The same compute, paid twice.
Pool miners supply the GPUs. Every hour our scheduler puts each one on the work that pays most. On the Peryn API, batch and fine-tuning, every AI job also mines, so the same compute is paid twice.
Three steps, every hour.
-
Miners supply GPUs to Peryn Pool.
They run our agent on their GPUs, and each GPU joins the pool.
-
Every hour, our scheduler picks what pays most per GPU.
It compares serving AI, which earns for the tokens and mines while it serves, with pure mining, and runs whichever pays more.
-
The miner keeps the $PEARL, plus a USD share.
Peryn takes the pool fee and a margin.
AI work that also mines.
USD for the AI work, $PEARL for the mining: two kinds of mining run on every job of the Peryn API, batch and fine-tuning.
- Inside the work
- The AI work itself produces Pearl proofs, so the GPU mines while it computes the answer.
- In the gaps
- Between requests, and between the steps of a job, the GPU mines.
One pool, three ways to use it.
Every product runs on the pool’s GPUs, and every one of them mines.
Put your GPUs in the pool.
Run our agent on your GPUs. The scheduler does the rest, every hour.