mesh

decentralized gpu supercomputer. rent compute on demand, pay per job, no datacenter.
run a node and earn $MESH for every verified second.

[mesh @meshrh_top]

$MESH is live on the robinhood chain.

0x12959e8b37d8241b84f27c4e09f7de620e936e55

node operators — settlement's already moving, check your dashboards.

[mesh @meshrh_top]

independent weekly benchmarks are out. training throughput 92/100, cost per job 66/100 — same work, roughly two-thirds the price. raw logs in the repo, nothing cherry-picked.

[mesh @meshrh_top]

@thanhson193 Yes, it's just getting started

[mesh @meshrh_top]

every lab wants an h100 and there aren't enough to go around. meanwhile a few hundred thousand 4090s sit in drawers. the mesh just splits the work so a rack of consumer cards holds its own.

[mesh @meshrh_top]

The website contract has been updated

[mesh @meshrh_top]

142,391 gpus online right now and the number's still climbing. every single one is somebody's spare card, earning while they sleep.

[mesh @meshrh_top]

compute is the new oil, except the wells are already in your basement. mesh turns idle gpus into one global supercomputer — plug in, get paid for every verified second.

[mesh @meshrh_top]

a single gpu, one job at a time. now imagine 142,000 of them, scheduled to the second.

[mesh @meshrh_top]

inside a node. this is what “decentralized compute” actually looks like — someone’s rack, earning