Here's the thing most people get wrong about a safety company going public: they assume the safety layer is the low-margin commodity and the robot is the prize. The filings say the opposite.
Two robotics companies filed to go public in the same window through the same discredited vehicle—the de-SPAC—but only one of them looks like it could have survived a real IPO. And it's not the one making the robots.
Agility Robotics, the humanoid maker behind Digit, is merging with Churchill Capital Corp XI at a $2.5 billion pre-money valuation. FORT Robotics, which sells the safety stack that keeps machines like Digit from hurting people, is merging with Newbury Street II Acquisition Corp at $556.6 million.
FORT reportedly disclosed real numbers. Agility, so far, reportedly has not disclosed comparable unit economics.
Start with the tell. Both companies chose a route that carries a stigma from the 2021 boom, when Berkshire Grey, Sarcos, Symbotic, and Vicarious Surgical all de-SPAC'd. You take the SPAC shortcut when the traditional IPO process—with its demand for an auditable financial track record—is a door you can't walk through.
For Agility, that logic is coherent. It's reportedly capital-intensive, pre-profit, and building a category whose floor economics don't exist yet. Per The Robot Report, Michael Klein's Churchill XI brings roughly $420 million in trust and a Foxconn-led PIPE. This is a company buying itself runway to prove a thesis it can't yet demonstrate on a spreadsheet.
FORT is the strange one. Its 2025 revenue grew 62 percent. Operating expenses grew 19 percent. That gap—revenue compounding more than three times faster than cost—is the textbook signature of operating leverage, the thing public-market investors say they want and rarely get from a robotics company.
Gross margins were 66 percent in 2025, down from 70 percent in 2024 but still firmly in software territory. Revenue per employee was $276,000. The platform is deployed across more than 19,500 units globally, and pre-2025 customer cohorts contributed an estimated 68 percent of 2025 bookings—the profile of a sticky, low-churn base, not a company sprinting on new logos to paper over retention.
No single customer is more than 9 percent of revenue. Six-figure accounts have grown 3.8x since 2021. These are IPO-grade metrics. Most bankers would kill for this book.
So ask the obvious question: if FORT has the numbers a real IPO demands, why did it take the shortcut too?
The answer reframes the entire listing. FORT isn't a safety vendor that happens to be well-margined. It's a pick-and-shovel play on the risk of the whole sector—and it chose the SPAC to sell that story on its own terms, with private valuation negotiation and a controlled narrative, rather than exposing itself to a public book-build in a market where SpaceX, Anthropic, and OpenAI are vacuuming up every dollar of institutional attention.
Look at who FORT sells to. Its customer list names Agility, Zoox, Google DeepMind, Ocado, and DoorDash. Read that again.
FORT's customers are, arguably, the competitors in the sibling stories. The humanoid maker, the robotaxi player, the warehouse automation giant, the delivery fleets—FORT sits underneath all of them, agnostic to which one wins.
And Agility—the $2.5 billion humanoid company whose unit economics have, reportedly, not yet been disclosed—is FORT's own customer. The pick-and-shovel firm is arriving at public markets in better financial shape than the miner it equips.
The pick-and-shovel firm is arriving at public markets in better financial shape than the miner it equips.
That's the asymmetry the two listings expose. One company is asking you to buy the physical-AI thesis directly, at $2.5 billion, on faith. The other is asking you to hedge it—to own the safety layer that gets paid regardless of which robot company survives.
On paper, the hedge is the better trade. The load-bearing question is whether the hedge is actually load-bearing.
Here's where FORT's own pitch turns on itself. CEO Samuel Reeves describes the mission as building "a shared framework for trust that robot manufacturers, integrators, end users, regulators, insurers, governments, and any other interested party can rely on."
Notice the list. Regulators. Insurers. Governments. FORT is pitching a universal trust layer that these authorities can rely on—an appeal to institutions that have not, as of this filing, mandated anything.
The Trust Layer is certified to Safety Integrity Level 3 per IEC 61508. That's a real standard, backed by 25 patents, and FORT's CTO sits on the U.S. Technical Advisory Group to ISO TC 200. The engineering is not vaporware.
But certification is not a mandate. A voluntary standard that FORT's customers choose to meet is a very different asset from a regulatory floor that every robot deployment is legally required to clear.
The moat FORT is selling depends on the second thing existing. If regulators or insurers make a certified safety layer non-optional—a condition of operating autonomous machines around people—then FORT's 19,500 deployed units and sticky enterprise cohorts become the infrastructure everyone must pay to cross. That's a genuine moat.
If that mandate never lands, FORT is a well-margined vendor selling peace of mind to customers who buy it because it's good practice, not because they're forced to. Good business. Not a toll road.
And the pitch itself concedes the mandate doesn't exist yet. You don't build "a framework regulators can rely on" if regulators already require one—you'd point to the rule. The entire investment thesis is that the rule is coming, and that FORT is positioned to be the default when it does.
This is safety being financialized ahead of the regulation that would make it non-optional. Investors are being invited to buy safety-as-an-asset-class as a hedge on humanoid economics that don't yet exist—before any regulator has mandated the safety layer that hedge depends on.
The irony is that Agility agrees with FORT's premise, loudly. CEO Peggy Johnson: "You can't build your robot and then make it safe. That's a redesign." Safety embedded in design, not bolted on after the lab demo.
That's an endorsement of exactly the risk FORT is monetizing. Both companies are telling you safety is the gating constraint on scale. Only one of them has turned that belief into a 66-percent-margin business.
So the two US physical-AI listings resolve into a single, uncomfortable read. The humanoid maker at $2.5 billion needs the SPAC because it can't yet show you the unit economics. The safety layer at $556.6 million took the SPAC despite being able to—because its story is easier to sell privately than to defend in a public offering against the question at its core.
That question: is the trust layer a regulatory inevitability, or a premium product waiting on a rule that may never come?
FORT's financials are real. Its moat is a bet on regulators who haven't shown up. Buy the numbers if you want them. Just know you're also buying the assumption that the government eventually agrees with the pitch deck.