The $8 Billion Question: Is Hadrian a Factory or a Story?

Hadrian raised $1.37 billion at a $7.87 billion valuation. The company's revenue is small relative to the valuation. The valuation is not connected to revenue.

The more interesting question is whether that matters. Most people look at a number like $8 billion for a company with revenue that is small relative to the valuation and conclude the market is crazy. But defense startups don't work like SaaS companies. They don't have a standard multiple. What they have is a bet about how many factories a country needs and whether one company can build them faster than anyone else.

Chris Power, Hadrian's founder, is Australian. He moved to the U.S. and spent six months touring aerospace and defense machine shops before founding the company in 2020. What he found was a supply chain run by operators averaging 63 years old, where the knowledge to make a critical part lived inside a single experienced machinist. His first attempt to fix this was ADSC, a private equity firm that rolled up precision manufacturers. That didn't work well enough. He concluded you can't retrofit automation onto a legacy shop the way you bolt software onto an old car. The right move was to build factories from scratch with the software already inside them.

That insight is worth thinking about. Because if it's right, it explains why Hadrian's product is called Opus — not a CNC machine or a robotic arm, but the software layer that ties them together. Opus converts hand-drawn blueprints into digital instructions, schedules machine time, runs quality inspections with lasers, and tracks tool wear automatically. Hadrian says it's five software companies and one robotics company stacked on top of each other. The machinery is off-the-shelf Hermle CNC machines and standard robotic arms. The moat, if there is one, is the software and the labeled data it generates.

Here's the thing about that data claim. Hadrian says it's the only company generating scalable, labeled data for AI in U.S. manufacturing, because legacy U.S. manufacturing has been offline for thirty years with no open-source equivalent. That's the kind of claim that's either a real moat or a press release. The ITAR (International Traffic in Arms Regulations) requirement means defense manufacturing data can't be shared openly, which does make labeled datasets scarce. But whether Hadrian's particular stack is irreplicable is the question, not the answer.

The machinists on Reddit had a simpler theory. They compared Hadrian to WeWork — taking something that already exists, repackaging it with better branding, and selling it to startups. The primary difference they identified was Hadrian's software for onboarding and shop-floor operations, though they weren't sure it was better than established systems like ProShop and Jobboss. You can see why they'd think that. From the outside, a factory with robots and CNC machines looks like a factory with robots and CNC machines. The software layer is invisible unless you're inside it.

Power is aware of this risk. It suggests he knows the gap between the story and the factory floor is where companies die.

What's happened so far is that the factory floor has expanded. Three facilities now. Hawthorne started as a small R&D lab, Torrance became the main production hub, and Mesa opened in late 2025. Power wants to build ten to twenty mega-factories.

But here's where the story gets complicated. Anduril — Palmer Luckey's high-valued defense tech company — just rolled the first Fury combat drone off its Arsenal-1 production line in Ohio in July 2026. Anduril isn't buying parts from Hadrian. It's building its own factory. Other primes are doing the same thing. SpaceX has vertical integration. Northrop Grumman brings production in-house. The companies Hadrian wants to serve are building the capability themselves.

Power's answer to this is direct. Power has argued that the country needs many Hadrians rather than more SpaceXs or Andurils, because the capital requirements are too high for that model. His thesis is that the factory-as-a-service model — where Hadrian operates the facility and the prime gets the output without the overhead — scales better than every company building its own Arsenal-1.

I suspect there's something right about that. Building a factory from scratch is hard even when you know what you're making. If a shake-out period for new facilities exists, it shows the friction in scaling. If Hadrian can compress that friction through Opus, there's a superlinear return to each new facility. The second factory learns from the first. The fifth learns from four. That compounding doesn't happen for the primes building one-off facilities for their own products.

The limitation is obvious. Hadrian has about $2 billion raised to date. Its revenue base is still tiny relative to the valuation. And the company needs to prove that factories opened in six months actually run at 75-80% uptime, not just on day one but over years of production. If Hadrian can't sustain its claims across three geographically dispersed facilities, the software-first narrative collapses into a capex play — expensive real estate and machinery that doesn't compound.

Also worth noting: the investor list reads like a political document. WCM Investment Management, Washington Harbour Partners, Valor Equity Partners, and a long list of other well-known investors. This is not a normal Series D. JPMorgan's involvement through its Security and Resiliency Initiative signals something closer to infrastructure financing than venture investing. The round is structured to look like the beginning of something that resembles a public utility rather than a startup.

What would change my mind? Two things. First, evidence that Opus is actually being used inside a prime's own facility through Hadrian's Factories-as-a-Service model, not just at Hadrian-owned locations. That would prove the software is portable and the moat is real. Second, a sustained demonstration that equipment uptime stays above 70% at the Mesa facility over a full production year, not a launch quarter. The difference between a 30% and 75% uptime rate is the difference between a margin business and a commodity business. It's the whole thesis in one number.

The test for investors is simple. Don't look at the valuation. Look at the Mesa factory in twelve months. If it's running at the utilization rates Hadrian claims, the software is the product and the factories are just the distribution mechanism. If not, you're funding expensive square footage at $8 billion per square foot.

I don't know which it will be yet. But that's the kind of question that matters.