Horizon Europe HORIZON-CL5-2027-02-D3-24 Large-Scale Validation and Upscaling of Generative AI Tools for a Next-Generation Digital Energy System

This Horizon Europe Innovation Action topic funds consortia that build data-sharing frameworks and train AI foundation models for energy-system use cases such as grid planning and operation, forecasting, congestion management, anomaly detection, fault diagnosis, predictive maintenance and flexibility management. Projects must develop effective, secure and transparent data-sharing strategies with governance that stays operational after the project ends, follow open-source practices where feasible, reach TRL 7-8, and join the BRIDGE initiative. Consortia should include data owners such as utilities, grid operators and asset owners. The topic has a EUR 30 million budget for about 3 projects of roughly EUR 10 million each at a 70% funding rate. It opened August 4, 2026, and the single-stage deadline is December 1, 2026.

This Horizon Europe Cluster 5 Research and Innovation Action funds improved understanding, seamless prediction, projection and attribution of climate-driven extreme events and their compound and cascading effects, supporting early warning and EU adaptation and preparedness strategies. Projects should use a hierarchy of models, including high-resolution Earth system models and digital twins, and the topic explicitly welcomes innovative AI-based approaches to predict, emulate and reveal unprecedented high-impact events. The single-stage topic opens November 17, 2026 and closes March 4, 2027, with EUR 18 million for about three projects.

The EuroHPC Joint Undertaking launched its AI Gigafactories call on 30 July 2026 as a joint procurement, not a conventional research grant, and that distinction shapes who should respond. It seeks consortia or Special Purpose Vehicles bringing together companies, public entities, investors and other partners to establish up to seven AI Gigafactories across Europe - facilities that combine massive computing power and state-of-the-art AI processors with the software and cloud technology stacks, high-bandwidth connectivity and energy-efficient data centres needed to deliver frontier AI training, fine-tuning and inference at scale. This is the tier above the 19 AI Factories and 13 AI Factory Antennas EuroHPC already oversees, which serve SMEs and startups with free customised support; the Gigafactories are aimed at frontier-scale sovereign capacity. The financing model is blended: EuroHPC expects total private investment of more than EUR 20,000,000,000 across the EU, with successful projects receiving varying levels of EU and national co-funding calibrated to the size and range of sovereign services offered. Per-facility EU contributions have deliberately not been published, so applicants must build their own capital stack and negotiate co-funding rather than apply against a fixed grant envelope. Facilities may be established in a single Member State or span several, and cross-border developments are explicitly welcomed. The reference is EUROHPC-2026-CEI-AIGF-01 and the deadline is 12 November 2026 at 16:00 CET. Realistically this is a call for infrastructure-grade consortia with credible access to billions in capital, energy siting and hyperscale operations experience - not for individual research groups or early-stage startups, who should instead route through the existing AI Factories network.

HMGCC Co-Creation and Dstl are seeking an edge device that processes video and audio from off-the-shelf smart glasses and stitches frames into high-fidelity 3D maps for reconnaissance. The minimum deliverable is a TRL 4 proof of concept. AI, computer vision and machine learning are listed technology themes, for example algorithms that compensate for head movement while stitching frames. Essential requirements are a report on where smart-glasses data flows and who can access it, good daytime urban imagery, and a camera-based 3D map. Desirable features include low-light imaging, audio fusion, testing across several manufacturers and a three-year manufacturer roadmap, and the stretch goal is fusing audio and visual data to handle reflections and solid walls. Redesigning or physically modifying the glasses is out of scope, and data handling must comply with GDPR and manufacturer terms. The total budget is up to £60,000, about US$80,000, excluding VAT, with on average two providers funded for a 12-week project. Questions are due 29 October, pitch day is 25 November, and kick-off is targeted for January 2027. Applications close on 12 November 2026 at 17:00 UK time.

HMGCC Co-Creation and Dstl are seeking proof-of-concept small robots that can move through confined built environments across at least two of water, land and air, at TRL 4 or above. Essential requirements are a payload of at least 100 g, a demonstration in a representative built-environment scenario, and a final report with a development roadmap. Desirable features include operation in all three domains, up to one hour of operation, autonomous navigation, operation with intermittent GNSS, travel through ducts or pipes, a volume under 0.01 cubic metres and a mass under 1 kg. Humans must keep responsibility for operational decisions, and the system must not depend on continuous public cloud access. Funding is up to £60,000, about US$80,000, excluding VAT, and on average two providers are funded per challenge. The call opened on 12 October 2026. Applicants are notified on 23 November, pitch day is 7 December, and kick-off is targeted for January 2027 for a project of about 10-12 weeks. Applications close on 12 November 2026 at 17:00 UK time.