Europe · Neural-compression enabling software research
Embed2Scale neural-compression research
This project focuses on neural-compression enabling software research.
Embed2Scale researches neural embeddings for EO and weather-data federation, using ground HPC training and reusable compressed representations. DLR explicitly connects this research to AI analysis on constrained satellite hardware. NeuCo-Bench evaluates application quality. No selected onboard computer, flight deployment or measured orbital compression is established.
Martel Innovate / Embed2Scale consortium · Research and development
Scope: Enabling research · Named neural-compression research with explicit prospective onboard-AI relevance and no defined shared orbital platform.
Figures retain their disclosed units and scope. Launched hardware does not establish demonstrated performance; targets may change. Not disclosed means no figure in the reviewed sources.
PROJECT RELATIONSHIPS
Suppliers & project partners
German Aerospace Center (DLR)
Consortium participant
Program organization: Martel Innovate BV
Partnership
Named participant or associated partner in Embed2Scale research.
Research participation, not procurement. IBM Research, Zurich and Oxford are associated partners rather than EU grant signatories.
Announced Publication date not supplied · Evidence checked 2 Nov 2026
Complete research validation and exploitation work within the program ending December2026. No orbital mission date is disclosed.
Manual review 2 Nov 2026. Compared DLR onboard-AI rationale with the formal CORDIS ground-training/data-federation objectives. Captured research targets, consortium/adviser roles and EU contribution without treating this as an operating platform or allocating all funding to orbital work. Primary body last observed not available. Source monitoring: unavailable. Source collection and substantive profile review are separate checks. Plans and dates may change.
CORDIS records a USD 2,936,402 (EUR2,618,047.50) EU contribution to Embed2Scale, coordinated by Martel Innovate. The grant was signed on December11,2023 and runs January2024–December2026. It develops HPC-trained neural compressors, targeting up to1000-fold compression for EO and weather-data exchange across four application families.
DLR describes Embed2Scale neural embeddings as a way to make AI analysis feasible on constrained satellite hardware. Its July2025 account reports NeuCo-Bench and Earth2Vec activities and discussions with ICEYE and OroraTech advisers. The research also serves distributed ground data centers and does not establish an onboard implementation or flight result.