Energy per Successful Goal: Goal-Level Energy Accounting for Agentic AI Systems

Deepak Panigrahy, Aakash Tyagi

arXiv:2605.22883 · 2026-05-25 공개 · arXiv · PDF

agentic-ai tool-augmented energy-accounting orchestration-overhead workflow-energy inference-energy ai-benchmarking a-lems

Abstract

Current AI energy benchmarks measure consumption at the granularity of a single model invocation or training run. For classical single-turn workloads this unit remains coherent. For agentic systems - where a single user goal may trigger multi-step orchestration, tool calls, retries, and failure-recovery cycles - the invocation count is an implementation artifact rather than a task property, and inference-level normalization misrepresents the energy cost of goal completion. We present A-LEMS (Agentic LLM Energy Measurement System), a cross-layer measurement framework that redefines the unit of AI energy accounting from energy per inference to Energy per Successful Goal (EpG). EpG aggregates total workflow energy across all execution attempts, including failures and retries, normalized by successfully completed goals. A-LEMS formalizes energy attribution through a temporal boundary model, a five-layer observation pipeline mapping RAPL signals to workflow-level energy, and a reproducibility protocol binding every measurement to hardware and runtime configuration. Building on EpG, we define the Orchestration Overhead Index (OOI), isolating the energy cost of orchestration relative to linear execution under identical task criteria. Across five reasoning and three tool-augmented task families, agentic workflows consume 4.33x higher mean energy per successful goal than linear baselines (888.1 J vs 205.3 J). This overhead is driven by orchestration structure, not inference compute. For tool-augmented tasks, OOI inverts below 1.0x: agentic execution is cheaper than linear, confirming the metric captures orchestration structure rather than a fixed upward bias. These findings establish that energy-per-inference is insufficient for agentic AI. EpG and OOI provide the measurement foundation for accurate benchmarking, where orchestration structure is the primary determinant of energy cost.

한국어 요약

📋 한 줄 요약

**[Agentic AI 에너지 측정]** A-LEMS가 inference 단위 대신 Energy-per-Successful-Goal(EpG)·Orchestration Overhead Index(OOI)로 agentic 워크플로우를 측정 — reasoning 태스크에서 linear baseline 대비 4.33×(888.1 J vs 205.3 J) 에너지 소비.

🎯 핵심 기여도

💡 핵심 아이디어

Agentic AI는 단위 inference가 아닌 "성공한 목표"로 에너지를 회계해야 하며, 이를 통해 orchestration 구조 자체가 에너지 비용의 1차 결정 요인임이 드러난다 — 측정 단위 변경이 곧 benchmarking 패러다임 변경.

🔬 기술적 접근법

📊 주요 결과

💭 의의 및 한계

**의의**: Agentic AI의 에너지 평가 표준을 제시, orchestration 구조가 비용을 좌우함을 정량 입증, 그린 AI·운영비용 분석의 새 기반. **한계**: RAPL 의존 — 비-Intel·GPU 중심 워크로드에 직접 적용 한계, 8개 task family로 일반화 추가 검증 필요, OOI 정의가 비교 가능한 linear baseline에 의존.

🚀 실용적 활용