$CCXI / $AGLT Part 2: The Fence Around the Robot
I pulled the lobbying disclosures, the 118,000 word merger agreement, the patent ledger and the Taiwan supply chain on Agility Robotics.
I told you in Part 1 that the robot already clocks in. That was the easy part, the part you can watch on a video. This is the part nobody has done the work on.
For this one I did not read the press release. I read the 118,000 word merger agreement and its exhibits. I pulled every quarterly filing on the Senate lobbying database. I read the patent recordation ledger at the reel and frame level. I mapped the fleet backend out of public DNS records. I traced the camera supply chain to a listed company in Taiwan. I read the actual Section 232 tariff docket. Almost none of what follows is in the coverage, and a few of the loudest things being said about this deal are simply wrong. I will show you both.
Here is the thesis in one sentence. Agility Robotics is not just building a humanoid. It is quietly building a wall around the American market for humanoids, on three separate fronts at once, and it is timing that wall to its own public listing. That is the story. The robot is the thing standing behind it.
The three-front campaign nobody has connected
Start with the piece that took the longest to find, because it is spread across three federal agencies and no single filing spells it out.
Front one is a bill. The GUARD Act, introduced in Congress this year, extends the mechanism the government used to ban Huawei and ZTE, the FCC Covered List, to humanoid and quadruped robots. The bill text names no company. But the sponsors' letter that accompanies it names Unitree, the Chinese humanoid leader, by name, and calls for it to be run through the Covered List, the Defense Department's 1260H list, and the Commerce Entity List. Agility is a public endorser of that bill. Its CEO put her name to a supporting quote.
Front two is a standard. There is no finished safety standard in the world today for a humanoid robot working around people. The first one, ISO 25785-1, is being written right now, and the person listed as its Project Leader is an Agility Robotics engineer named Kevin Reese. When that standard publishes, certified compliance becomes the de facto license to put a robot next to a human worker, and it will have been drafted, in part, by the company that most needs it to exist. The same Kevin Reese is also the inventor on Agility's "cooperatively safe" fall-behavior patent. One person is building the moat in the patent office and in the standards body at the same time.
Front three is a tariff. On the second of September last year, Commerce opened a Section 232 national-security investigation into imported robotics and industrial machinery. Agility filed a comment in that docket, through its Chief Business Officer. A Section 232 investigation runs on a 270-day statutory clock, which puts the report to the President at roughly the end of May 2026. That is inside the window this deal is expected to close in. A tariff on imported, which in practice means Chinese, humanoids is a direct subsidy to a maker that assembles in Oregon.
Put the three together and the picture is unmistakable. A bill to ban the Chinese competitor, a standard to certify the domestic one, and a tariff to price the import out, all pushed by the same company, all landing in the same year it goes public. I have not seen anyone else lay these three next to each other.
The lobbying that front-ran the policy
The campaign above is not an accident, and you can prove that from the lobbying record.
Agility retains two Washington firms. The first, New Lantern, is the CEO's personal government-affairs shop. Its principal was Microsoft's chief lobbyist for sixteen years during her time as a Microsoft executive, then lobbied for her next company, and now lobbies for this one. The second firm, SIO Advocacy, registered to represent Agility in the same quarter the SPAC deal was signed, and its lobbyist is a fresh exit from Capitol Hill with an armed-services background. One firm works the executive branch. The other works Congress. That is a deliberate two-vector build.
Look at where the money points. Across its filings, Agility's lobbying targets are the Defense Logistics Agency, the Maritime Administration, the Defense Department, the Office of Science and Technology Policy, and the Office of the Vice President of the United States. The total spend is modest in dollars, about 110,000 across the filings, but the target list is not modest at all. A humanoid company lobbying the Defense Logistics Agency and the Maritime Administration is telling you it sees military warehousing and shipyard logistics as a market.
And here is the tell that it is working. Agility lobbied the Maritime Administration in the back half of last year. In February of this year, the White House released a Maritime Action Plan that explicitly prioritizes shipyard automation and automated material-handling systems. They lobbied the door roughly six months before the policy walked through it.
They rented the hardest part of the business
The single most common objection to any robotics company is the one about service. A robot is not software. It breaks, it needs parts, it needs a human to show up. How does a company of roughly four hundred people keep thousands of machines running across a continent? The honest answer for most robot startups is that they cannot, and it kills them.
Agility's answer is that it does not try. It rented the answer, and it rented it from names you already know.
Field service, the physically hardest part, goes to Ricoh. The company whose technician already shows up to fix the copier in your office is the company that services the Digit fleet. That is a global, boots-on-the-ground support network Agility did not have to build.
The software problem, getting a robot to talk to the system that runs the warehouse, goes to Manhattan Associates, the dominant warehouse-management-software vendor. Digit plugs into the software that most large warehouses already run.
And the selling and installing, the systems-integration layer, goes to two national integrators, Tompkins Solutions and Zion Solutions Group. Agility does not have to knock on every door itself.
Read those four names together and you realize Agility quietly built the distribution and support layer of a mature robotics company while everyone was staring at the robot. This is not in a single piece of the coverage.
The cap table is a supply chain wearing a suit
Every article lists the same investors. None of them explains what those investors are actually doing, which is the entire point. This is not a group of financial backers. It is a value chain, and every major name owns a different link.
NVIDIA sells the onboard compute, provides the simulation environment the robot is trained in, and holds equity. Schaeffler, the German motion-technology company, makes humanoid actuators for the whole market, holds equity, and buys Digits for a US plant. Amazon runs a pilot and holds equity. And on the manufacturing side there are two Taiwanese strategics, Foxconn, which leads the financing, and the parent of an optics maker I will come back to in a minute.
The reason this matters is that it de-risks the three things that kill hardware companies all at once. Supply is de-risked because your suppliers are your owners. Demand is de-risked because your customers are your owners. Compute is de-risked because the company that sells the chips is your owner. When Foxconn and one of the sharpest strategic rosters in the industry all show up on the same cap table, each with a different commercial reason to want it to work, that is a different animal from a financial SPAC.
Here is the full funding history, reconciled from the filings, which itself has never been laid out cleanly in one place.
The listed way to play it that nobody has spotted
Come back to that optics maker. This is the closest thing to a tradeable, non-obvious edge in the entire file.
Agility's June 2025 strategic round included NT$300 million from a Taiwanese group and from Canon. The group's listed arm is Ability Enterprise, which trades in Taipei under ticker 2374. Ability, working with Canon, co-develops the vision modules that go into Digit. A humanoid needs a lot of cameras. Ability's own management guides that humanoid and drone optics grow from a single-digit share of its sales today to roughly twenty percent by 2027, and it credited Agility's Nasdaq listing for its second-half upside.
So there is a public company, in a different market, whose revenue re-rates with Agility's success, and I have not seen a single piece of CCXI coverage mention it. Be precise about it, because precision is the whole game here. The NT$300 million check came from the private parent, Abico Group, not from the listed 2374 directly, and the twenty-percent figure is a single trade-press source behind a paywall. But the relationship is real and the read-through is real.
The government leg is real, and it is a zero in every model
Every model of this company that you will read carries defense revenue at zero. That is not conservatism, it is a blind spot, because there is already a federal contract.
The United States Naval Research Laboratory bought Digits. The award is in the federal spending database, 510,000 dollars, described as bipedal humanoid robots for interaction, from 2024. It is the company's only federal award to date, and there has been no follow-on, so do not oversell it. But pair it with the lobbying targets above, the Defense Logistics Agency and the Maritime Administration, and with the fact that the Defense Logistics Agency is publicly working on robots for military logistics, and you have the outline of a defense-logistics leg that is worth exactly zero in consensus and could be worth something real. Any dollar of it is upside.
The backend is real infrastructure, not a slide
One of the ways you separate a real robotics company from a demo is whether the fleet actually phones home, and how. I mapped Agility's cloud out of public DNS and certificate records, and the answer is that this is a genuine, modern production system.
Every robot authenticates to the cloud through an Auth0 machine-to-machine identity, the same pattern a serious IoT fleet uses. There is a separate European identity tenant, which is the concrete mechanism behind serving a customer like Schaeffler in Germany under European data rules. There is a live SOC 2 compliance program running on a Vanta trust center, which is what enterprise buyers demand before they let a vendor's software touch their operations. And there is an invite-only partner sandbox on a separate domain that mirrors the robot-telemetry stack, which means there is a real third-party integration program at the plumbing layer. None of this is marketing. It is in the DNS.
The IP is not what the marketing says, and that is fine
Here is where I correct the bull story, because the "decade of leg IP" line is doing more work in the pitch than it should.
The foundational patents, the spring-mass legged-locomotion work that the whole Digit lineage rests on, are not Agility's. They are Oregon State University's, they were funded by a DARPA contract, and Agility licenses them rather than owning them. Oregon State also holds equity in the company. I verified at the patent recordation ledger that there is no lien on those patents, which removes a real bear hook, there is no hidden secured creditor sitting ahead of shareholders on the crown-jewel IP. But I also pulled the citation graph, and here is the uncomfortable part. Boston Dynamics, Honda, Ford and Toyota cite those patents. None of the humanoid competitors, not Figure, not Tesla, not Apptronik, cite them. The famous leg IP is pedigree and brand. It is not an offensive moat that blocks anybody.
What Agility actually owns is a portfolio of in-house actuator and gearbox patents, plus a shipping whole-body-control model that is an unpatented trade secret. The moat is real, it is just in the drivetrain and the software, not in the university patents the deck leans on. And the one genuine open risk on the IP is the license itself. The royalty, the exclusivity, the reversion terms, and the size of Oregon State's equity stake are all undisclosed until the S-4 prints. A running royalty on Digit revenue, or a milestone reversion that a delayed product could trip, would be material. That is a real thing to watch.
The deal mechanics, read from the actual agreement
Now the structure, read line by line from the exhibit, not the summary.
The float is not retail's. Seven institutional funds, including a Millennium stake that has not been reported anywhere I can find, hold about forty-five percent of the public shares. These are passive holders, not activists, and at a share price well above the trust value none of them will redeem, so I am not going to tell you they control the vote. But the concentration is the fact. This is a tightly held float.
The deal itself is close to un-terminable. I searched the entire 118,000-word agreement and found no break-up fee, no earnout, and no sponsor forfeiture. There is an outside date and there are regulatory outs, but there is no ordinary path to walking away. The fairness opinion, unusually, was not signed by a bank. It was signed by an intellectual-property appraisal boutique, and it is fair only to the SPAC shareholders other than the sponsor. That tells you the value here is being appraised as intellectual property.
And there is a cash tell in the fine print. The agreement authorizes a pre-close bridge financing capped at 12.36 million Agility shares, roughly forty-four percent of Agility's own diluted share count, and it explicitly excludes the proceeds of that bridge from the 200 million dollar minimum-cash condition. A large internal raise that dilutes Agility's own holders, sitting inside a fixed valuation, carved out of the cash test, is not the sign of a company swimming in money.
The pattern trade, and its trapdoor
This is a Michael Klein SPAC, and Klein has a long record you can study. It is bimodal. Some of his deals popped and held, the AltC-Oklo deal and Infleqtion among them. Others broke badly, MultiPlan collapsed on a short report, Skillsoft bled into a reverse split, Lucid ran and then gave it all back over a longer horizon. So do not draw a single line through his history.
The mechanical risk here is not a mystery, it is a calendar. Every insider lockup, the sponsor's promote and the entire rolled-over Agility cap table, releases either when the stock trades above twelve dollars for a stretch or, failing that, at the hard 180-day open. The arbitrage desks that hold the float today are mechanical sellers into strength before the vote. The honest way to hold this is to respect that calendar, be long into the catalysts and cautious into the supply cliff, rather than pretend the overhang is not there.
The honest part, in full
I am not here to sell you a robot. Here is everything I found that cuts the other way, because a thesis is only as good as the case against it.
The flagship product has slipped about two years. The uncaged, cage-free Digit that carries the entire order book was pitched to ship in the fall of 2024. It is now guided to broad availability in 2027, and that slip is in the CEO's own words across two years of interviews. This is the cleanest bear tell in the file.
There is no audited revenue anywhere, because the S-4 has not been filed. Every revenue-multiple argument you have read, including the eye-watering ones, is built on estimates and on a misread of a CEO quote. When she said thirty dollars an hour, she was describing the human wage Digit competes against, not Digit's price. Digit's actual price, in her words, is several thousand dollars a month. The real number resolves only when the S-4 prints.
The order book is not a single anchor. The often-repeated line about a thousand-robot order from one customer is not in any filing. The filings say more than 300 million dollars in orders across more than thirty customers, converting as the delayed product reaches broad availability in 2027. And one of the loudest assumed anchors, Schaeffler, actually signed its big multi-thousand-unit European rollout with a different, competing robot company, not with Agility. Agility's Schaeffler footprint is a single US plant.
The factory is a nameplate. RoboFab is rated for ten thousand robots a year, but on the company's own factory tour it will make hundreds in its first year. The ten-thousand number is theoretical make-capacity, not sell-guidance.
The robot is more human-in-the-loop than the pitch admits. Agility is actively staffing a teleoperation and human-demonstration organization and building a data flywheel around it. Operations are autonomous once the robot is taught, in the CEO's framing, but the teaching and the supervision are real and ongoing, and the frontline operators are not even on the public job board.
And the people signals are mixed. The management team is a Magic Leap reunion at the top, unproven together on hardware commercialization. The board is founders and two early venture partners, with none of the strategic investors holding a seat. Glassdoor carries pointed reviews about management and a culture that soured through a 2024 reorganization. There is even an anonymous insider post from late last year, right at the original ship window, saying the company would be sold or go bankrupt. That is one person's opinion, but it is a dateable one, and I am not going to hide it. On the other side, the pre-IPO hiring is a genuine, accelerating surge, headcount at an all-time high with nearly half the open roles posted the week the deal was announced, so the company is clearly spending against a real ramp.
What actually decides this, and how I am playing it
Strip it all down and this is a company with a real, deployed product, a rented go-to-market that solves the hardest problem in the sector, a strategic consortium that de-risks supply and demand and compute at once, and a three-front regulatory campaign to wall out its Chinese competition, timed to its own listing. Against that sits a flagship that has slipped two years, no audited revenue until the S-4, a factory that makes hundreds not thousands, and a supply calendar that punishes the impatient.
The single event that resolves most of the open questions is the S-4. It will print the first audited revenue number, the identity of the customers behind that 300 million, the Oregon State license terms and equity stake, the real share count, and the post-close board. Everything I could not close in this research resolves there. Read it the day it drops.
Until then, the way I hold this is simple. The unpublished edges are the reason to own it, the honest bears are the reason to size it carefully, and the day-180 supply cliff is the reason to respect the calendar. I would rather be early and small on something this asymmetric than wait for the S-4 to make it obvious and pay up.
I told you this would be the deepest file on this name anywhere, and I meant it. Everything above is from public records, none of it is priced in, and most of it is not written anywhere else yet. When the S-4 makes it obvious, it will not be a secret anymore.
Disclosure. This is research synthesis built entirely from public filings, public databases, and public web sources, for educational purposes. It is not investment advice, and I am not a registered investment advisor. Positions and scenarios discussed are illustrative, not forecasts, and the inputs can be wrong. The company has not yet filed its S-4, so several figures here are drawn from secondary sources or are estimates and are labeled as such in the text. Do your own work.