On October 27, Arthur Hayes published a detailed essay discussing the intricacies of futures and perpetual swap contracts in relation to the Flop Network’s proposed computing market. This concept revolves around utilizing AI agents compensated with FLOPs—a specific cryptocurrency designed for computing power transactions.
Overview of the FRCR Futures Model
In his essay, aptly titled “The Trifecta,” Hayes outlines a framework intended for entrepreneurs eager to develop derivatives connected to an AI-driven economy. A key aspect of his proposal is the creation of a viable spot market for computing power, which would serve as a pricing benchmark for futures and perpetual swaps associated with the resources consumed by AI systems.
At the heart of this proposal is the Flop-based Compute Rate (FRCR), which is calculated on an hourly basis through a volume-weighted price of FLOPs per gigaflop of computational capacity. Quarterly futures contracts would utilize US dollars for margin requirements and profit calculations, with settlements occurring on the last Friday of March, June, September, and December.
Importantly, Hayes emphasizes that this dollar-centric structure necessitates the use of quanto derivatives—contracts that handle payments in a different currency from the asset in question. Although the computing market will be priced in FLOPs, traders may prefer their returns to be denominated in US dollars, necessitating adjustments that reflect the relationship between the FRCR and the FLOP-to-dollar exchange rate.
Perpetual Swaps and Funding Model
The essay also discusses a proposed perpetual swap mechanism, which would calculate funding disbursements based on an 8-hour time-weighted average that compares the transaction price to the FRCR. Similar to existing perpetual contracts in the cryptocurrency market, the pricing dynamics would result in higher-cost traders compensating those on the lower-end, based on the premium paid.
Market makers would have the capability to replicate futures exposures by buying or supplying Flop Computing Units (FCUs)—reservations of inference power allocated for specific AI models and timeframes. Additionally, to ensure accountability, suppliers would be required to issue FLOP bonds which could be forfeited should they fail to fulfill their obligations regarding computing tasks.
Foundation of a Spot Market
Hayes stresses that the success of the derivatives proposal hinges on establishing a functional spot market within the Flop Network. He critiques existing price indexes that monitor chip rental fees and individual AI token values, asserting that they provide an incomplete understanding since both represent derivatives of the actual costs associated with computation.
In his design, network blocks would document inference assignments, completed compute tasks, and the fees miners receive. These records would facilitate the calculation of exchange rates between FLOPs and available computing power, effectively offering a standardized basis for both buyers and suppliers in the market.
Introduction of Flop Labs and Token Utility
In August 2026, Hayes launched Flop Labs, introducing the project’s native token, FLOP, described as “bait for AI agents.” This initiative is grounded in the belief that autonomous agents require a monetary system allowing for direct exchanges of computing resources, rather than payment frameworks tailored primarily for human users and businesses.
