Its development starts from the analysis of different existing architectures, with the aim of ensuring an efficient interaction among the various project components, adhering to well-known standards, and ensuring interoperability. In particular, the I-NERGY architecture takes in consideration the following reference architectures: COSMAG (Comprehensive Architecture for Smart Grid), BRIDGE reference architecture recommendations for a business process agnostic data exchange reference architecture in the energy domain, the International Data Spaces (IDS) initiative and related Reference Architecture Model (RAM) together with GAIA-X and FIWARE RA to ensure data sovereignty in the data supply chains from data acquisition to data analysis and use.

I-NERGY architecture includes the following layers:

• Data Service layer to enable the I-NERGY full interoperability allowing the integration, pre-processing, semantic annotation and streaming of heterogeneous data, guarantying traceability, provenance tracking and accountability of the I-NERGY data and external data source.
• The AI Trained models layer to facilitate the development, training, evaluation and serving of the project ML models to be easily used in new services.
• The I-NERGY Analytics Application layer to constitute the foundation of the I-NERGY integrated solution developing and validating the selected pilot cases/small-scale experiments, descriptive, predictive and prescriptive/anticipatory data analytics along a variety of significant applications, by making use of AI-based ML Suite. 
• The I-NERGY AI-on-Demand Interconnection layer to leverage and make use of the already existing services of the AI-on-Demand platform, as well as create novel services and useful assets that can be integrated in the AI-on-Demand platform. 
• All the layers are secured by the Security and Access Control layer to guarantee security mechanism within the I-NERGY ecosystem below the architecture building blocks.