AI Disclosure

This page provides the transparency information required by Article 50 of the EU AI Act (Regulation (EU) 2024/1689), which applies from 2 August 2026. We publish the same information at any point of first interaction with our AI features and inside reports we deliver.

Noda generates preliminary grid risk reports with the assistance of large language models. The system analyses public grid data and project documents you upload, then produces a structured assessment covering:

• bankability indicators for the planned connection
• N-1 exposure of the relevant transmission lines
• capacity competition with nearby developments
• comparative analysis across the customer's pipeline

The public website also runs an AI chat assistant for general questions about Noda.

Noda reports do not control, dispatch, protect, or otherwise operate any electricity transmission or distribution equipment. They are advisory documents read by the customer's engineers, investors, and advisors before commercial decisions. On that basis they fall outside the Annex III "high-risk" classification of the AI Act and benefit from the carve-out for systems that perform a preparatory task to a human assessment (Article 6(3)). We keep an internal Article 6(3) self-assessment on file.

The system is probabilistic. Reports reflect data available at the time of generation, contain uncertainty, and may be wrong. Every numeric output ships with a confidence band and a data-completeness flag. A Noda engineer reviews each report before delivery.

Each report is reviewed and signed off by a named Noda engineer. The reviewer can override the system, refuse delivery, or request additional data from the customer. The reviewer's name and timestamp appear in the report metadata.

Before making any financial, contractual, or regulatory commitment, you must independently verify all material findings with the relevant DSO, TSO, or licensed grid engineer. Reports are not a substitute for a connection study, grid study, or engineering opinion of the DSO or TSO.

We use Anthropic and Mistral AI as upstream model providers. We pass through their acceptable-use rules to you. We do not allow the Service to be used to facilitate the destruction, disruption, or unauthorised access of critical infrastructure.

AI-generated text and images in reports are marked with machine-readable metadata where technically feasible, in line with Article 50(2) of the AI Act.

Noda does not make decisions producing legal effects on individuals or similarly significantly affecting them through purely automated means. Reports are an input to human decisions. You can request a human review or explanation at hello@noda.energy.

AI safety or transparency questions: hello@noda.energy.