Nominal is like one of these examples of you look at the ROI and you're like, okay, this isn't a nice to have. This is actually like saving us money. It's creating efficiencies in the business. If you are growing, then you're probably undervalued. I frankly fell in love with this problem and was exposed to it was when I was at Anduril in the early days. This was when Anduril was a much, much smaller company, like around 100 people. Wake up really early. We would go to Apple Valley, California, train like...
The trailer in the desert.
The trailer in the desert. You wake up super early before, you know, before the sun comes up. Pack a lot of these Anvil interceptor drones and the tower system, and you drive up, set up in a trailer. You'd get like Wi-Fi pucks. It was really hard to get. That group of humans at Anduril, I think, has gone on to start incredible companies, new products. Like it was just an awesome group. Really, really smart people.
I heard you barely even touched your $75 million Series B.
You know what's better than $75 million? $155 million. Move fast and break things. If you move fast and break things, you're also breaking people. People are going to get hurt. I would say it's not magic, it's sorcery.
Cameron, Trae, welcome to Sorcery.
Thank you for having us. Pumped to be here.
Well, it's been just 10 months since you were last on Sorcery for your $75 million Series B. Now you're back for your $80 million led by Founders Fund. What happened?
I always like to say, if you are growing, then you're probably undervalued. We were approached by the Founders Fund team, who's been on Nominal's cap table really from the beginning, over the sort of December holidays with a really, really attractive offer to lead a financing. And I was sort of thinking about it in terms of, I think you always want the sort of external realities to match kind of the internal momentum that we were feeling at Nominal over the course of 2025. We had really a breakout year, and so when the Founders Fund team kind of approached us with a really good proposition of leading a financing, it just made all the sense in the world. So yeah, we're pumped.
Might I add, your last financing happened in 10 days. This one was preempted. So there's some sort of magic at work here. What do you think that is?
I mean, I really think we sometimes joke internally at Nominal, but like I think we're building in a time where it's hard to find a tailwind in the industry that I think we are like not benefiting from. And so I think re-industrialization, physical AI, people getting back to building, you know, hardware in a serious way, all of those greatly benefit the business that Nominal is building. And I think really like, you know, in 2025 we saw sort of the hockey stick, you know, inflection point in the curve. A lot of that I think cracking commercial, you know, motion, going from a product to really finding product market fit and starting to scale that. And I think the, you know, good investors who are around the table can spot that. So...
And Trae... I would say it's not magic, it's sorcery.
Good, good, good, good. We love that. I'm going to clip that out.
Well, Trae, you've been tracking, and Founders Fund has been tracking Nominal for quite some time. You were involved in the seed round. So what exactly are you tracking throughout time to make this sort of preemptive decision?
Yeah, I mean, we're in kind of a unique position in that, you know, we invest in a lot of industrial companies. We don't have a thesis, but we just feel like this is an area that needs increased attention. But beyond that, we've incubated multiple companies that are kind of roughly in this sector that are customers of Nominal. So we had an inside view on what was going on, and I think that gave us the ability to say, wow, I mean, we're hearing a lot from companies that we started ourselves, from portfolio companies, about how important and necessary this software stack is, and that enabled us to lean in. Not only that, but, you know, Cam was actually at Anduril for a couple of years in the very early days, and so he and I have known each other for a while. He has a strong relationship with many of the partners on the team, and so we were just kind of in a position where we had an information flow and we knew that it was a time that we wanted to come in and make a high conviction, highly concentrated bet.
From an underwriting perspective, why was it worth taking the risk now versus waiting? And how do you kind of diligence this?
Yeah, I mean, the answer is anytime you can get in earlier at a lower price when something, to Cam's point, is growing, that's the right time to do it. That doesn't mean that there won't be future bets moving forward. We're still putting capital into some of the latest stage growth companies in Silicon Valley. We have a growth fund specifically intended to approach investments in that way. So we're just excited to be along for the ride at this point, and we'll continue to be a part of that ride moving forward as well.
You've recently tweeted that you're an investor/operator, and so with this specific... a slashy reference for those of you that aren't aware... with this particular company and your involvement in Anduril, can you take us behind the scenes on what it's like from the Anduril perspective of working with Nominal and what they've unlocked for you?
Yeah, you know, when I first started at Founders 12 years ago, over 12 years ago, I had no idea what I was doing. Arguably, I still don't, but I definitely didn't 12 years ago. And I think the only reason that I was useful in any way to Founders Fund or portfolio companies is that I had spent the prior six years at Palantir. And so I had this kind of depth of experience gathered mostly from making a ton of mistakes in the early days of Palantir. I think that especially in venture, this is probably less true of growth stage investing, but in venture investing, earlier stage, having operational exposure to startups in any given moment, you know, there's like epics that shift over time. The knowledge that I had from Palantir from 2008 to 2014 kind of becomes stale over time, but being in it at Anduril day-to-day gives you an incredible perspective of what tooling, what software is actually necessary to building a company in the modern day, which I believe gives you a tremendous edge when you're out doing diligence and finding companies, sourcing potential investments. So I think, you know, this is probably underappreciated. It's not a formal part of our strategy at Founders Fund, but it would be crazy to believe that having Peter as a co-founder at Palantir and Delian at Varda and Scott at General Matter and me at Anduril isn't a huge advantage in that specific way. Not saying we have every advantage, but in that specific way, we have a tremendous amount of operational exposure to building tech companies in 2026.
Before we go too far, Cameron, for people who might not be familiar with Nominal, can you explain what Nominal is?
Yeah, absolutely. And I'll actually hit with an Anduril origin story vignette too, because it's really relevant. So Nominal, we're building a data platform all around hardware testing and hardware validation. You can kind of think of us as like, you know, go back in time. If you for the last two decades have been building software and you weren't using GitHub, like you would have been crazy. And I think that that analogy is actually really apt to the moment in time we're in for software-defined hardware. This is a new industrial sector, autonomy, robotics. And I think I'll pull that thread a little bit more. You know, it used to be that VCSs, you know, version control systems for people building software, like those were just, they weren't like locally, like they were locally managed rather, they weren't centrally managed. So people were just building software and like, you know, it was kind of the wild, wild west. People started to realize like, oh, we should actually standardize this, but it was also internally managed. Then, you know, eventually that got productized and outsourced and venture dollars, you know, flew in. A lot of what we consider like the modern stack for testing and validating pure software was built on top of that, like billions in market cap created. Our whole thesis is that that exact same thing is happening for the software-defined hardware world in the new industrial world. It's going to rhyme. It's going to happen really fast because there's a lot of prior art and prior learnings that can be ported over, but there's substantial differences in how that sort of data platform and that stack needs to look. And that's exactly what Nominal was doing.
Where I frankly fell in love with this problem and was exposed to it was when I was at Anduril in the early days. So this was when Anduril was a much, much smaller company, like around 100 people. And the program that I was working on was the counter-UAS, the anti-drone system in its sort of infancy. And, you know, you always get excited to build something when you, I think, experience like sufficient pain. And just to get perspective about like what that looked like, it was like we would wake up really early. We would go to Apple Valley, California, train like...
The trailer in the desert.
The trailer in the desert. You wake up super early before, you know, before the sun comes up, pack a lot of these Anvil, people familiar with the Anvil interceptor drone and the tower system, and you drive up, set up in a trailer, you'd get like Wi-Fi pucks. It was really hard to get, you know, get proper connectivity to do what you needed to do. But like everything was, it was just like highly manual. And obviously like that group of humans at Anduril, I think has gone on to start incredible companies, new products. Like it was just an awesome group, really, really smart people. But like the software tools that the industry had available at that time were insufficient, right? And so naturally people started building things, tinkering, like ideating. And I just lived that. It was a lot of using MATLAB, it was a lot of using Excel to like track things. There's a lot of writing on whiteboards, like very manual workflows to sort of architect how you go through and really iteratively test and validate these products. And just started thinking like there's got to be a better way. And why is this, you know, what's happening? And that was then, you know, 2019, 2020 timeframe. So yeah, that is where Nominal is playing in the world, and we've seen a good bit of, I think, success expanding on that vision.
Yeah, I think like one of the threads to pull on this is that with so many products, whether it's, you know, software, enterprise B2B SaaS, or consumer software, or even consumer hardware, you know, there's like the old adage of move fast and break things. It's like, you know, just try it out, try it out, try it out in the real world, see what happens. The problem is, is that if you're flying Anvils or Roadrunners or something like that, if you move fast and break things, you're also breaking people. Like people are going to get hurt. It's going to be really messy. And then you're, you know, dealing with all the fallout of injuries at test ranges and stuff like that, which nobody wants to deal with. But more so, it's like once you get things productized and put out in the field, the important thing is that you are protecting lives. You're preventing the loss of, you know, warfighter lives. And so it is incredibly important in aerospace and defense in particular to ensure that your test and evaluation stack is preventing those edge cases from happening. And I think this is something that's like has largely been unappreciated in Silicon Valley as we've been focused on new and better ways to share 140 characters with your friends or optimize ads. We're moving into a different paradigm now where we're actually much more concerned about like how do we preserve, you know, our democratic and, you know, capitalist system for decades to come. And unfortunately, for better or for worse, that means that we have to get much better at doing real world things. We can't live in our software fantasy land on the internet anymore. There's a real world out there, and Nominal is an incredibly important part of the stack to ensuring that we can do that test and evaluation stuff as efficiently and elegantly as possible.
Last time we talked, we went over the TAM and how when you were working with Sequoia, you had no idea how big the TAM was, and they put some work behind that and they discovered it. Have you started to lay out as you go across categories and customers where you are going to see, I guess, most customer conversion and how big those markets are?
Molly, you're referencing a funny thing when we were raising the Series B after we had pitched Sequoia. I think Alfred Lin pulled me aside whatever and was like, really good. You're like way underestimating the TAM. And I just thought it was really funny. I kind of smiled. And I think he's right for what it's worth. But I think on the TAM piece, just because this comes up a lot where, you know, in Founders Fund, I'll make a Peter Thiel reference because I think it's like a really good one, is like, you know, starting with small markets, that sort of philosophy I think is super powerful. And I think it's like, you know, almost four years into the Nominal journey, I'm very grateful that we started with a like very simple problem that is like there's very acute pain and we could make it 10x better. And I think that is born out of this Anduril experience was particularly post-test analysis data review. This is a term of art. It's this like process, but basically every time you test validate a hardware system, you generate this like explosion of multimodal telemetry, sensor data, logs, video, audio, like all of this data. And that process just sucks. Like after you've been flying drones for 12 hours, you then have to figure out like, what just happened? Like, has the system been better? Did it improve? And that it was happening from 9:00 p.m. to 1:00 a.m., right? Like those like hours and improving it. So really like that is what Nominal started doing. Like we were just like, let's like build a product that makes that much, much better. And knowing that we were building in this massive industrial, you know, market that was one growing, I think, very quickly and having software like collide with it in innovative ways and now AI. And I think that has sort of, you know, proven very, very true.
In the earliest pitch deck slides of Nominal, we said we were going to start with hardware testing. I think to the point you just made, Trae, like that is where, like hardware testing is where software that is governing the performance of those hardware systems is first like meeting physics. It's like this beautiful like dance. It's like that's when you figure out like, is it going to work? Is it going to crash? Is it going to break? And so I actually think rather than simulation or design tools or other areas, like that's actually our belief is like that's the entry point to providing a ton of value to customers there. And then we're already in the process of expanding into more manufacturing, production, manufacturing testing, and into more like deployed fleet operations. And the market just grows. I think the global manufacturing TAM is like, I don't know, 200 billion or whatever McKinsey would say. It's big. So like I'm not worried about the addressable market at all.
You had some big customers at the last funding announcement. What does that scale to today?
Well, Anduril is one of those now. We publicly announced the partnership with Anduril, which has been great. Yeah, we are now working with four of the five largest traditional defense primes. That's how people refer to them these days. But which has been great. I think really something that's changed in the last 10 months, even since we spoke, I think is, you know, at some point we just needed the clock to like run sufficiently. Like there's, these are heavily regulated industries. There's a lot of bureaucracy. There is rightfully so like this is mission critical, the definition of mission critical capability. And so it's not a move fast and break things world. And I think we just needed the clock to run and to build past performance that we could demonstrate both with commercial customers and the government and just build trust and like word of mouth is a real thing. And people start, I think, talking and saying, oh, are you like, are you guys using Nominal? They're using Nominal. Everyone like, oh, everyone's using Nominal, right? So I think we're starting to sort of like feel that that's a big kind of inflection point that's happening. If you're not using Nominal, you're not testing properly.
Just one more time. You heard it here. Apart from Anduril, are there any other differentiated use cases that you could share? I mean, I just find this to be one of my favorite shows growing up was like How It's Made. And I feel like this is very much How It's Made because you're really direct on the hardware and you're learning from the systems and you're helping them and you're helping kind of, you're helping prevent things exploding and things breaking. And I'm just really curious, could you break down any more use?
Yeah, let me, I mean, let me give an example maybe of the federal testing landscape. I think that's actually a really interesting one. It's like, what is the status quo today for how massive, you know, government weapons and acquisition programs like perform testing? Because I think it's very, very eye-opening, right? And so I think the process is designed for exquisite systems. You'll hear that language a lot. But programs that take 10, 15, 20 years to develop, billions and billions of dollars in those contracts, the government works in what are called test matrices, but it's essentially a massive series of test points and conditions that these hardware platforms need to abide by. You know, the voltage must be between this and this, the angle of attack of this plane must perform this, the range, you know, all these things. When they go to then test those systems, everything is like sequential. There's really no concept of like parallelism and like multitasking. And the tools that are used, the software that's used is quite literally Microsoft Excel, MATLAB for performing analysis, maybe Python to glue things together, software like it's called IADS, it's like been around for a really long time performing like flight testing, but PDF and PowerPoint. So I always think it's like crazy. It's like this cognitive dissonance of like you look at an Anduril Fury or any just like new platform that companies are building, the end capability, and then you kind of like pan over to the engineer that's like performing the analysis and it's like you rewind 30 years and it's just like so obvious to me. This feels like an inevitability that has to change. And I think it's like what that forces is, I think this is an Elon type thing, but I think it's Parkinson's law that basically is sort of suggests like processes take as long as you like allow them to. And so I think that is like the definition of why we have so much delay in how we perform testing. Like every product, every weapon system is like, is massively delayed. I think unpacking that, Parkinson's law sort of states like the effort that has to go into testing something is like, you know, proportional to time. And I believe that. I don't think there is maliciousness here. I think it is like software-defined hardware systems today generate petabytes of data and 20 years ago they didn't. And so I think the tools have just not kept up and people are overwhelmed and not able to like, you know, keep pace. And that's where Nominal I think comes into play is like we can sort of flip the paradigm of effort in Parkinson's law and actually be like, no, it's actually a lot easier to do this. There's really powerful software you can query across all of your generated data for this like hardware test and convince yourself, you know, right then and there that you've like met all of the conditions in this matrix just by flying your Fury for, you know, 20 minutes autonomously and realizing that you actually put that platform in all of these vehicle states. Like that's the paradigm that we're, you know, trying to shift. And I think it starts with like really, really good software. But yeah, that's an example on the federal side.
Sorcery is brought to you by Brex, the financial stack trusted by more than 30,000 companies, including one in three venture-backed startups in the US. Nearly 40% of startups fail because they run out of cash. Brex is literally built to help founders avoid that. Unlike traditional banks that let your money sit idle, chipping away at it with fees, Brex is designed to help you spend smarter and move faster. Their all-in-one solution combines checking, treasury, and FDIC protection into one powerful account. You can send and receive money globally at lightning speeds. Get 20 times the standard FDIC coverage through their partner banks, and even high yield from day one with same-day and even same-hour liquidity. Access your funds anytime. Companies like Scale AI, DoorDash, Service Titan, Hims, Anthropic, Flexport, Robinhood, and Plaid trust and use Brex. Start today at brex.com/sorcery. That's brex.com/sorcery.
Turing is training the next generation of AI with tasks that require real expertise and real-world judgment. That's why companies like NVIDIA, Anthropic, Salesforce, and Gemini partner with Turing. Turing builds realistic reinforcement learning environments and data systems based on real operational traces. The kind of infrastructure frontier labs need to train superintelligence. Visit turing.com.
How much time are you saving?
Oh my God, a lot. I think like some fun stats are like some of these major test campaigns at some of the larger primes are like $3 million a day just for testing.
And often those test campaigns can be three, four, five, six months. We've been able to shave 50 to 60% off of those. So when you have those types of conversations with customers, you're saving the minutes that compound to hours that compound to days. There's huge, huge cost savings. And so it really makes the business discussion, I think, a lot easier.
Yeah, I think one of the things that is not talked about very often in the tech community is that enterprise SaaS balloons unbelievably as organizations grow. Where these companies, like you start off and it's like probably what as a consumer you're like, well, I have a Netflix subscription and I have a Paramount Plus subscription and I have a Hulu subscription. Before you know it, you're paying like $500 a month for your entertainment. Whereas like if you had just had cable 20 years ago, it all would have been done for 120 bucks or whatever. This is constantly happening inside of technology enterprises. And it usually takes years before the operations side of the business, you know, takes a step back and looks at their spend and goes, "This is crazy." Like the amount of money I'm spending on all of these different pieces of software has ballooned to an extent that I have no control over it. I haven't negotiated any of these contracts. And so at Anduril, my co-founder Matt Grimm, who's like I think maybe the world's greatest COO, he's very intense, but he's, you know, as dialed in as they come alongside our CIO, Tom Boswell. They have gone into this project of figuring out how to cut spend on SaaS because it's completely out of control, like it is at every other company at scale. You know, they go through and they identify stuff, they cut it, they consolidate, like they're going through this exercise. And usually what ends up happening is that, you know, a lot of nice-to-haves are sacrificed on the altar of efficiency. It's like, okay, yep, that was cool. It was a nice experiment. Too bad you're out of luck. We're not using that. Everyone shut up and go back to work. Nominal is like one of these examples of you look at the ROI and you're like, okay, this isn't a nice-to-have. This is actually like saving us money. It's creating efficiencies in the business. This is one of those things that you back up the truck to and really push to integrate it more deeply into the business. And I can honestly say like there's just not that many vendors that fit that level of commitment. Most things are, you know, begrudgingly investments like legacy software that you literally can't get away from using because you're in the aerospace and defense sector or, you know, ERPs or something like that that everyone's unhappy. Except the legacy software vendor that realizes that they have you over a barrel. But this is one of those unique cases where everyone feels super aligned, very happy, and we feel like we're getting the return on investment that we want out of that spend.
And I think just building on that point too, I think, you know, Anduril is this amazing case of, like, you know, we often in Nominal, we're like competing with sort of internally built solutions or maintained tools. And I think really like satisfying and fulfilling for us, I think, to sort of like know that what we had built, we'll put it this way, I was very familiar, I think, with like the level of capability that existed just natively inside of Anduril having worked there before. And so waited until I thought, you know, Nominal had sort of hit escape velocity on the product offering and the capability to sort of bring it back and have it met with that type of reception that Trae just described. So I think like, you know, this is an area where people will, companies as they grow, will try to assemble sort of their Frankenstein version of an internal stack. But at some point when you get very serious about scale and operations, you have 30 different hardware programs and product lines, it makes sense to buy a productized solution and back the truck up around it.
And now you're saving Anduril billions of dollars.
I mean, that would be great because if he was saving us billions of dollars, we would be profitable, which we are not.
We're doing our small part.
Yeah. You're doing your part to helping us eventually get to profitability. Yes.
What a beautiful partnership.
I'm so curious from both of your perspectives on this as AI rips through markets, public markets, hundreds of billions of dollars are being decimated. Do you think the hype, the effect and this kind of destruction or disruption from OpenClaw and Clawed and all these different types of new AI solutions that are causing autonomous agents and creating things instantaneously, is this going to affect you?
I think it will affect everything. But let me actually go back to the, I think the GitHub example is a good one. It's like when you whip out your cloud code and you're, you know, vibe coding or whatever you're building, you're still in almost all instances, you're still referencing GitHub. Like I mean, it's GitHub, right? It's like interfacing and that core infrastructure functionality like does not exist in this software for hardware space, right? Like that is what Nominal is building. And so I think where we sort of see is like we talk a lot about how we've like taken this like industrial software and software for hardware world from, I don't know, 2003 to 2021, like that's the crawl phase of Nominal. And so much of that is just like...
Good software engineering is like organizing data, associating metadata, tagging it, storing it thoughtfully — you know, like all of those things that enable the work that engineers can do using Nominal. And now we think we have, you know, a pretty sufficient moat to sort of build the next phase. So we're incorporating, you know, AI into our products. We already have beta features out there where if you're doing this post-test analysis, you can, you know, just use a chat interface to start to interact with your test data, which blows engineers' minds. But you can say something like, 'Plot the kinematics of the drone,' right? You don't have to manually do that, right? And it built on the building blocks of Nominal, what already exists. It will start to just do that type of function.
And so really, like the paradigm shift, I think, is moving from like this exquisite hardware system where, you know, 50 engineers had to be involved for like one piece of hardware. I think right now we're sort of in this zone of like one-to-one, you know, man-on-man. And I think Nominal will help enable a world where one human engineer can actually start to affect and test and validate and interact with, you know, 50 pieces of hardware through like, you know, agentic workflows, a lot of parallelization, distribution like that. I think that is where the world is going. I do not worry about the sort of de-Nominalization of Nominal. I think the software-for-hardware world is ripe for a lot of really, really good innovation.
So what's your long-term defensibility?
I mean, I think being that sort of like core piece of infrastructure is that. I think continue to build closer and closer to the hardware as well, like getting more into the embedded systems layer and stuff like that. And then ultimately, like, I mean, we kind of have a pretty central north star of like we will do whatever it takes. That sounds kind of ominous, but we'll do whatever it takes, I think, to satisfy the north star for our customers, which is like helping them build and validate hardware better, faster, and sort of cheaper. And I think having that flexible mission is really, really powerful as we adapt and react to how technology is changing.
Whatever it takes. Write that down.
Yeah. Whatever. It's important. It's really important.
Yeah. His customers are going to appreciate that at some point down the line.
Yeah. No, I think to your point about the industrials, is that the sector is different than like software companies building software where, you know, you can pretty easily in many cases — not in every case — you can kind of rip something out and replace it with an updated thing where you're just kind of integrating the data again via an API.
This is one of those industries where there's stickiness. Once it's in and it becomes the trusted kind of source system for the data, not only are the partners and vendors going to also be plugged in and trust it, but the government, which is oftentimes the customer of these commercial organizations, they will look to that as, you know, a trusted system that they can rely on that data being, you know, represented in a similar way that they can track over time. So these things don't end up getting pulled out, and you can look at other companies that have demonstrated success in this space. Palantir, most notably. Palantir doesn't have customer churn — like once Palantir is in, it's sort of in. There's not a whole lot that the customer would want to do to change that. And I think that is like a moat and defensibility not only for Nominal but for like really anyone that's working in this space. The downside, of course, is that these are really difficult customers. They're difficult customers to win, and they're difficult customers to service. And it requires a bit more of a, you know, a bit more courage than going out and, you know, being a YC company selling software to other YC companies. That's like a very different model.
Yeah. I think on that point, I mean, just to build on that, I think like we — this is a fun one because I think I know Trae categorically is very skeptical of dual-use businesses. Is that a fair statement? Yeah. I think like Nominal is going to be the dual-use business that gets Trae to be like, 'Okay, fine.' [laughter] That's my goal. That's my goal. No, but I think on that point, like I think it's been really — we've done a really good job, I think, of engaging directly with, you know, federal customers, which are ultimately the ones setting standards, setting regulations, like interpreting that test data as the final sort of outcome, and sort of like working with them, you know, building alongside things like, you know, deploying the product in the different environments where it needs to be, doing like high-side deployments, getting credibility, having people be like, 'Oh, actually, like I can look over the shoulder of vendors using Nominal and be like, "Oh, that's like what industry is using the best."' Like we've worked really hard, I think, to connect those two audiences because I think to your point, when that demand signal is like coming all the way through the ecosystem, it's just extremely powerful. And I think we're really feeling that pull now. Like we've already had some sort of organic hops where people are like, 'Oh, this other program, you know, this vendor participating in this other program is using Nominal. Like we should probably use Nominal and get on board.' And I think we just start to move the market. So it's been a case where I think the strategic nature of being dual-use is working so far.
So what is your sales strategy?
Yes. Software. [laughter] Yeah. Really do whatever it takes.
