Helicyn research

• Coordinating compute, energy, cooling, and grid signals together beats optimizing any one of them in isolation.
• Consolidation was the single biggest lever, cutting simulated energy, carbon, and cost by roughly 8% by reclaiming idle server capacity.
• No policy wins on every metric: the integrated coordination policy traded higher energy and more deadline misses for lower carbon and better thermal safety.
• Everything here is simulated, under documented assumptions, and not yet validated on real hardware.

Read the full report →

The next constraint on artificial intelligence may not be model capability but the coordination of the physical infrastructure needed to power, cool, and operate large-scale computation. As cloud computing, AI training, and large-scale inference grow more common, data centers are no longer passive back-end facilities. They are becoming active nodes in a wider infrastructure system, shaped continuously by workload demand, GPU availability, thermal headroom, electricity prices, carbon intensity, cooling capacity, and regional grid conditions. In the United States alone, data center electricity use reached 176 TWh in 2023 and is projected to rise to 325–580 TWh by 2028, driven partly by the rapid deployment of GPU-accelerated AI servers and partly by uncertainty about future operating practices. The central limitation of existing research is that these methods usually act as isolated optimization tools rather than connected layers of a single coordinated system. This paper argues that energy efficiency, thermal stability, and resource utilization are emergent properties of coordinating workloads, regions, facilities, and local control systems together, and proposes a four-stage hierarchical framework: detection, coordination, optimization, and verification. To test the argument concretely, the paper pairs this framework with a reproducible, simulator-based evaluation that runs seven scheduling policies across six operating scenarios and five random seeds, using public workload, grid, carbon, and weather data. The results are deliberately modest and specific: no single policy dominates every objective. Consolidation produces the largest efficiency gains, cutting facility energy, carbon, and cost by roughly 8% by reclaiming idle server capacity, while an integrated multi-objective policy does not reduce energy outright but shifts the operating point toward lower carbon and thermal risk at the cost of higher energy and more deadline misses. The evaluated integrated policy is a fixed-weight heuristic controller inside an ML-ready architecture, not a production-proven autonomous system.

Read the full abstract and paper →

Three findings carry the argument. First, the largest efficiency lever was reclaiming idle capacity: consolidation reduced energy, carbon, and cost by roughly 8% by reducing active server-hours, and facility energy tracks active server-hours more tightly than it tracks any timing decision. Second, single-objective policies improved their target metric but could regress others; the DVFS-aware policy regressed on every headline metric precisely because it optimized power states in isolation from placement. Third, the integrated coordination policy did not dominate; it traded higher energy and lower deadline reliability for lower carbon and better thermal safety, which is the honest signature of coordinating competing objectives rather than eliminating their conflict.

All findings are simulated, under documented assumptions, and research prototype; none represent production savings or real facility validation. The full results, comparisons, and figures for every scheduling policy and scenario are in the report.

Read the full results and figures →