# Arthur Hayes Says FLOP Rewards Will Be Weighted by Inference Volume

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/60248
Published: 2026-11-18T08:30:34.000Z
Updated: 2026-11-18T08:30:34.000Z
Section: Technology

> In Hayes’ example, a miner handling 100 GFLOPs of a block’s 200 GFLOPs receives half of its reward.

Flop Labs CEO Arthur Hayes said a miner’s share of a FLOP block’s inference volume is used to weight its share of the block reward, tying payouts to completed computing work.

Hayes’ example uses floating-point operations (FLOPs) to measure inference volume. A miner handling 100 GFLOPs out of 200 GFLOPs in one block would receive half of that block’s reward.

He said Flop Labs cannot recommend hardware or configurations after retail and enterprise miners asked what to buy and how to set it up. Hardware that runs only AI models with too little demand may receive no inference requests, leaving its operator with no income.

Hayes suggested miners review the most popular models on OpenRouter to gauge what agent applications may need. He also advised trying a current setup first and adjusting it based on actual results.

The formula adds detail to Hayes’ earlier explanation that [FLOP mining income would track demand for the models miners can run](<https://www.tokenpost.com/news/technology/57174>).



## Links in this article

- [FLOP mining income would track demand for the models miners can run](https://www.tokenpost.com/news/technology/57174)
