FINOPS MEETS GREENOPS: A CARBON-PRICED GOVERNANCE FRAMEWORK FOR AI WORKLOADS IN PUBLIC-SECTOR CLOUD
International Journal on Cloud Computing: Services and Architecture · 28.10.2026 · 16 (5) · 01-17
DOI: 10.5121/ijccsa.2026.16501 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Cloud Computing: Services and Architecture, online erschienen am 28.10.2026. Verfasst von Lu Yan.
Abstract (Originalsprache)
Public-sector organisations are moving AI workloads onto public cloud under two pressures managed separately: spend, through FinOps, and carbon, through GreenOps. GPU services aggravate both, and dataresidency rules remove the lever on which carbon-aware scheduling relies, regional shifting. This designand-protocol paper proposes pricing cost and carbon on one scale: each workload is charged its cloud cost plus a shadow carbon price taken from the organisation’s appraisal values, delivered through the existing chargeback channel with a budget-neutral recycling rule. We state when the rule aligns team and organisational objectives, specify how energy and emissions are attributed to workloads and graded by evidence, and map the framework onto an Azure reference architecture. A scenario analysis finds the carbon charge is 0.5-1.5% of GPU price on the GB grid, so it reverses only near-tie decisions; a sensitivity analysis gives the reversal condition.
Zitieren
APA
Yan, L. (2026). FINOPS MEETS GREENOPS: A CARBON-PRICED GOVERNANCE FRAMEWORK FOR AI WORKLOADS IN PUBLIC-SECTOR CLOUD. International Journal on Cloud Computing: Services and Architecture, 16(5), 01-17. https://doi.org/10.5121/ijccsa.2026.16501
Vancouver
Yan L. FINOPS MEETS GREENOPS: A CARBON-PRICED GOVERNANCE FRAMEWORK FOR AI WORKLOADS IN PUBLIC-SECTOR CLOUD. International Journal on Cloud Computing: Services and Architecture. 2026;16(5):01-17. doi:10.5121/ijccsa.2026.16501
BibTeX
@article{Yan2026,
author = {Yan, Lu},
title = {FINOPS MEETS GREENOPS: A CARBON-PRICED GOVERNANCE FRAMEWORK FOR AI WORKLOADS IN PUBLIC-SECTOR CLOUD},
journal = {International Journal on Cloud Computing: Services and Architecture},
year = {2026},
volume = {16},
number = {5},
pages = {01--17},
doi = {10.5121/ijccsa.2026.16501}
}
EUVIMED