Ankündigung

Earthian tritt Nvidias Inception-Programm bei, nimmt Kapital bei 112 Mio. $ auf. Mehr erfahren

Zuletzt aktualisiert:

20. September 2026·10 min read

The Corporate Sustainability Reporting Directive (CSRD) and European Sustainability Reporting Standards (ESRS) require detailed sustainability disclosures integrated with financial filings for many EU and EU-listed undertakings. Double materiality—impact and financial perspectives—means reports must explain not only what happened, but what could plausibly happen under forward scenarios. Earthian provides inference-driven environmental, climate, and policy intelligence so CSRD narratives are specific, comparable, and audit-friendly.

Materiality assessment is the front door: topics must be justified, stakeholders considered, and metrics chosen with definitions that survive assurance. Earthian supports this by supplying forward-looking risk and opportunity narratives grounded in models that reason about physical climate, catastrophe, transition, and policy channels—reducing reliance on generic sector templates.

Data warehouses collect metrics; CSRD asks for analysis. Earthian bridges the gap with structured explanations—how metrics move when scenarios change, which assets or subsidiaries drive exposure tails, and where governance actions are tied to measurable targets. Outputs can align to ESRS climate, pollution, workforce, and governance themes where Earthian models apply.

CSRD owners sit in multiple functions. Earthian Hub orchestrates model outputs so finance, risk, and sustainability see the same scenario set—avoiding conflicting climate paragraphs in the annual report and the risk report. APIs allow metrics and narratives to flow into disclosure management tools.

Assurers increasingly test controls over sustainability information. Earthian outputs include explainability and versioning patterns familiar to model risk teams, helping CSRD preparers document how AI-supported sections were reviewed, challenged, and signed off.

Strengthen ESRS chapters with coordinated inference models: