Sustainability
Last Updated: April 23, 2026
Shine is built on AI. These systems consume significant water and energy, and the impact falls disproportionately on communities near data center infrastructure.
We are building a model where our compute usage contributes directly to water restoration, with measurable, verifiable impact.
Our principle
Sustainability claims should be grounded in real, measurable impact, not performative commitments or unverifiable offsets.
We apply the same bar to our own company that we apply to the customer claims we help prove.
Our footprint
Shine operates as a fully remote, distributed company with a low physical footprint. We do not maintain offices and minimize travel.
Our largest environmental cost comes from AI compute: the inference and model runs that power story extraction, claim analysis, and the rest of our product.
Our approach
As our product runs, a portion of our compute spend will flow to organizations working on measurable water impact.
When the model is live, we will publish exactly what it is and how it is calculated on this page.
Our partners
When the compute-linked model begins, our initial recipients will be organizations focused on water systems in the Americas that publish outcome data we can evaluate.
The Nature Conservancy
Watershed health and groundwater systems across the Americas. Impact and accountability data published at nature.org/accountability.
Bonneville Environmental Foundation
Measurable, project-level water restoration in U.S. watersheds, including their Water Restoration Certificates program.
We will add or remove partners based on the quality of evidence behind their impact.
Our standard
Every figure or claim on this page links to either our own underlying data or the partner organization's published impact reports. If something here cannot be backed by evidence we can point at, it does not belong on this page.
We hold our own sustainability statements to the same proof standard we hold our customers' claims to.
Questions
For questions about this page or our approach: