企业温室气体清单中的 AI 推理采购核算:一种基于 token 的方法
Accounting for Purchased Artificial Intelligence Inference in Corporate Greenhouse Gas Inventories: A Token-Based Method
AI 导读
一篇论文提出基于 token 的方法核算企业 Scope 3 类别 1 中采购的 AI 推理排放,替代按支出折算(后者为物理估算的 1.7 至 12.3 倍)的做法。
正文 · 原文
Abstract:Purchased artificial intelligence (AI) inference belongs in Scope 3 Category 1 of a corporate greenhouse gas inventory, yet practice omits it or applies a spend-based factor that gives 1.7 to 12.3 times the physical estimate and cannot tell one model or region from another. This paper sets out a token-based method and open dataset: a cradle-to-gate boundary, four tiers of activity data, an effective-token basis, a four-term factor (serving electricity, embodied hardware, data-centre construction, disclosed final training runs) and a rule resolving any model identifier and region to a factor. Inference footprints span twelve times between model classes and nine between regions. The mid-class factor, 0.039 kg of carbon dioxide equivalent (CO2e) per million billed tokens at the world-average grid, is the low end of a stated structural range: a less efficient serving configuration raises it 2.8 times, and replacing the training term with all model development, on its own boundary, gives 1.5 to 2.9 times it. Input bands alone give -18% to +29%. Published per-query energy estimates that can describe a mid-class model are 2.4 to 7.2 times the mid-class energy, above its structural range in four of five rows; those that can describe a frontier-class model bracket it at 0.7 to 1.6 times. Worked examples give 8.5 kg CO2e a year for a 200-person firm, and 2.1 tonnes for fifty developers using agentic coding tools on a mid-class model or 8.3 tonnes on a frontier-class one. The contribution is a traceable figure; accuracy against metered energy is untested.
| Comments: | v4: accepted manuscript, with the author's proof-stage corrections. Published in Sustainable Production and Consumption 70 (2026) 320-347, this https URL. Dataset v3.0.0 (this https URL). 139 pages, 11 figures, 39 tables including the supplementary material |
| Subjects: | Computers and Society (cs.CY); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.10660 [cs.CY] |
| (or arXiv:2606.10660v4 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2606.10660 arXiv-issued DOI via DataCite |
|
| Journal reference: | Sustainable Production and Consumption 70 (2026) 320-347 |
| Related DOI: | https://doi.org/10.1016/j.spc.2026.11.003
DOI(s) linking to related resources |
Submission history
From: Guillermo Llopis [view email]
[v1]
Tue, 9 Jun 2026 10:08:36 UTC (1,238 KB)
[v2]
Thu, 15 Oct 2026 14:25:34 UTC (3,088 KB)
[v3]
Tue, 3 Nov 2026 09:21:18 UTC (3,089 KB)
[v4]
Wed, 18 Nov 2026 08:11:06 UTC (3,049 KB)
来源:arXiv cs.AI · arxiv.org