Resource Consumption
Artwork (Unknown) from Cosmos
We recommend using ByteThirst to track your AI activity.
Orelier has announced a long-term R&D initiative to build an open-access AI resource awareness tool, targeting a public release in 2029. The software would help users estimate the water and electricity footprint of their active AI use while introducing resource-aware prompt literacy. This guidance helps people recognize when AI is useful versus when a task is better handled through human thought or convos.
This project responds to a critical global reality. Generative AI is no longer an experiment, with the DataReportal 2026 digital update estimating 2.42 billion active users globally. Meanwhile, the International Energy Agency reports that global data centers consumed roughly 415 terawatt-hours of electricity in 2024 and could more than double that demand by 2030, driven heavily by AI. Large data centers consume millions of gallons of water daily, straining local watersheds and electrical grids. While we do not support excessive AI use or a future where every ordinary act of thinking is routed through energy-intensive systems by default, AI is already woven into daily life. Rather than pushing back with empty criticism, we accept the reality of the situation. We are contributing our niche angle as a tech nonprofit to drive necessary change.
Important backend and tracking innovations already exist in this space, including developer emissions libraries like CodeCarbon and EcoLogits, alongside other early-stage consumer tracking extensions. Rather than building a competitor from scratch, Orelier is highly interested in pairing with these existing technologies to integrate their resource-tracking backends directly into our platform. This integration will allow us to focus our development on the user interface and prompt-engineering layer, directly addressing token waste where it starts. Instead of offering false precision, the integrated tool would track estimated water and electricity usage in transparent ranges, acknowledging that exact prompt-level data requires disclosures that tech companies currently withhold. It would help users identify low-yield AI use (including vague or impulsive prompting) to shift the focus from how to use AI more to whether to use it at all.
To fund and build this tool responsibly, we will pursue a staged, four-year roadmap. In 2026, we will focus on research, publishing, and coalition-building. In 2027, we will secure fiscal sponsorship, philanthropic support, climate and public-interest tech grants, and university partnerships. In 2028, we aim to develop a modest working prototype with advisors across UX, environmental data, and AI policy. In 2029, we plan to release the open-access version with clear methodology notes, limited beta testing, and honest explanations of what the tool can and cannot know. AI sustainability cannot be solved only at the data center, it must be addressed at the interface and at the user's moment of decision. Our goal is to contribute a public-facing tool that protects human curiosity from becoming unconscious consumption.
