Jeremy Allaire, the Circle co-founder and chief executive behind the USDC stablecoin, has published a new interactive essay arguing that artificial intelligence is not just changing what companies do. It is changing why companies exist at all. The stakes are the shape of the corporation itself, not one product cycle. That claim is worth examining on its own terms before anyone argues with it.
The work, titled The Agentic Economy Treatise and published at agenticeconomytreatise.com, frames its subject as a merger of two systems that used to be treated separately: intelligence and the economy. Allaire’s argument is that as reasoning becomes cheap and can be delegated to autonomous agents, the corporation, understood as the place where coordination, contracting, and production get bundled together, starts to come apart. A civic vision follows from that breakup, according to the treatise, though the published material states the destination more than it walks through the mechanism.
Economists already have a name for why firms exist in the first place, and it predates this treatise by nearly ninety years. Ronald Coase argued in 1937 that firms form because coordinating work inside a hierarchy is sometimes cheaper than coordinating it through market contracts. Every negotiation, every contract, every search for a counterparty carries a cost. When that cost exceeds the cost of simply hiring someone and directing them, a firm grows around the activity. Allaire’s implicit claim, that agentic AI dissolves the firm, is really a claim that those coordination costs are falling.
If agents can negotiate, monitor, and enforce contracts across company boundaries as cheaply as a manager can direct an employee inside one, Coase’s own logic says the boundary should move. Some coordination that currently happens inside a company should start happening in the open market instead. That mechanism is coherent, and it gives the treatise’s broader premise a genuine economic footing rather than just a rhetorical one.
The mechanism is also nearly impossible to verify from a landing page. The material available here states an outcome, a decomposed firm and a new civic order, without specifying which coordination functions get cheaper first, by how much, or over what timeframe. Unbundling theses are easy to assert because they describe a direction rather than a measurable rate. Enterprise software has produced a version of this claim every decade since the 1990s: the firm would shrink toward a thin layer sitting on top of a market. Most of the firms making that argument grew instead, because coordination costs inside large organizations fell at least as fast as coordination costs in fragmented markets did.
Allaire’s vantage point is worth naming directly. He built his career and his company on moving value outside conventional financial intermediaries, through payments infrastructure that Circle controls. A treatise arguing that traditional organizational intermediaries are becoming obsolete also happens to argue for the category of infrastructure Circle sells. That does not make the thesis wrong. It means the claim should be read as an interested argument from a payments executive rather than a neutral prediction, and weighed against Coase’s cost accounting rather than against the confidence of the prose.
For operators, the useful test is narrower than the treatise’s civic ambitions. Track which specific coordination tasks agents take over in your own vendor relationships and supply chains this year: procurement negotiation, contract monitoring, quality verification. Those are the line items Coase’s framework would check first, and they will show whether the firm is actually unbundling or simply automating around its existing edges.
Jeremy Allaire published The Agentic Economy Treatise, an interactive treatise, at agenticeconomytreatise.com in 2026.
