Grayscale's research chief said derivatives tied to artificial intelligence computing remain an immature market, while describing AI compute as a significant financial theme in a discussion with an a16z partner.
Separately, BitMEX co-founder Arthur Hayes has outlined a cash-settled derivatives proposal linking the Flop token (FLOP) to AI computing capacity.
The proposed Flop Reference Compute Rate would measure FLOP tokens per GigaFLOPS each hour, with trading volume used to weight the calculation. It would serve as the reference for a quarterly futures contract and a perpetual swap, with U.S. dollars used for margin.
Hayes's rationale is that computing lacks a worldwide reference price, even as chip and AI-token prices reflect compute costs. He has said a shared measure could take shape if AI agents use Flop to make payments. The proposed derivatives rate differs from the network's proposed spot rate, which would express FLOP relative to computing capacity.
