PADO, an energy-orchestration platform backed by LG NOVA, has announced a partnership with VESSL to develop a joint grid-aware Machine Learning Operations (MLOps) solution for AI workload orchestration. The companies say the goal is to reduce the operational impact of fast-growing AI loads on the electric grid, energy pricing, and data center operations by aligning when and where compute runs with energy conditions.

According to the announcement, the joint platform is positioned as an “energy-oriented MLOps workflow” aimed at data centers, hyperscalers, and tenants. PADO and VESSL say it will automate orchestration that ties machine learning workloads to “energy economics and renewable availability,” with an emphasis on maximizing compute usage while also creating “new revenue streams for data center operators.”

The companies say the integrated approach combines PADO’s grid-aware workload scheduling with VESSL’s multi-cloud AI orchestration into a unified “energy-smart MLOps layer” that operators and tenants can deploy without “overhauling existing infrastructure.” PADO and VESSL position this as a way to increase flexibility in data center operations as AI usage grows.

In the release, PADO and VESSL say the partnership enables automated GPU and AI workload shifting to lower-cost and renewable-abundant windows; grid-based dynamic workload routing across clusters or regions; continued machine learning experiment reproducibility and tenant Service Level Agreements (SLAs); monetization of flexible compute as a grid resource for demand response, firm frequency response, capacity, and energy arbitrage; and optimized energy consumption without interrupting AI research and development (R&D) or production pipelines.

Source: PADO