspective from a facto's perspective [3:04] and now from a systems perspective. So [3:07] one theory that a person could develop [3:09] after looking at the list is okay we [3:12] built out this industrial base for you [3:15] know defense like building missiles and [3:17] drones. Um but it's actually a tailwind [3:20] now that's helping a lot of different [3:22] industries in America. So there is [3:24] energy, there's robotics, there's AI. [3:26] Can you kind of talk about how all of [3:28] these things fit together? Yeah, I mean [3:31] uh it's funny because if you look back [3:32] in history, I mean even you know 70 80 [3:35] almost 100 years ago if you look at how [3:38] we've developed technology in America a [3:40] lot of times it's very much been from [3:42] the government. If you look at how we [3:43] you know went to space for the first [3:44] time, if you look at how we developed [3:45] the internet, um how we developed [3:47] nuclear energy, it was the government [3:48] saying we're going to do this for the [3:51] purpose of the government and then the [3:52] long tale of that technology was super [3:54] valuable for, you know, you name it, the [3:56] internet being created, you know, energy [3:57] being proliferated, you know, across the [3:59] world. Um what's different now to some [4:02] extent is now private industry as well [4:04] as private capital was saying, you know, [4:06] we're actually going to be the impetus [4:08] that that develop something. And so to [4:10] your point um you know the many of these [4:13] markets is to serve the federal [4:15] government to serve the department of [4:16] war in some way. But the beauty about [4:18] these things is when you build you know [4:20] an industrial base either you know down [4:22] the supply chain or end systems itself [4:24] those inherently have many use cases. I [4:26] mean the the term dual use is talked [4:28] about a lot in in investing and you know [4:30] to some extent like uh I think it's a [4:33] little bit of a folly like focusing on [4:34] too many things at once is not good for [4:36] a startup but it is 100% true of what's [4:38] happening with any factory that's being [4:41] built today any neoactory being built [4:43] any any uh energy production anything we [4:45] we we touching kind of minerals or [4:47] materials in the ground it's not just [4:49] going into a defense system it's going [4:50] to go into my uncle's car it's going to [4:53] go into like my kid's future you know [4:55] toy that they have in some way, shape or [4:57] form. So I I I do think it is a uh you [4:59] know uh maybe the tip of the spear is [5:01] focus on the government but I think in [5:03] the long run to your point like what's [5:05] happening is these systems will be used [5:06] you know in the Whimo that we took today [5:08] or or anything else. [5:09] >> I think a good a good example of this to [5:11] to look at is actually DJI in China. [5:13] >> Yeah. [5:14] >> Um which start which has gone through [5:16] you know multiple arcs of of of its [5:19] lifetime and its journey. you know, [5:21] started out building consumer drones and [5:24] recognized over the late 2010s that [5:27] there was a large market for drones in [5:30] um both public safety and national [5:31] defense and so built out, you know, a [5:34] real highcale drone production business [5:37] for defense. And now the company behind [5:40] DJ or the the folks behind DJI are now [5:43] the largest producers of robots in the [5:45] world because what is very similar to a [5:47] drone motor is a robot actuator which is [5:49] the hardest part of building a robot. So [5:51] you have this company that over the [5:52] course of growing a large defense [5:55] business that itself spun out of a [5:56] consumer business is now you know the [5:59] world's leading expert on how to build [6:00] and produce hardware for kind of the [6:02] next wave of physical AI both consumer [6:04] and uh industrial. Um, and so I think, [6:08] you know, this technology transfer and [6:10] process knowledge transfer between [6:12] military use cases and non-military use [6:14] cases is really important and has always [6:16] been how it works. Um, and it's great to [6:18] have the federal government be sort of a [6:21] a a subsidizer for figuring out this [6:23] process knowledge and figuring out the [6:25] new technology. And then, you know, I [6:26] think I like our world's greatest gift [6:29] in the United States is the discernment [6:32] of the American consumer. like we are [6:34] going to push things to be more capable [6:36] and cheaper than the government ever [6:38] will. So that symbiotic relationship um [6:40] in the industrial base is actually [6:42] really important and we're certainly [6:43] seeing that on the energy side. [6:45] >> We're seeing that on the electronic [6:46] side. uh we're seeing that you know ac [6:49] across the entire sort of industrial [6:51] stack [6:51] >> and and and to some extent I I I I don't [6:53] think this is a super complex you know [6:56] topic but it's important to kind of [6:57] restate which is when you build the [6:59] first of a kind of anything it's quite [7:01] expensive like when you are building a [7:03] bespoke you know whether it's a military [7:05] system or a consumer product [7:06] >> zero to 100 it's going to cost a lot [7:08] >> it's going to cost a lot and and to some [7:10] extent like you you know we'll get into [7:12] this a bit a bit later I bet but you [7:13] know we talked about vertically [7:14] integrating or using the supply you know [7:16] chain base that you have. Um the [7:18] beautiful thing about what's happening [7:20] right now is we have a federal [7:21] government and a department of war is [7:23] saying, you know, number one, we're [7:24] going to we're going to try to actively [7:26] rebuild some of this this industrial [7:28] base. And by doing that, they're [7:29] actually changing that unit economic [7:31] decision from the start for these [7:33] companies. So, you know, again, it's [7:35] instead of building a 100 things, you [7:36] can say, okay, maybe build a hundred, [7:39] but we're willing to pay a cost that the [7:41] consumer or the commercial market [7:42] wouldn't pay. And all of a sudden then [7:44] you know that means more machines you [7:45] can put in your factory more capex that [7:47] you know the government is helping you [7:48] pay at the start and then that leads to [7:50] things that are physic you know both [7:52] like financially viable but also like [7:54] you know at a scale that actually can [7:55] touch the consumer. So I just it's [7:57] important to to to identify that and [7:58] that is definitely happening right now. [8:00] >> Yeah. Like like if you're building a new [8:01] PCB fab in the United States like that's [8:03] a pretty expensive operation. Those are [8:05] pretty expensive machinery pieces of [8:07] machinery. Um, and if you have a defense [8:10] buyer as the first buyer to help you [8:12] kind of pay for or subsidize a lot of [8:14] that capex investment, that's great [8:16] because you know what else has PCB [8:18] boards? Robots. You know what else has [8:19] PCB boards? [8:21] >> Hearing aids. Um, and so the ability to [8:23] to kind of pass on some of that capex to [8:26] the defense industrial base, I think is [8:28] uh will be really powerful as we as we [8:30] uh rebuild kind of industrial capacity [8:31] in the United States. So what [8:33] specifically are you seeing startups do [8:35] to drive down costs and move faster? [8:38] >> Yeah. Um I mean I I think uh like many [8:41] things we can look to uh Elon for a lot [8:43] of answers here. I mean it's not a crazy [8:45] statement to say really the few [8:47] modern-day companies that have built [8:49] anything at scale are probably SpaceX [8:51] and Tesla. Um and if you look at how [8:53] Elon and and his team have built those [8:55] businesses, it's been one very [8:57] vertically integrated. Uh and and look [8:59] like I I take the the ideology that you [9:01] shouldn't vertically integrate [9:02] everything. I mean again if you're [9:03] building a super exquisite system like a [9:06] missile that is you know doing something [9:07] very unique you kind of have to because [9:09] you have to you know control the [9:10] exquisitness you have to control the [9:12] cost. [9:13] >> Well well I mean SpaceX and Tesla were [9:14] doing first of their kind like there was [9:16] there didn't exist an industrial base in [9:18] the US to support them and so they had [9:19] to build it themselves. I think where [9:21] there are high quality suppliers like [9:23] you should use for sure suppliers. [9:25] >> Yeah. Yeah. I mean again I I I I think [9:27] the the the learning though is number [9:30] one if you don't have an industrial base [9:32] that you can rely on you know quickly [9:34] reliably you know etc then you have to [9:36] do it yourself and inherently that is a [9:38] hard journey like that that's really [9:39] hard to do this concept of saying okay [9:41] I'm going to focus on the end system I'm [9:43] going to focus on you know building the [9:45] most you know cost effective you know [9:48] effective system that I can and oh by [9:49] the way I got to go figure out the ball [9:51] bearing I got to go figure out the [9:52] solenoid valve I got like that that's a [9:54] lot to do and in case of you know Tesla [9:56] and SpaceX I would argue the the auto [9:58] industry was a little bit different [10:00] because there was a but there wasn't [10:01] there wasn't a battery supply chain they [10:03] had to you know Elon had to recreate [10:04] that [10:05] >> well the whole the whole power [10:06] electronics the the whole drivetrain you [10:08] know they had to really build completely [10:10] from scratch [10:11] >> uh based off papers that were published [10:12] in the 70s that you know had been [10:14] collecting dust for for 40 years. Um in [10:17] terms of uh companies that we talk about [10:20] in the 80 portfolio, there's Mariana [10:23] Minerals which bought a copper mine and [10:26] is kind of reinventing things end to [10:28] end. Would they be a good example to to [10:30] talk about here? I think that they they [10:33] recognized that um the you know the the [10:36] existing players in the mining industry [10:39] just were not going to pick pick up [10:40] technology fast enough uh for it to [10:43] really make a difference and really move [10:44] the needle that and they also recognized [10:46] that they wanted to actually own the [10:48] entire value chain to be able to really [10:50] put their money where their mouth was um [10:52] and capture the end value of the new [10:54] technology they were bringing. So yeah, [10:56] they do everything from you know [10:58] acquiring the asset, developing the [11:00] asset, mining the asset, operating the [11:02] asset and refining the material [11:04] themselves across multiple different uh [11:06] multiple different minerals and then [11:08] they sell a final commodity product in [11:09] the market. Um and have really leaned in [11:11] that's another trend I would point to is [11:13] how much founders and entrepreneurs [11:16] building in these spaces are leaning [11:17] into AI and software as a huge huge [11:20] component um of their build. Um [11:22] >> yeah I mean I I I think when you look at [11:24] the physical world applying software and [11:26] AI will look different than what has [11:28] happened in say legal or customer [11:30] service or or any other category where [11:32] you see AI taking off in a in a crazy [11:34] way today. But but to Erin's point like [11:37] many times what you have to do is you [11:39] have to own the asset in order really to [11:42] change that asset. Not every time, but a [11:44] lot of times, you know, kind of showing [11:45] is believing in the physical world. [11:47] Whereas, you know, when you when you're [11:48] purely a digital AI product, you can [11:50] kind of just see it and it's very easily [11:52] transferable and kind of understanding. [11:54] >> You know, what Marion is doing with this [11:55] copper mine is is really impressive in a [11:57] lot of ways. It's not just saying, "Hey, [11:59] we're going to be better at extracting [12:01] copper." It's it's rethinking [12:02] everything. It's [12:03] >> entire process of how you [12:05] >> plan, operate, mine, and refine [12:08] materials. And you know what's cool is [12:11] once you once you own that process end [12:13] to end and are using software at every [12:16] stage of that journey, you actually [12:17] capture that data and then it feeds [12:19] back. Um, [12:20] >> and this is this is an important point. [12:22] I mean this is not just Mariana, but if [12:23] you look at another business like Hrien, [12:25] I mean this is a perfect example of how [12:27] one you can utilize software, you know, [12:29] but but more importantly, um, it's not [12:32] one silver bullet that all of a sudden [12:34] makes a factory or a mine perfect. It's [12:37] actually 15 or 20 different things. And [12:39] when you look at what Chris has done at [12:40] Hrien, you know, yes, it's kind of [12:42] upfront what it takes to like quote and [12:44] even take an order of a part. It's, you [12:46] know, it is the actual physical, you [12:48] know, making of the part. It's how you, [12:50] you know, inspect it and transfer it and [12:51] and, you know, there's there's all these [12:53] different intram steps that exist within [12:55] a factory. And the beautiful thing when [12:57] you kind of either start from a you know [12:59] a fresh slate like Chris has in building [13:01] his factories or in Mariana's case when [13:03] you kind of own the thing and take it [13:05] over you can look at every single node [13:08] from a first principles p perspective [13:10] and say hey was this process that's been [13:12] done the last 100 years the right [13:14] process and that's what's happening [13:15] across every layer of the physical world [13:17] right now. It's not just mines it's not [13:19] just factories. I mean again it exists [13:21] in many different permutations and and [13:23] you know my personal belief and we see [13:25] this in a lot of our portfolio companies [13:27] is look AI will be a phenomenal tool [13:29] that can be used in the physical world. [13:32] The reality is AI cannot turn a wrench [13:34] yet. It will at some point but we're [13:36] we're far off that you know again we can [13:38] maybe have a discussion about the you [13:40] know robotic models and autonomy that [13:41] that that we see in the physical world. [13:43] We're getting we're getting closer. I [13:44] just think the the nuance here and and [13:46] the portfolio companies many that are on [13:48] this list or or many many non-portfolio [13:50] companies on this list that they [13:52] understand that um look the the physical [13:54] world is not about just putting a silver [13:56] bullet forward like I said that some you [13:58] know shiny AI product it's it's it's [14:00] deep understanding of what a process [14:03] looks like you know physically how to [14:05] like you know manipulate the the process [14:07] or how to you know you know change [14:09] something that's happening in the world [14:10] is just going to look different based [14:11] off [14:12] >> well maybe like a specific [14:14] example or to talk through is like um [14:19] like low mix high volume versus high mix [14:21] low volume and how how AI can kind of [14:24] change change the game there. So, you [14:26] know, historically like if you're [14:27] spending a lot and setting up a line [14:29] with a lot of automation, [14:31] >> you really need to produce be producing [14:33] like a lot of one single thing. Mhm. [14:35] >> Um and that's how you get the cost down [14:37] and that's how you justify the cost of [14:39] invest investing in you know a pretty [14:41] bespoke manufacturing line. Um in the US [14:44] we do a lot of lowmix high volume [14:46] manufacturing which is very hard to [14:48] automate because it's very expensive [14:49] when you have to change the line all the [14:51] time. U but what we're seeing across our [14:53] portfolio is that uh manufacturing [14:56] technicians themselves [14:58] are have this have the intuition and now [15:00] have the software tools to be able to to [15:02] to change the economics of what it looks [15:04] like to produce low volume high mix on [15:06] the line. So you can have you know a [15:08] manufacturing technician. This is not [15:10] someone with a college degree. This is [15:11] not someone maybe they are but like not [15:13] necessarily who's like studied software [15:15] engineering is now able to use these [15:18] tools to reprogram and reconfigure their [15:20] manufacturing lines on the fly using the [15:22] fact that they understand the [15:23] manufacturing process and they know how [15:25] to do what they're doing in the physical [15:26] world. they now suddenly have all these [15:28] tools. So they can, you know, you can [15:30] retool a line instead of having to call [15:32] a system system integrator, you know, [15:34] and it takes three months and hundreds [15:36] of thousands of dollars to kind of [15:37] reconfigure a line for a new setup. You [15:39] can have a manufacturing engineer use [15:40] claude and do that in, you know, 20 [15:42] minutes. Um, and so that's been a real [15:44] gamecher across lots of companies that [15:47] we've seen seen in this space. And and [15:49] why that's important is you know when [15:51] you look back at at how we've [15:52] articulated AD50 this year uh you know [15:55] it's it's you know kind of three [15:56] categories and a lot of times what we [15:58] focus on at least you know I have and [16:00] and many people in the in in the sector [16:02] are focused on is just the systems [16:03] companies and look there's a lot of [16:05] really interesting systems companies I [16:07] mean Andrew Sran Castillian I mean you [16:09] know the list goes on and on of who's [16:10] building really good end systems but the [16:13] problem is is if you can't you know uh [16:15] iterate on that on the design and [16:18] iterate on that prototype, you don't not [16:20] only not get to an end system that is [16:22] like, you know, different, better, but [16:24] you also just move at a slower pace. So [16:26] then what happens is by the time you [16:28] finally fix a production level, you [16:30] know, missile or plane or kind of [16:31] whatever, then all of a sudden then you [16:33] got to, you know, go back and rejigger [16:34] and build up all these kind of, you [16:36] know, subsystems in order to meet the [16:37] scale. But if you have, you know, a more [16:39] transferable line, a more adjustable [16:41] line, number one, you can also build the [16:43] prototype faster, but then two, you can [16:45] then, you know, set up each one of these [16:46] different components in a way where it [16:48] actually can get to rate production [16:50] quicker. [16:50] >> Yeah. You're you're designing for you're [16:52] designing for manufacturer from the [16:53] outset. [16:54] >> Exactly. Exactly. And that that is again [16:56] I I keep going back to learnings from, [16:58] you know, Tesla and SpaceX. In a lot of [17:00] ways, they they they had to make those [17:02] choice because they didn't have any [17:03] other opportunity. But in the end it's [17:05] actually just that mantra of design for [17:06] manufacturing that now exists within all [17:08] these systems companies. Um and to some [17:11] extent like we are now in the age of now [17:13] seeing them produce. I mean again like [17:15] go look at what Andrew's built in you [17:17] know in Ohio and uh and now on the east [17:20] coast in in Baltimore you know go look [17:22] at what uh you know Seronica's built [17:24] with the ship. I mean these are massive [17:26] huge facilities that are at some point [17:29] you know in a very soon soon timeline [17:31] going to be producing thousands tens of [17:33] thousands hopefully hundreds of [17:34] thousands of systems. We haven't done [17:36] that America in a long time. [17:38] >> Yeah this is no longer an on paper [17:40] theoretical exercise. Uh this is [17:42] happening in in in the real world and [17:44] and real stuff is coming off the lines. [17:46] something that I think is uh like [17:48] special about AD Founders is they make [17:51] any problem an Elonsized problem. Um and [17:56] and I'd love to I'd love to hear sort of [17:59] that applied to the different industries [18:01] that we're talking about here. Yeah, I [18:04] mean I think when you look at defense [18:06] specifically, uh this has actually been [18:08] one of the most historically hard, you [18:10] know, uh sectors to just build the thing [18:13] that you think the government needs [18:15] because unfortunately and again, you [18:16] know, I I I experienced this in my past, [18:19] you know, when I served in the army, you [18:20] know, the buyer is not the user and the [18:24] b and the appropriator of the dollars [18:25] for the buyer is completely removed. So [18:27] you have all these like disjointed [18:29] decision makers or groups that matter. [18:31] And so what happens is, you know, [18:33] companies, you know, back then we didn't [18:34] really have neopimes or ad companies to [18:36] serve, but what you had is you had [18:38] primes would just sit there and listen, [18:39] like tell me what you want. I'll develop [18:41] the thing after you give me the money. [18:42] And you know, the the age-old problem is [18:44] you get a system that's like actually [18:46] not, you know, technology forward. You [18:48] get a system that was developed for 10 [18:49] years ago and then, you know, the end [18:51] soldier gets something that actually is [18:52] not super valuable to them. What's [18:54] different now back to the like you know [18:56] you can think through an Elon problem [18:58] now is you can say okay what does the [19:01] military actually need and and the [19:02] beautiful thing is you know venture [19:04] capital and what we do at American [19:05] Dynamism but also you know the the rest [19:07] of the capital markets is saying we will [19:09] actually fund that in advance and you [19:11] can build in advance of the government [19:12] saying here's the thing for you to build [19:15] and and and the outcome of that is the [19:17] soldier gets a better product gets [19:19] something that actually fundamentally [19:20] changes the battlefield and you know [19:21] helps with deterrence or security or [19:23] safety [19:23] uh and then more importantly you get a [19:26] time frame of a of a type of system that [19:28] is actually built you know it you know [19:31] aligned with the requirement of the [19:32] market and what technology can do today. [19:35] So it's just it's it's a fundamentally [19:36] different way of like instead of sitting [19:38] and waiting for the government to tell [19:39] you what they want you you the [19:41] government signals and you see it and [19:43] you build an advance by using private [19:45] capital and that just has not happened [19:47] in the history of defense. So you can [19:49] build, you know, like Castellian [19:50] building a a hypersonic missile that [19:52] does a, you know, a not a super [19:55] exquisite thing, but it's really hard to [19:57] build a missile that goes that fast and [19:59] is reliable. And, you know, the [20:01] government, you know, now has given them [20:02] contracts and kind of, you know, showed [20:03] up and and and bought missiles. But if [20:05] you go look at the last three years of [20:07] that business, I mean, it was private [20:08] venture capital. It was, you know, [20:10] entrepreneurs like taking the risk and [20:12] building the thing in advance. I think [20:13] if you see that again and again in [20:15] defense, you're just going to get better [20:16] products and fundamentally like, you [20:18] know, you'll have the neopimes and the [20:20] new companies start to take a lot of [20:22] that market share. But it's not just the [20:23] defense thing. I mean, Aaron sees it on [20:25] energy as well. I mean, we're rewriting [20:28] our entire energy grid right now. And [20:30] we're again like Aaron would know better [20:31] than I would, but like we are in on the [20:33] one yard line. We are just starting [20:37] that process. I I think you know another [20:39] part of it is [clears throat] [20:41] is a little a little more intangible but [20:43] like these are really hard companies to [20:46] build. These are really really hard [20:48] problems to solve. Um you know when I [20:50] talk to Turner who's the CEO of Mariana [20:54] Minerals and you know he's flying all [20:56] he's flying to the middle of nowhere [20:58] places all over the world to build his [21:00] company. he's like out where there's no [21:02] cell phone signal, where there's no [21:03] electricity, you know, touching dirt, [21:06] you know, spending like weeks and months [21:09] away from his family. Um, like it has to [21:13] be worth it. You're building a physical [21:15] thing. You're not sitting at a desk on [21:17] your computer. You're out in the world. [21:20] It sounds kind of sexy and romantic, but [21:22] when you're actually doing it, like, [21:24] it's rough. The travel is really hard. [21:26] the people you're engaging with, you [21:28] know, you're not talking to, you're [21:30] calling up your suppliers. These are not [21:32] like other startups who operate in the [21:34] same kind of time scale and culture as [21:36] you. These are like really tough people [21:39] to negotiate with and work with who, you [21:41] know, have a very different culture. [21:43] >> And if you're going to be putting [21:45] yourself through that, like it better be [21:47] damn worth it. Um, and so I think people [21:52] people really gravitate towards these [21:53] problems that feel huge and enormous [21:56] because they're, you know, these are [21:58] real missionary founders that we see [21:59] building in this space even as even as [22:02] it's kind of heated up and there's more [22:03] capital and it's gotten cool. [22:05] >> Um, it's still brutal. I I I think the [22:08] biggest thing that has changed in [22:10] American dynamism land in the last few [22:12] years is we always knew that the [22:14] companies being built in this category [22:16] were really important to the fabric of [22:18] America. I mean again like when we talk [22:19] about defense systems, when we talk [22:21] about energy, I mean when we talk about [22:23] building, you know, manufacturing [22:24] widgets that go into everything in the [22:26] like that stuff's really important. Like [22:27] it affects our safety, it affects how we [22:29] live our life. But the biggest change is [22:32] those companies weren't just the most [22:33] important thing to us. Now they can be [22:35] the most valuable companies and that's [22:37] like a big distinction that that has [22:39] happened which is you know it's not just [22:41] like you know solving a problem for [22:43] safety or security it's solving a [22:45] problem for safety and security and oh [22:46] by the way those can be some of the [22:48] largest companies that exist in the [22:49] world and and that's what technology has [22:51] changed in these industries I mean [22:53] mining has always been a big big [22:55] industry defense has always been energy [22:57] >> it hasn't been economically worth it for [22:59] a young smart person you know who has a [23:02] lot of ambition to go after these [23:03] categories but I think [23:04] >> companiesactly like Palunteer, Anderl, [23:07] SpaceX have proved that you can build a [23:09] really big business operating in in [23:12] these categories. And if you I mean if [23:13] you look at the you know the top 10 most [23:15] valuable companies in the world like [23:17] that right now like the majority of them [23:18] are hardware companies. They're building [23:20] a physical product. um you know we've [23:22] kind of come full circle and so [23:24] >> you get the most ambitious founders of [23:26] our generation who you know who want to [23:28] run towards that hard problem and know [23:30] it's going to be worth it [23:31] >> and and and the beautiful thing about [23:32] many of these industries like mining or [23:34] you know Riotinto has been around for a [23:35] long time I mean Boeing's been around [23:37] for like I mean you know good companies [23:38] that are building good products but I [23:40] think when you take this you know first [23:42] of all like a first principles [23:43] technology approach about like what the [23:45] audacious thing that needs to be built [23:46] is and then two you get entrepreneurs [23:49] like these these folks that that are on [23:50] this 8050 [23:51] >> like yeah I'm going to rebuild the [23:53] American energy grid that sounds that [23:55] sounds fun [23:56] >> and you just that that that didn't exist [23:58] 50 years ago that didn't exist 5 years [24:00] ago I mean it is something that is [24:01] fundamentally you know our America's you [24:03] know best and brightest are building [24:05] towards this category for the first time [24:07] and then importantly like Erin said like [24:09] the opportunity is there you see these [24:11] examples of these companies that have [24:13] redefined space launch that have [24:14] redefined you know kind of military [24:16] systems have redefined you know all all [24:17] these these categories of American [24:19] dynamism that we care out. Um, and yeah, [24:21] I think it's exciting. I mean, uh, you [24:23] know, as as again, as someone who served [24:25] in the military, you always think of, [24:27] you know, like what you do as something [24:28] greater than yourself. I mean, look, [24:30] like the capital markets exists to, you [24:32] know, make money. Like it's we're we're [24:33] not a charity. This is a capitalist [24:35] society, but in the end, a lot of the [24:37] companies on this 8050 list are doing [24:39] something that is both going to be large [24:41] and valuable, but also going to be [24:42] really inherently important to the [24:44] fabric of America. And that's uh that's [24:46] pretty exciting. What's it what's it [24:48] like to to visit one of these sites like [24:50] a manufacturing site or a mining site or [24:53] you know drone manufacturers those kinds [24:56] of sites like [24:57] >> I mean Connor and I have the best job of [24:59] all time. I did have to I I did I do own [25:01] steeltoed boots now because the number [25:03] of the number of site visits I've been [25:04] to with portfolio companies where you [25:06] where they're like yeah you got to show [25:07] up in steel [laughter] toaded boots. [25:10] >> Um yeah [25:11] >> I mean uh uh it's not all sexy to be [25:14] honest. [25:15] These are these are locations that are, [25:16] you know, a lot of times hard to get to [25:18] in the middle of nowhere. And in the [25:21] end, like we're talking about physical [25:24] things that require, you know, earth [25:26] moving. Like I used to have uh in my [25:28] time in the military, I used to have [25:29] this this kind of senior officer who [25:31] would always tell me like the physical [25:33] world doesn't care about you. So just be [25:35] careful in the physical world. like, you [25:37] know, these are, you know, moving, you [25:38] know, tons of dirt and, you know, [25:40] massive, you know, these factories being [25:42] built out, you know, Hrien or or or what [25:44] AMA will do. Um, I mean, these things [25:46] are huge. Um, and and it's impressive. [25:49] You know, my one of my favorite things [25:50] is maybe you don't see this at the [25:52] remote sites. Maybe you do, but uh uh [25:54] it's this uh kind of running uh [25:57] competition between these companies who [25:59] can have the bigger American flag at [26:00] their site. I actually don't know who [26:02] currently holds that today. Uh, I mean, [26:04] Hadrien's was pretty big last time I [26:05] saw. [26:06] >> I think Valor is pretty big. [26:07] >> Valor's big, too. Okay, good. Yeah. [26:09] Yeah. Yeah. And like again, I mean, um, [26:12] that's that's awesome. Like the fact [26:14] that you have that type of like one [26:16] passion for for something greater than [26:18] just the the product or the the economic [26:20] opportunity that you're building. I [26:21] mean, again, it it drives purpose behind [26:24] a lot of these teams. when you show up [26:25] at, you know, Mariana, when you show up [26:27] at Hrien, when you show up at [26:29] Castellian, like you you literally get [26:30] people who are so passionate about what [26:32] they're doing and and you know, that [26:34] produces really good results. When you [26:36] have to do something really hard in the [26:37] physical world, if you have people that [26:39] are willing to run through walls, [26:40] sometimes literally to solve that [26:42] problem, you're just, you know, you get [26:44] really audacious, great outcomes. One of [26:46] my favorite things to do, you know, is [26:48] uh visit the same company sort of [26:50] multiple times within a 12-month period, [26:53] whether they're in the portfolio or not, [26:54] because you learn a lot by how fast [26:57] >> a factory buildout happens. [26:59] >> And so, you know, that that's one thing [27:01] we do is like we, you know, you don't [27:03] just go once, you go once, and then you [27:04] go three months later, you go three [27:05] months later, you go three months later. [27:07] And like I'm not I'm not a manufacturing [27:08] expert. I spent enough time on the [27:10] factory floor on a factory floor between [27:11] my time at Palunteer and now doing this [27:13] job. Um, you know, I like sort of know [27:16] what some of the machines are and kind [27:17] of like how a line roughly works, but [27:19] like there's just something about seeing [27:21] stuff accumulate in a factory and it [27:24] goes from the process of watching [27:25] something go from like empty warehouse [27:27] with epoxy floors and the big American [27:30] flag and maybe like a couple of pallets [27:31] and the founders just really excited to [27:33] show you their forklift that they [27:34] rented. [laughter] Um, and they're kind [27:36] of like, you know, they're they have the [27:38] basketball hoop set up or like the gym [27:39] set up in the corner to like, you know, [27:41] fast forward six months and it's just, [27:43] you know, they're freaking out about how [27:44] to expand. They're like, "Oh, we need to [27:46] get, you know, the the city out to to [27:48] give us more power because we have too [27:50] many machines plugged in. It's like kind [27:51] of chaos. There's technicians and [27:53] operators everywhere, people building [27:55] stuff." That's really satisfying and and [27:57] and really fun and something I think we [27:59] we're lucky to get to see that journey [28:01] and and and how quickly that can happen. [28:03] >> Yeah. Has anyone ever done like a time [28:05] lapse of uh [28:06] >> Oh, there's a lot. It's a whole That's a [28:08] whole like category of film making and [28:10] everyone kind of compares theirs and [28:12] it's like a, you know, a bit of a [28:14] >> I do I do like the gym thing. I remember [28:16] when I when I first visited many years [28:18] ago, Isaiah and the Gundo, uh he made me [28:21] do a rep on his uh on his [28:23] >> Yeah, they they all they all have like [28:25] they all they all have their bench press [28:26] set up, [28:27] >> which is cool. Is is there a like a [28:30] contagion effect to things being made in [28:32] America where it's like oh because the [28:34] actuator is made here now we can make [28:37] this and now we can make that and sort [28:40] of when you're able to compress one [28:42] timeline you can compress multiple [28:44] timelines. [28:45] >> I don't even think it's just timeline. I [28:46] also think it's cost and quality. I mean [28:49] it's interesting because when we look at [28:51] defense autonomous systems in particular [28:53] like looks making a lot of boats. [28:55] Covenant's making a lot of missiles. [28:57] Castillian's making a lot of missiles, [28:58] you know, Andre's making a lot of [29:00] things. Like when you build many [29:02] systems, you also require many [29:04] subsystems below that. And you know, [29:05] again, like there are really interesting [29:07] military systems being built today. The [29:09] the the correlary though, back to the [29:11] comment we had earlier, is like, you [29:13] know, by building towards defense, you [29:15] also build for a different industry. We [29:17] actually don't have a lot of humanoids [29:19] or robotics that exist in the world [29:20] today. I mean there are thousands maybe [29:23] I don't know the exact number but when [29:24] you think about the the subsystems that [29:26] go into some of our military systems [29:28] it's the exact same subsystems that are [29:30] going to go into the future robotic [29:31] systems it's the actuators like it's the [29:33] motors it's the you know you know valves [29:35] and switches and those type of things [29:37] and I think when you can build those to [29:38] quality and to cost then all of a sudden [29:41] you make other markets that maybe [29:43] weren't viable or real all of a sudden [29:46] makes sense and again I think that's [29:47] going to happen in the robotics world [29:49] which is you know We're going to spend [29:51] all this time on development of all [29:53] those subcomponents. We're going to [29:54] build for a lot of these military [29:56] systems, which is good. We need it for [29:57] security and safety and deterrence, but [29:59] then all of a sudden you're going to [30:00] have this, you know, kind of robotic [30:01] wave that will come in America. I said [30:03] it, you know, AI can't turn a wrench [30:05] yet. They can. They will eventually. Um, [30:07] and that will be enabled by all that, [30:10] you know, actuators and switches and [30:11] subsystems that we built. It's going to [30:13] serve a different market and and it's [30:15] going to make it, you know, financially [30:16] viable and we're not going to have to [30:18] rely on factories in China. And and [30:19] again I think that is that is net good [30:21] because you know as you build an [30:23] industrial base an industrial base can [30:25] be used for many things and you know [30:27] down to the component level you know I I [30:29] just think we don't have enough of it in [30:31] America and it it's a it's a decision [30:33] like you talk to a robotics company [30:34] today that wants to build you know some [30:36] robotic deployment there's a 99% chance [30:39] that majority of what they are buying is [30:41] coming from China and some [30:43] >> they for many reasons not just [30:45] geopolitical want to onshore that [30:49] purchasing for quality, cost, speed of [30:51] iteration um yeah and geopolitical [30:54] reasons. I think like one thing it you [30:57] know it's really over the last three [30:59] years even just it's remarkable how much [31:03] of a supplier ecosystem has developed [31:06] like [31:08] >> one of the companies in our portfolio um [31:10] you three years ago would order two [31:13] times the number of PCB boards PCBs that [31:16] they needed because they assumed that [31:18] they would have to throw 50% of the them [31:20] away for quality issues. They would get [31:22] the batch from their supplier. they [31:24] would do the testing and 50% of them [31:26] wouldn't pass a quality test. Um, you [31:28] know, then startups have emerged in this [31:30] category and are like, we're going to [31:31] fix this problem. We're going to use, [31:33] you know, software and modern technology [31:34] to, you know, redo the way PCBs are [31:37] manufactured in the US. We're going to [31:38] do it faster. We're going to do it [31:39] cheaper. And guess what? We're going to [31:41] do it better. Uh, and you know that now [31:43] that same company um buys you know PCBs [31:46] from new suppliers and you know they're [31:48] keeping 70 80 90% uh passing quality [31:51] test. So it's I think like we [31:54] underestimate how much the you know [31:57] close collocation and co-development of [32:00] some of this technology with um the end [32:02] customer and their suppliers really [32:03] matters to to drive quality higher. [32:05] >> So something really interesting is [32:07] happening. I I believe in the industrial [32:08] base. Uh like we talked about earlier, [32:10] you get this top town demand signal from [32:13] these new systems companies like we're [32:14] building new military systems but oh by [32:16] the way like you know Joby and Archer [32:19] we're building new planes too like we're [32:20] building new end systems that that is a [32:22] fact like cars are are being built [32:24] different planes are being built [32:25] military systems are being built [32:26] different when you build these new [32:28] systems inherently the inputs of those [32:30] systems need to change as well and when [32:32] you look at the industrial base in [32:34] America today like historically if you [32:36] go look at how Boeing makes a 737 you [32:38] know Boeing's a a company important to [32:40] the fabric of America, but they just [32:41] pick up the phone and they call [32:42] Honeywell and Honeywell helps them kind [32:44] of build all the different subcomponents [32:46] and it's not Honeywell's fault. But what [32:47] Honeywell does is they say, "I don't [32:49] want to build a new landing gear again. [32:51] We have this landing gear that meets the [32:52] spec and does the thing. So, let's just [32:54] keep going and doing that same thing [32:55] again and again and again." It's [32:57] rational. Like, you know, they're going [32:58] to build more 737s. They're going to get [33:00] paid the same way if it's a for 50-y old [33:02] landing gear or if it's a new landing [33:04] gear. The problem though is when you [33:06] keep iterating on these new requirements [33:08] for end systems, at some point the [33:11] 50-year-old subsystem just doesn't work. [33:14] And maybe it it it physically works in [33:16] in terms of meets the requirements, but [33:18] it's been designed and it's been [33:19] developed in a way that inherently uh [33:22] you know, we haven't innovated on it. [33:25] And and and why does that matter? It's [33:27] because you know, we've developed new [33:29] materials. like yes physics hasn't [33:31] changed but but the inputs to the [33:32] physics and the inputs to that design [33:34] has changed. So I think what's happening [33:36] or what with the opportunity for a lot [33:38] of these manufacturing companies is to [33:40] do what I call co-engineering and you [33:43] know we talked about you know the long [33:45] tale of these like subtier uh you know [33:47] uh suppliers they they they've been in [33:50] receive mode for the last 40 years. [33:52] Honeywell's picked up the phone and [33:53] they've said hey I need another landing [33:54] gear. Go make that landing gear. and [33:56] they've executed on the specs with that [33:58] IP and building it. When you want a new [34:00] requirement, uh that middle layer [34:03] cannot, you know, think on behalf of [34:05] Honeywell and change the requirement [34:08] because again they're not in the [34:09] position to do it. But I think what's [34:11] going to happen in the future is you're [34:13] going to have this middle layer of [34:14] supply companies that you know will [34:17] effectively be co-engineers with the N [34:20] OEM of rethinking not just what the end [34:23] system can do by way of speed or kind of [34:26] you know weight or mass but the [34:29] individual subcomponents will be [34:31] rethought will be re-engineered and will [34:33] be rebuilt and so you know the end [34:35] system will change from a top level the [34:37] the the the individual parts and [34:39] components will and and I think that [34:41] just hasn't happened in the last 40 or [34:43] 50 years. So, you're going to see a [34:45] completely new wave of what I believe [34:47] like will be called like the [34:49] manufacturing middle. Uh that will be, [34:51] you know, co-engineering and pushing uh [34:54] design and development upward in the [34:56] supply chain as opposed to just being in [34:58] receive mode and saying, "Okay, I can [35:00] build the same thing I built 40 years [35:02] from now." And I just I I think that's [35:04] an opportunity. I mean, I defined it, [35:06] you know, previously as like we're [35:07] creating the American manufacturing [35:09] asset class right now. I mean, it, you [35:11] know, private equity has been playing in [35:13] this ecosystem for a long time and [35:14] looking at an asset and buying it and [35:16] improving the cost and, you know, [35:18] selling it off. Again, that's not [35:19] fundamentally changing the things that [35:21] we build. Right now, companies will [35:23] change the the the different things that [35:25] we'll build. [35:26] >> A good example of this is kind of the um [35:28] on the on the grid. And when you think [35:30] about sort of like the electronic stack, [35:32] >> um, and we're about to go through a huge [35:35] step change in how we manage and [35:37] transfer power on the grid, moving from [35:39] a system of sort of static and [35:41] mechanical systems that were deployed in [35:44] large scale across the US in like the [35:45] 70s and 80s to now sort of modern power [35:48] electronics, battery storage [35:50] architectures which are much more [35:51] software native and flexible and you [35:55] know that is is just when you have a [35:56] such a complete step change in [35:59] technology where the technology that's [36:01] replacing the previous thing is just [36:03] completely different. you have this [36:05] opportunity to rethink the entire supply [36:07] chain behind that and the way all of [36:09] that design and engineering work happens [36:11] >> and and in the energy world is the same [36:12] thing that's happening in the defense or [36:14] the the commercial systems world which [36:15] is you have this massive amount of [36:17] demand for you know data centers and you [36:19] know new factories and new build when [36:21] you get this top down level you have the [36:23] ability to innovate like Aaron's talking [36:25] about and rethink the entire [36:26] architecture and look like [36:28] >> because it's a completely new [36:29] architecture so you know you can't rely [36:31] on these components that worked in the [36:33] previous worldly to just reconfigure [36:36] like you actually have the opportunity [36:37] to rethink that from the ground up. [36:39] >> And another important point here is like [36:41] you know you could be pessimistic and [36:42] you can say okay Connor all these new [36:45] military or end systems like not all of [36:47] them are going to be adopted so like [36:48] that that top down demand or or or or [36:51] Aaron like we're not actually going to [36:53] increase energy by four times like [36:55] there's not going to be that much [36:56] demand. I actually don't think it [36:58] matters because I think once you start [36:59] the snowball of the demand cycle moving [37:02] in one way even if you deliver half of [37:04] that you know uh requirement [37:06] >> you're not going back [37:06] >> you're not going back and it doesn't [37:08] make financial sense and again it's it's [37:10] like when you rethink something first [37:11] principles and you make the like new [37:14] design new engineering decision and a [37:16] lot of times that comes with better cost [37:17] better efficiency you're just going to [37:19] buy that thing now going [37:20] >> better serviceability easier to [37:22] manufacture [37:23] >> yeah yeah you're never going back [37:26] >> I'm curious Curious kind of what you're [37:27] seeing in terms of founders being second [37:31] generation Anderole or second generation [37:33] Palunteer or something like that and [37:36] maybe they they see some of these [37:37] learnings and they see the opportunity [37:39] to be the middle layer supplier in some [37:44] something that has never been touched [37:46] since you know decades ago and and kind [37:49] of like what what are you seeing in [37:51] terms of you know those opportunities [37:54] and those kinds of founders? Yeah, for [37:56] sure. I mean, if you you get a great way [37:58] to light a fire under an entrepreneur is [38:00] for them to be banging their head [38:02] against the wall on a sort of [38:04] engineering or supply chain problem [38:05] where they are and then they're, you [38:07] know, they're like, "Well, okay, I [38:08] really care about this. This is a a pain [38:10] in my ass. I know it's I know it's [38:12] bothering other people. I know it's a [38:13] problem that I don't just have here at [38:15] Ander or SpaceX or Tesla. It's a problem [38:17] more broadly. I'm going to go solve that [38:19] more generally." um and the kind of [38:21] lineage of founders that have sort of [38:22] lived and breathed this on the [38:24] manufacturing floor and not just [38:25] founders but even engineering teams who [38:27] are like I saw this problem I solved [38:28] this you know within Andre or Seronic or [38:31] whatever and I'm going to actually like [38:33] go out and build this as a company and [38:34] then I'm going to you know I know [38:36] everyone in the ecosystem so I'm I I [38:38] have you know automatic access to a [38:41] customer base that's captive um you know [38:44] we certainly see that trend happening [38:46] and you know this is how ecosystems get [38:48] built like people make fun of YC see for [38:50] all YC companies kind of selling to each [38:52] other. But like that's how it works [38:54] because you have, you know, it's it's [38:56] your friends. They're tolerant. They're [38:58] understanding. They're willing to like [39:00] work with you and help you and and you [39:01] know, put up with something breaking [39:03] because they're working on their own [39:05] really really hard problem. And if you [39:06] know the thing you're doing is not the [39:08] most existential thing to them, they're [39:09] like very willing to outsource it to [39:11] you. That's beautiful. That's how [39:12] Silicon Valley was built. And what's [39:14] really like what makes me really [39:16] optimistic is the extent to which we're [39:18] starting to see that in these kind of [39:19] physical world ecosystems and supply [39:21] chains. [39:21] >> Yeah. I I I think that the few companies [39:23] you mentioned, you know, the SpaceX's, [39:24] the Palunteers, um look, they weren't [39:27] overnight successes. Like those [39:28] companies took a long time to get to [39:31] both the product scale and just the the [39:33] size and success of the company that [39:34] they're at. [39:35] >> Overnight successes that took 20 years. [39:37] >> Exactly. I mean, and I think the the [39:39] learning point and you know, we invest [39:40] and a lot of the entrepreneurs on this [39:42] list like come come from those [39:43] ecosystems is you just learn this like [39:46] relentless focus to do the one thing as [39:49] well as you can, you know, down a path [39:52] until you've succeeded. And like, you [39:53] know, I've I've seen this with with the [39:55] Castellian guys, for example. It's like [39:57] they focused for two and a half, three [39:58] years on the first hypersonic missile [40:00] that was really hard to build. And, you [40:02] know, other people showed up, oh, be [40:03] multi-product, go build this small drone [40:05] here. and they're like, "No, we need to [40:06] do this one thing really well." And I [40:09] think, you know, that comes from the [40:10] fabric of SpaceX. Like at SpaceX, the [40:12] concept of being a responsible engineer [40:13] and like if this thing doesn't work, [40:16] like it's my ass. I'm going to make sure [40:18] like [laughter] I'm going to obsess over [40:19] this thing and again and again. Like [40:20] that's what it takes to build really [40:22] hard things. And again, like, you know, [40:24] there will be different, you know, [40:25] derivations of of success. Not everyone [40:27] will think just like the SpaceX mindset, [40:29] but if you look at a lot of these [40:30] businesses like that, you know, in the [40:32] physical world, you kind of need to do [40:33] that to build something really hard. [40:35] >> And but what what you do in doing that [40:37] is you learn, you know, you learn that [40:39] process knowledge. You learn how to do [40:41] it. And so think two, thing three, think [40:43] four [40:44] >> are a lot easier. You've developed a [40:46] supply chain. You have your [40:47] relationships. You've developed [40:48] manufacturing for design for [40:50] manufacturer processes. You know, you've [40:52] learned how to build a factory to a high [40:54] quality, high spec. you've learned how [40:55] to build that trust with customers and [40:57] so if you can do that one thing really [40:58] well [40:59] >> that sets you up to do other things [41:02] really well but it doesn't work unless [41:03] you deliver on that on that one core [41:06] thing [41:06] >> uh so we talk a lot about competition [41:09] China what does the United States do [41:12] uniquely when it comes to our capital [41:13] markets that that really enables these [41:16] startups [41:16] >> well I I I think this may arguably is [41:18] one of I mean Brian Shimp was talking [41:20] about it the founder of Android you know [41:22] a couple days ago which is this concept [41:23] that you know we of literally trillions [41:25] of dollars of private capital in our [41:27] capital markets that inherently follow [41:30] you know success follow both technology [41:32] success and and you know to some extent [41:34] like production success as well and I [41:36] think if you look at the United States [41:38] um number one like we do not have a top- [41:41] down directed you know uh industrial [41:43] policy that says the state will put this [41:45] amount of money in in this and I think [41:47] to some extent that's a good thing [41:48] because the money will follow the [41:50] entrepreneur the money will follow the [41:52] success I think that the the distin [41:54] distinction though we've kind of always [41:55] had a good capital markets in the United [41:57] States. I think the distinction is the [41:59] capital markets all of a sudden have [42:00] said you know what the category of ad is [42:02] really interesting and and you know [42:04] Aaron and I you know have been living [42:05] and breathing this for a long time but [42:07] if you go look you know not just up and [42:09] down you know Silicon Valley and San [42:10] Hill Road but if you go to you know the [42:12] major banks I mean JP Morgan has has [42:14] committed you know trillions of dollars [42:16] towards you know both investing in the [42:17] physical world not just defense but [42:19] factories and and other things in the [42:20] physical world and you see that again [42:22] and again and again across not just our [42:24] our venture capital firms but you know [42:26] private private equity now is is is is [42:28] starting to play in this category in a [42:29] big [42:30] >> big credit firms [42:31] >> big credit firms and and and that's so [42:33] important because you know the the the [42:35] reality is many of these companies in in [42:37] the AD world they require a lot of [42:39] capital when you build physical systems [42:41] when you build a Falcon 9 rocket or when [42:43] you build a CCA and you build a Blackird [42:46] missile um those are expensive [42:48] inherently to build up those lines and [42:49] build that physical system. So I I think [42:51] part of this is you know the capital [42:53] markets you know speak with you know [42:55] where the success happens and the [42:57] reality is is it's the capital markets [42:59] are not top down driven by our [43:01] government unlike China's and and what [43:03] that allows is that allows us to be far [43:05] more flexible uh you know again to [43:07] follow success like I said and then more [43:08] importantly um to fund things in advance [43:11] of it being obvious and and we talked [43:13] about that earlier the concept of [43:15] building in advance of the requirement [43:17] the need you know maybe we're a little [43:20] bit late on the energy side maybe a [43:21] little bit late on on some of the [43:22] defense systems. But look, the capital [43:24] markets have woken up and we're we're [43:26] putting hundreds of billions of dollars [43:27] if not trillions of dollars, you know, [43:28] you know, towards this category for the [43:30] next, you know, 10 plus years. [43:31] >> Yeah. I mean, if you look at just the [43:32] first half of 2026, more uh more risk [43:36] dollars went into the defense ecosystem [43:39] than all of last year combined, which [43:41] was many times what went into the [43:43] defense ecosystem in all of the previous [43:45] year combined. And I think historically [43:48] there's, you know, in defense [43:49] specifically, for example, there's been, [43:50] you know, a narrative around the valley [43:52] of death and like how do you get the [43:55] government contract to like propel you [43:57] through this period um of needing to [44:00] build out capex and and what's really [44:02] exciting is that almost that that that's [44:05] starting to disappear with the support [44:07] of of capital markets and and and of [44:09] risk and of risk capital. And that's [44:11] certainly extending even beyond the [44:13] defense ecosystem to energy and robotics [44:15] and critical minerals and some of these [44:17] other sectors that are really important [44:19] which is super super exciting to see. Um [44:22] yeah, [44:23] >> I had one question. And I'm not sure if [44:24] it's too much of a tangent though. Just [44:27] um there's that very famous chart of [44:29] like US power production versus like [44:33] China new new power new energy [44:35] production like America's line is like [44:38] basically flat and China's is you know [44:40] up and to the right like like I I guess [44:43] some question uh around like like what [44:46] would it take to close that gap or how [44:49] how fast can we build new power? Yeah, I [44:52] I don't think that is I don't think that [44:54] that is a um technology problem. I think [44:58] that that's a we need to do some [45:00] soularching about our will as an [45:02] American people and do we want [45:04] conforming projects to be built? Do we [45:06] want to grow? Um and if we do, I think [45:09] the the the technology and the capital [45:11] will follow. [45:12] >> Yeah. And it it's funny because you know [45:15] this is not just an energy problem. It's [45:17] you know it's it's an everything [45:18] problem. [45:18] >> It's an everything problem. and and and [45:20] look I you know we're we're optimists [45:23] inherently I do think that snowball has [45:25] started to roll. I think the question is [45:27] you know back to our discussion on [45:28] capital markets earlier uh will will [45:31] there be emergent winners in these [45:33] categories that then fuels more capital [45:35] to flow into the ecosystem and you know [45:37] it's this kind of self-fulfilling [45:38] prophecy that happens like you know yeah [45:40] if you build it you know that you know [45:42] they will come but at some point there [45:43] needs to be a pot at the end of the [45:44] rainbow and if there is for a few [45:46] companies I think you know one capital [45:48] will fall like I mentioned but but I [45:50] think the last thing I'll say just real [45:52] briefly is uh anything in the physical [45:55] world takes time. Um, you know, we we, [45:57] you know, to set up a factory, you know, [45:59] at a scale where really you either [46:01] affect our our power generation or you [46:03] affect our military systems or or you [46:05] name it. You can't take a couple of [46:08] engineers, you know, throw them in front [46:09] of a a couple servers and computer [46:11] screens and come up with an answer. I [46:12] mean, it's it's moving dirt. It is [46:14] physically understanding a supply chain. [46:16] Like, these things take years. And so, [46:18] you know, there's a little bit of [46:19] patience that we need to have in the [46:21] capital market side, but again, I I [46:23] think I'm optimi optimistic. I think [46:24] it's happening and you know to some [46:26] extent like it's ours to lose like it's [46:28] ours as a government and a nation to to [46:30] mess up and to get in the way of this [46:32] and that's a risk don't get me wrong but [46:34] >> I do think it's I do think it's another [46:36] uh long will be another long-term [46:38] benefit of capital markets even if it's [46:40] short-term painful just people need to [46:42] feel I mean you mentioned that these are [46:45] physical things they take time they're [46:47] also physical things in that they have a [46:48] physical manifestation to the people and [46:50] communities that live around them and [46:52] the most important thing is that you [46:55] know American society at large feels [46:57] like they're participating in the [46:58] economic growth that's coming as a [47:00] result of this development. Software you [47:03] know was sort of like had this ability [47:06] to be sort of siloed because it really [47:07] existed in just a few cities. It was [47:09] kind of this virtual thing. you sort of [47:11] experienced it in the way that your your [47:14] interaction with digital services got [47:16] better a lot better over the last 20 [47:18] years but it you know it it was a little [47:19] bit more it was a little bit more [47:21] abstract physical world like someone's [47:22] building a data center a power plant a [47:25] mine in your backyard um and you need a [47:29] you know you you have a say and you have [47:31] a stake in that you phys you physically [47:34] see it it affects your life and if you [47:35] don't feel like you have participation [47:37] in that economic upside it's hard I do [47:40] think that this is one thing that the [47:41] capital markets longterm will solve [47:42] really well because companies will [47:43] figure out how to bring communities [47:45] along for the journey and make it [47:46] financially and economically worth it [47:48] for them. [47:48] >> And I think there are some good examples [47:49] of that. A few companies on the AD50 [47:51] list who like, you know, build real [47:52] things in their communities and show up [47:54] the right way. There are also bad [47:55] examples that have existed in the last, [47:57] you know, few years. I mean, if you look [47:58] at what AMA does, like they show up to [48:00] these these really small machine shops [48:03] that exist. They partner with them, you [48:05] know, they don't they don't uh, you [48:06] know, cut the management and fire [48:08] people. They say, "How can we build new [48:10] products?" and grow the kind of you know [48:12] business that you have today. And it's [48:14] that type of uh you know concept of [48:17] working with communities as opposed to [48:18] working against them that I think we [48:20] still need to you know wrestle with a [48:21] little bit uh in in the tech world and [48:23] understand that just because we sit in [48:25] Silicon Valley I mean we we we don't [48:27] understand everything that's happening [48:28] in the world and you know fundamentally [48:30] I think a lot of these 80 companies will [48:33] build stuff that that touches middle [48:34] America. I mean it will physically exist [48:36] in middle America and you know I I I [48:39] think that's net good. I think that [48:40] means more jobs, more prosperity. I [48:42] think that means more energy and [48:43] abundance world where you know we have a [48:46] safer America and [48:47] >> less concentration of wealth in these [48:49] pockets in these elite pockets on the [48:51] coast. It's a good thing. [48:52] >> Yeah. [48:52] >> But it will take time. [48:54] >> Yeah. Well, Erin Connor, thank you so [48:56] much for joining. Uh and everyone can [48:58] check out the 8050 list today. [49:00] >> Awesome. Thanks. [49:01] >> Awesome. Thanks so much.