Unscarcity
Sign in for free: Preamble (PDF, ebook & audiobook) + Forum access + Direct purchases Sign In

Unscarcity Research

Wage-Anchored Pricing: Why Cheaper Robots Don't Mean Cheaper Goods

Agility's S-4 rents Digit for $8,500 a month, priced against human labor cost, while its build cost drops from $150,000 to $30,000. Who keeps the gap?

11 min read 2525 words Updated September 2026 /a/wage-anchored-pricing-automation

Note: This is a research note supplementing the book Unscarcity, now available for purchase. These notes expand on concepts from the main text. Start here or get the book.

Wage-Anchored Pricing: Why Cheaper Robots Don’t Mean Cheaper Goods

The most important sentence in the prospectus of the first American humanoid-robot company to file to go public is not about the robot. It is about the person the robot replaces.

On September 4, 2026, Churchill Capital Corp XI filed its Form S-4 for the $2.5 billion merger that will list Agility Robotics, the Oregon company whose Digit humanoid has been moving totes at a GXO warehouse in Georgia since June 2024. Its “Illustrative Unit Economics” section, around page 132, states the pricing rule in a single clause:

“Under the RaaS model, subscription pricing is generally structured at a discount to a customer’s fully burdened labor cost, allowing customers to realize a positive economic return immediately upon deployment while avoiding significant upfront capital investment.”

RaaS is robots-as-a-service: Agility keeps the robot, the customer rents it. And the rent is not derived from what the machine costs to build. It is derived from what the human it stands in for costs to employ. The customer pays a bit less than a worker. The vendor keeps everything between that number and the bill of materials.

That is a legitimate way to sell a new capital good. It is also, if it holds, the mechanism by which the abundance this book describes fails to arrive.

The table the securities law made them publish

Securities law has one great virtue: it forces a company to write down what it tells investors in private. So for the first time, a humanoid maker has published its own cost curve next to its own price.

The price side, RaaS model: $8,500 a month per Digit, software and maintenance included, plus a one-time deployment fee of about $25,000, on a robot with a five-year useful life. Over that life, one robot bills about $535,000.

The cost side, quoted from the filing: “The cost to build a robot will be approximately $150,000 at commercial launch. At the annual production of 1,000 robots, Agility targets a robot BOM of approximately $75,000. At the annual production of 10,000 robots, Agility targets a robot BOM of approximately $30,000.” Add Agility’s own running costs, about $15,000 a year to deliver the software and maintenance and $15,000 to deploy each unit.

Now do the arithmetic the filing invites. At launch cost, a robot that bills $535,000 costs Agility about $240,000 to build, deploy and support over five years: 2.2 times its cost. At the 10,000-a-year cost, the same robot costs about $120,000 to build, deploy and support: 4.5 times its cost. The customer’s invoice is identical in both rows. Eighty percent of the bill of materials disappears between the first row and the last, and not one dollar of it is scheduled to reach the customer, let alone the warehouse worker whose payroll set the price in the first place.

The company is candid that the curve is a target. Its risk factors warn that “there can be no assurance that anticipated bill of materials cost reductions or manufacturing efficiencies will be realized within the expected timeframe.” Fair enough. But notice what the target is a target for. It is a target for margin, not for price.

What “a discount to labor” actually buys

The benchmark in Agility’s June 2026 investor deck is a fully burdened warehouse worker at about $30.50 an hour: wages, benefits, overtime, and the standing recruiting cost of a job whose turnover often runs above a third of the workforce a year. Set $8,500 a month against that and something odd appears. $8,500 buys 279 hours of that human. A full-time person works about 173 hours a month. On a single shift, Digit costs $49 an hour, half again as much as the worker. On two shifts it costs $25. Around the clock, about $12.

So the “discount” is real only when the robot works the hours of roughly 1.6 people, and the pitch is honest about that: the robot does not go home. That is what machines are for. But it tells you exactly where the price is pegged. Not to the cost of the robot. Not even to the cost of one worker. To the payroll of all the workers a tireless machine can cover across a 24-hour day. The anchor is the wage bill.

For contrast, look at a humanoid priced without a wage in the room. 1X rents its Neo to households for $499 a month, because a household has no payroll to peg to. Digit is a different class of machine, industrial-duty and built for double shifts, and some of the gap is engineering. But a seventeen-fold difference in monthly rent is not a duty-cycle difference. It is a difference in what the price was anchored to.

An old trick with a new body

None of this is new. It is value-based pricing, the oldest move in the industrial playbook: charge what the alternative costs, not what your product costs.

Xerox did it in 1959. The 914 copier was leased for about $95 a month plus a few cents for every copy past the first 2,000, a price set against the cost of a typist retyping the page, not against the cost of toner. Software did it for two decades with the per-seat license: Microsoft’s 365 Copilot is $30 per user per month, a price that scales with headcount and therefore with payroll. And now that AI is shrinking headcount, watch the software industry sprint to a new anchor. Salesforce launched its Agentforce agents at $2 per conversation. The unit changed; the logic did not. Peg the price to the work replaced, never to the cost of doing it.

Economists have a name for the space between what something costs to make and what a buyer would pay rather than go without: surplus. Under competition it flows to buyers, which is the entire reason falling costs have ever turned into falling prices. The Model T went from $850 to $260 in seventeen years because Ford had to pass the assembly line through to the customer to sell ten million cars. A wage anchor is an engineering decision to route the surplus the other way. William Baumol explained why a haircut gets dearer every decade: its productivity stands still while wages elsewhere rise. Wage-anchored pricing is Baumol run backwards. Productivity rises fivefold, and the price stays exactly where the wage was.

Why this matters for abundance

The argument of Unscarcity rests on a chain: the marginal cost of a unit of work, a kilowatt-hour and a token collapse; prices follow cost toward zero; and a civilization can then afford a Foundation that hands every Resident food, shelter, healthcare, energy and compute unconditionally. The Labor Cliff is the first link. Wage-anchored pricing attacks the second.

When the input goes free, this site argued, the money does not vanish; it migrates to whichever adjacent layer stays scarce: energy, compute, distribution, permission. Here is a case where the scarce layer is not physical at all. It is the pricing power to keep the price where the wage used to be. Agility’s cost curve is doing exactly what the book predicts, $150,000 to $30,000. Its price curve is flat by construction.

Three things follow, none of them good.

First, the displaced worker’s wage does not turn into cheaper goods. It turns into vendor margin, and then into a $2.5 billion valuation whose entire premise is the spread between a wage-indexed price and a collapsing cost. The prospectus is asking public investors to buy that spread.

Second, the deflation the book counts on arrives muted. A warehouse operator sees its labor line fall by a fifth or a third, not its cost line fall by four fifths. The customer’s savings are real, and small, and they stop at the customer.

Third, and worst, a Foundation that has to buy embodiment at wage parity is not a Foundation. It is an employer with a different logo. The book’s floor works only if bodies, like roads, are priced at what they cost.

This is Law 3 in action: Power Must Decay. A five-year lease on a machine the vendor still owns, running software the vendor still controls, at a price pegged to the wage of the person it replaced, is a structure for accumulating surplus at one layer for as long as nobody breaks the anchor. Compute as Collateral described the same pattern one layer up: an asset priced on scarcity cannot be the instrument that delivers abundance.

What breaks the anchor

Anchors do not hold by themselves. Three things cut them.

Competition. Unitree sells its G1 humanoid for about $16,000, and Chinese manufacturers ship roughly nine of every ten humanoids on Earth. Tesla’s stated Optimus target is $20,000. Against those numbers, $8,500 a month survives only behind a wall, and there is a wall. On July 28, 2026, the FCC added foreign-made humanoid robots to its Covered List, denying new Chinese models the equipment authorization they need to be sold in the United States. Five weeks later, the S-4 landed. The tariff that protects a wage anchor turns out not to be a tariff at all. It is a certification rule. The Humanoid Robot Revolution has the full story of that order and of the fleet already inside the wall.

Ownership. The same S-4 quietly publishes the alternative. Buy the robot outright for about $200,000, pay $20,000 to deploy it and $36,000 a year for software and maintenance, and five years cost about $400,000 instead of $535,000. Owning the body already cuts the bill by a quarter, at today’s $150,000 build cost. At the $30,000 cost, the purchase price becomes the next negotiation. And the $36,000-a-year software line is the next anchor, because a vendor who loses the wage peg on hardware will try to reattach it to the subscription. The book’s Mission Guilds, production organizations that build at cost for the Foundation rather than at value for a customer, exist for exactly this reason.

Visibility. You cannot negotiate against a cost you cannot see, and until this month nobody outside Agility had seen the cost curve. Securities law did what no customer could: it made the vendor publish the bill of materials next to the price. Law 2, Truth Must Be Seen, is not a slogan here. It is the precondition for the other two levers. A Foundation buying robots by the thousand should demand cost-plus disclosure as a term of sale, the way public utilities have had to open their books to a rate board for a century.

The honest caveats

Agility is a company with $1.8 million of 2025 revenue and an operating loss above $140 million. The valuation rests on a deployment schedule of roughly 7,000 robots by 2030, a quarter of the 28,000-robot “unadjusted pipeline” the same filing floats. The 10,000-a-year cost row may never be reached, in which case the anchor never gets tested. And an early vendor pricing against labor to fund its own factory is not villainy. It is how every new capital good, from the copier to the mainframe, was first sold.

The question is not whether Agility may price this way in 2026. It is whether the pricing rule is still standing in 2035, when the cost is $30,000, the fleet is 25,000 units, and the sentence in the unit-economics section has become an industry norm that nobody thinks to read.

The bill for abundance is a pricing decision

The book treats abundance as deflation, and it is right: nothing about a $30,000 robot is scarce. But deflation is not a physical constant that falls out of a cost curve. It is a price, and prices are decided by whoever holds the anchor. The Abundance J-Curve showed the buildout’s bill landing on households first; this is the same fight seen from the other side, the payoff held at the vendor’s layer while the cost falls away underneath it.

That is the case Unscarcity makes: the machines will get cheap on their own. Cheap goods, cheap floors, a Foundation that can afford its own bodies, those have to be won, one anchor at a time, by competition, by ownership, and by making the cost visible enough that someone can point at the gap and ask who is keeping it.

Sources

Share this article:

Share Share Share