Macrocosmos launches iota SDK and Liquid Compute to train on disaggregated global compute

markjeffrey · x · 2026-09-29

At Montreal's Exploit Summit, Macrocosmos (built on Bittensor/TAO) unveiled Liquid Compute, a disaggregated compute platform turning the long tail of global compute into trainable capacity, and the iota SDK, which lets developers define a workload normally while the platform handles distribution, execution, and recovery across heterogeneous compute. The company says its flagship results reproduce in 500 lines of code, and it will go to market in the coming weeks with early-access registration open.

Original post →

More from Infra

Infra channel →