Voice for Purpose Enters A New Phase with EuroHPC JU Support

Rome – October 1, 2025

Translated today presented the latest progress of the Voice for Purpose initiative at EuroHPC User Days 2025, the European event where researchers, enterprises, and institutions demonstrate how high-performance computing (HPC) accelerates scientific discovery and industrial innovation.

Launched in February 2023, Voice for Purpose is now the largest European initiative dedicated to providing natural-sounding voices to individuals with voice impairments. To date, more than 6,000 voice donors have contributed to the project, which has already improved the quality of life for hundreds of individuals affected by ALS, SMA, Rett syndrome, stroke, and other conditions. By combining advanced large language models (LLMs) with next-generation voice generation technologies, the initiative is scaling assistive communication to a European level.

After winning the LocWorld’s 17th Process Innovation Challenge in late 2024, Voice for Purpose has entered a new phase. Through the AI and Data-Intensive Applications Access Mode program, a EuroHPC Joint Undertaking program designed to provide industry, SMEs, startups, and public institutions with access to supercomputing resources for AI-driven research, Translated was awarded 200,000 hours of compute time on Cineca’s Leonardo supercomputer.

This access is being used to power research that advances voice technologies for people with speech impairments, with a focus on:

• Developing scalable, multilingual voice cloning that requires only 5–10 minutes of recordings instead of 2 hours.
• Optimizing LLM interactions for faster, more natural communication with less effort by the users.

By combining Europe’s HPC infrastructure with AI innovation, Voice for Purpose is pioneering a new frontier in assistive communication, empowering individuals and strengthening inclusion at scale.