Skip to content

The public AI Energy Index

Venturi publishes a free, no-login AI Energy Index at venturi.systems/ai-energy. It is a public, navigable view of model energy, water, carbon, and eco-efficiency, built from the same signed catalog that powers the numbers in your own Venturi instance. This page documents what the viewer shows and where its data comes from.

The methodology behind every number is the same one documented in the energy, water, and eco-efficiency pages.

Where the data comes from

The viewer is a static page that reads a signed, versioned JSON dataset. That dataset is byte-mirrored from the platform control plane at api.venturi.systems/control/v1/energy-catalog and is Ed25519-signed, so the public number and the number in a customer's instance are produced from the identical, verifiable catalog. There is no separate "public" methodology — it is the same catalog, mirrored.

Because the dataset is mirrored from the catalog, the viewer also honors any active sitewide availability state for venturi.systems.

The seven views

View What it shows
Overview / Leaderboard A ranked table of every model: energy (Wh per 1k output tokens), 1-to-5 rating, water (mL per query), carbon (gCO2e per query), task, size class, derivation method, confidence, and freshness date. Default sort is energy ascending; you can re-sort by water, carbon, or eco-efficiency.
Model detail A per-model card with the tri-metric (energy / water / carbon) across short, medium, and long prompt sizes (mean ± standard deviation), latency and throughput, benchmark scores, the provider multipliers (PUE / WUE / CIF), and full provenance and caveats.
Compare Two to four models side by side, with a guardrail that keeps comparisons within the same task and size class.
Time-series Day-to-day tracking of a model's energy, flagging when its watt-hours move.
Fleet-scale impact Energy, carbon, and water projected to fleet-scale query volumes, presented with the equivalence cards below.
Reasoning breakdown Energy by reasoning effort level, making the reasoning premium concrete.
Methodology The formulas, constants, tiers, provenance, and license attributions — the same material as these docs, with the AI Energy Score named and linked.

The six equivalence cards

The fleet-scale view translates abstract totals into real-world equivalents using these exact divisors:

Equivalence Divisor
Average household electricity 1.0950 MWh/yr
University campus electricity 1,202 MWh/yr
Drinking water per person 1.2 kL/person/yr
Olympic swimming pool 2,500 kL
Gasoline passenger car 4.6 tCO2e/yr
Transatlantic flight 504 tCO2e

Energy totals map to the household and university cards, water totals to the drinking-water and Olympic-pool cards, and carbon totals to the car and flight cards.

Honest disclosure on every row

The viewer surfaces the same honest-unknown discipline the engine enforces:

  • Per-row derivation_method and confidence are shown inline, so you can see exactly how each number was produced (measured, infra-aware estimate, class analogue, or unestimated) and how much to trust it. Confidence is encoded as a text-and-badge indicator, not by color alone.
  • An uncertainty band accompanies estimated rows.
  • Null is shown as null, never zero. A model with no available data is visibly unrated, not silently counted as emission-free.
  • A persistent coverage-ratio banner on the Overview states how much of the full model population the displayed catalog covers, so a partial catalog is never mistaken for a complete one.
  • A "how is this calculated?" affordance sits next to each metric, linking to the relevant methodology page.

Downloading the data

The dataset is downloadable in two formats:

  • Signed JSON — the full, Ed25519-signed dataset, so you can verify it came from the Venturi control plane unaltered.
  • CSV — a flat table for spreadsheets and analysis.

Every numeric field in the dataset is nullable (null, never a fabricated zero), and the provenance, derivation_method, and confidence fields are mandatory and non-null on every row.

Accessibility

The viewer targets WCAG 2.1 AA: ratings carry color-independent text (for example, "Model Energy Rating 4 of 5") rather than stars alone, provenance is exposed to assistive technology, charts have data-table fallbacks, and the tables and filters are fully keyboard-navigable. Viewer state — filters, sort, and the selected model — is encoded in the URL so any view is deep-linkable and shareable.

Where to go next