Arthur Hayes has laid out a blueprint for quarterly futures and perpetual swaps on AI compute. The benchmark price would come from the Flop Network, a blockchain that has not launched.

The BitMEX co-founder pitches the contracts as his bid for a third hit after the perpetual swap and synthetic dollars. However, compute buyers and market makers would hedge against a benchmark derived from a single token network.

Can One Token Anchor the Price of Compute?

Hayes argues that no global price for compute exists today. Prices are split across chip rentals and per-model token rates.

His proposed fix is the Flop Reference Compute Rate (FRCR), an hourly index built from Flop Network inference jobs. On that network, AI agents pay miners in FLOP, the project’s token, to run AI model queries.

The quarterly futures would take dollar margin even though FLOP denominates the index. This quanto structure pays in one currency on an asset valued in another.

A perpetual swap, a futures contract with no expiry, would follow with funding tied to an eight-hour premium index. Market makers, in turn, could replicate positions by rolling strips of 30-day forward compute units.

Sellers would post FLOP bonds above their fees, and a missed delivery within 24 hours forfeits the bond.

Does the Track Record Travel to Compute?

Hayes credits himself with the BitMEX perpetual swap in 2016 and the synthetic-dollar idea behind Ethena’s USDe. USDe is a stablecoin hedged with short perpetual swaps.

Its supply peaked near $14.8 billion in October 2025, however, and has since fallen below $5 billion, according to DefiLlama.

Meanwhile, the FLOP testnet timeline remains open, with Flop Labs’ page listing no launch date as of mid-October. Flop Labs targets genesis, the network’s first block, for the first quarter of 2027.

Hayes has also predicted a bailout after the AI data center boom. A hedging market could help price that risk, but only if traders accept a benchmark tied to one token.

The post Arthur Hayes Wants a Third Win With a Futures Market for AI Compute appeared first on BeInCrypto.