AIforE's operating framework for making AI sustainable at scale.
ROI isn't just return on investment — it's return on impact. Sustainable AI you can't measure is just a guess, so we build the tools that make it measurable.
We build tools, tech, and frameworks to right-size AI solutions — matching model and infrastructure to the task, not the hype.
We promote and facilitate renewable energy — including SMRs and safe nuclear power alongside solar, wind, and other clean sources — helping AI run on clean power, not just more power.
We promote and facilitate recycled cooling, reclaimed water, and thermodynamics — turning waste heat and water into a resource, not a liability. We're extending the pillar to a lifecycle standard for data-center hardware reuse and recycling.
We remediate ecological harm — cutting greenhouse gases, mitigating extreme weather and erosion, restoring habitats, and advancing desalination and space-based solar power.
Spotlighting organizations and projects that are getting AI sustainability right.
Videos, podcasts, and articles that bring sustainable AI to a broad audience.
Specifications, tools, and technologies to benchmark AI sustainability — including new tokens-to-carbon-output citations.
Independent ratings that let organizations prove — and improve — their sustainability footprint.
A standards contributor for AI sustainability, contributing to IEEE and more.
Hands-on guidance to help organizations design and operate sustainable AI.
Connecting best practices, products, and services across our community.
A running news archive that also works to debunk AI sustainability misinformation.
Flagship events plus regional event kits so local communities can host their own.
Practical guidance grounded in real showcase examples, not theory.
Independent research backing each pillar of our framework.
Most AI investment still can't prove its return, and almost none can prove its impact: How to Maximize AI ROI in 2026 (IBM) and AI ROI Statistics 2026 (Actigy). Measurement is the missing layer — which is what our benchmark framework and calculator are built to provide.
Matching model size to the task is one of the biggest levers available: Small Changes in AI Models Can Produce Big Energy Savings (UNESCO / TechNewsWorld) and Right-Sized AI: Why Small Language Models Outperform Giants (CloudGeometry).
AI's power demand is colliding with the grid, and clean power is the release valve: Global Energy Demands within the AI Regulatory Landscape (Brookings) and 2026 Predictions: AI Sparks Data Center Power Revolution (Data Center Knowledge). We consider SMRs and safe nuclear power part of that clean-power mix alongside solar and wind.
Server heat doesn't have to be wasted — it can warm homes and communities: Thermal Energy Networks Turn Data Center Waste Heat into a Hot Commodity (EESI) and Data Center Heat Reuse: How to Make the Most of Excess Heat (TechTarget).
From removing greenhouse gases to harvesting solar power from orbit: Greenhouse Gas Removal (The Royal Society) and China Conducts First Experiments for Space-Based Solar Power Plants (pv magazine).
AI's resource cost is usually invisible to the person making a query. Pick a model, a query size and where it runs, and this calculator estimates the cost, energy, water, carbon and heat behind your AI queries, in plain units — using the same method as the Stardust browser extension, and a first step toward the tokens-to-carbon-output methodology in our Benchmark Framework v1.
Our methodology and sources are published openly on GitHub, so the numbers can be checked, not just trusted.
Uses the same draft methodology (v0.2) as the Stardust browser extension. Energy per token by model size (small, mid, frontier), anchored to ~0.3 Wh for a typical frontier-model reply — Oviedo et al., Joule (2026), and Epoch AI (2025). Unlisted models use the frontier figures and are marked Estimated. Prices are list prices per token, September 2026. Grid carbon: U.S. EIA (2023) and Ember (2024 world average); France and Sweden are approximate 2023 mixes. Water: 1.8 L/kWh on-site cooling (LBNL, 2016) plus 3.14 L/kWh for power generation (Li et al., 2023). Heat: essentially all electricity used by servers becomes heat. Indicator code = impact grade (CO₂e per 1,000 tokens, A best to F) + cost tier (per query, 1 lowest to 5, X unknown) + size (tokens per query, XS to XL). Energy, grid and grade figures are draft placeholders pending review, published openly on GitHub. Forest equivalent: ~1.5 Mg C/ha/yr (≈0.55 kg CO₂ per m² per year) for 20–100-year-old Pacific Northwest conifer stands, from Fig. 5 of Gray, Whittier & Harmon, Ecosphere 7(1):e01224 (2016); excludes soil carbon and varies widely by site. Each icon ≈ 0.1 m². Chat apps resend the whole conversation with each message, so input grows as a chat goes on; reasoning models also use hidden thinking tokens.
Publish AIforE's first AI sustainability benchmark framework and tokens-to-carbon methodology.
Ship the rebuilt AIforE.org plus community tools like the token-cost calculator.
Host our first bi-annual event and launch the newsletter and weekly AI sustainability news.
Secure a formal contributor role with IEEE and allied AI sustainability standards efforts.
Grow the board toward 8 members, add part-time program leadership, and put D&O insurance in place.