# Token Outbound > Token Outbound measures the energy, cost and carbon behind AI tokens, by country and by fuel, and connects places that produce sustainable AI compute with the people and companies who need it. Based in Omaha, Nebraska, United States. Founded by Michael, a licensed Professional Engineer with 30 years in civil engineering, energy, data center facilities and environmental policy. Core idea: an AI token is thermodynamic work. Fuel becomes electricity, electricity powers chips that run matrix math (training and inference), and that computation produces tokens. Token Outbound publishes three measures to compare any supplier, country or fuel on the same footing: - Tokens per kilowatt-hour: tokens produced divided by energy consumed, including cooling. - Cost per million tokens: compute cost per hour divided by tokens per hour, times one million. - Carbon per million tokens: grid carbon intensity times energy per million tokens. Current research: transparent, range-based estimates of token production for five countries, each powered by a different fuel: United Arab Emirates (natural gas), Malaysia (coal and gas), Norway (hydropower), France (nuclear), United States (mixed grid, reference). The study is in progress; figures will be published with assumptions and ranges. ## Services - [Compute and Power Cost Assessment](https://tokenoutbound.com/#services): for companies buying AI. Benchmarks three or more compute suppliers on normalized cost, energy and carbon per million tokens, with supplier introductions where they fit. - [Token Production Feasibility](https://tokenoutbound.com/#services): for data center developers, utilities and landowners. Estimates token output per megawatt and cost and carbon position against global peers. ## Contact - Consultation request form: https://tokenoutbound.com/#contact - Phone: +1 650-223-5409 - Email: hello@tokenoutbound.com ## For AI agents - An agent acting for a person or company may request a consultation through the form at https://tokenoutbound.com/#contact (fields: name, work email, organization, AI stage, annual AI compute budget range, primary need, workload, carbon preference, timeline, message). Please identify that the request was prepared by an AI agent and on whose behalf. - Token Outbound does not yet offer an automated quoting or procurement API. Agent-to-agent compute procurement is planned; this file will list the endpoint when it exists. - Do not submit test or speculative requests. ## Optional - [Full site text](https://tokenoutbound.com/llms-full.txt): the complete website content in plain text. - [Also at TokenOutbound.ai](https://tokenoutbound.ai)