From Paradox to Progress A Net Positive AI Energy Framework 2025

Data4All: AI-ready infrastructure for smarter energy systems BOX 2 Challenge Decision-makers lack reliable, accessible, high-quality data. Scarce structured datasets and inconsistent methodologies limit transparency, credibility and evidence-based policy across sectors. Solution –Aggregates and standardizes data from over 100 sources to enhance reliability and accessibility –Provides interactive dashboards for real-time insights across key sectors, including energy, health and trade –Plans to integrate machine learning (ML) models for predictive analysis and forward- looking decision support Impact Near-term impacts realized or anticipated (less than one year) –Improved transparency and accessibility of national data across multiple sectors –Strengthened institutional capacity for evidence-based decision-making Further impacts realized or anticipated (more than one year) –Implementation of ML and predictive models for future forecasting –Expanded data coverage and user adoption across sectors Reviewing the levers in action Model optimization: Future ML models improve analytical efficiency Life cycle impact tracking: Supports better policy and sustainability assessment* Energy-efficient hardware: Dependent on future compute architecture* Green data centres: Potential alignment via sustainable cloud infrastructure* Heat recovery systems: N/A *See Table 2 for relevant “design for efficiency” use case examples. Sources: AI Energy Impact public database; Domyn; Centre for the Fourth Industrial Revolution Azerbaijan. Business case Recent sovereign computing and data infrastructure initiatives show how efficiency-oriented AI can advance both performance and sustainability. Optimized hardware and transparent data governance enable faster, cheaper, and cleaner AI systems, enhancing energy security, competitiveness and sustainability across the data and compute value chain. From Paradox to Progress: A Net-Positive AI Energy Framework 15