ALIA: The Public AI Infrastructure for Spanish and Co-Official Languages

2025-02-06

This project, backed by the Secretary of State for Digitalization and Artificial Intelligence (SEDIA) and coordinated by the Barcelona Supercomputing Center–National Supercomputing Center (BSC-CNS), is a strategic step toward strengthening Europe's technological sovereignty and ensuring an AI ecosystem that is transparent, responsible, and inclusive.

Artificial Intelligence (AI) is transforming society, and its impact on language is undeniable. However, much of the development in generative AI and Natural Language Processing (NLP) has been dominated by English. To bridge this gap, Spain has launched ALIA, a pioneering initiative within the European Union that aims to provide a public and open infrastructure for developing AI models in Spanish and co-official languages (Catalan/Valencian, Basque, and Galician).

Why Public and Open?

ALIA is 100% publicly funded to serve the public interest. Its goal is to provide universal access to AI resources for citizens, public administrations, universities, and businesses, ensuring that AI innovation benefits all of society.

As an open initiative, ALIA stands out for its transparency and adherence to the EU AI Act, promoting technological reliability and social and economic inclusion. The ALIA model family is also verified by the Spanish Agency for AI Supervision (AESIA), ensuring compliance with the highest transparency standards.

Who Can Benefit?

ALIA is available to everyone, for everyone. It provides public supercomputing resources to advance AI research and innovation in Spanish and co-official languages.

Businesses

Companies across various industries—from tech startups to large enterprises—can leverage ALIA’s open AI models and datasets to develop new products, improve customer interactions, and enhance automation in Spanish, Catalan, Basque, and Galician. By removing language barriers in AI, ALIA helps businesses reach broader audiences and ensure linguistic inclusivity in their services.

Public Administration

ALIA provides government agencies with AI tools that enhance public services, document processing, and multilingual communication. This enables administrations to offer more efficient, accessible, and inclusive services to citizens while maintaining data sovereignty and compliance with EU regulations.

Citizens

By promoting the development of transparent and responsible AI, ALIA ensures that AI-driven tools—such as voice assistants, translation services, and automated customer support—work accurately and fairly for speakers of Spanish and co-official languages. This initiative fosters a more inclusive digital landscape where linguistic diversity is respected and represented.

AI Researchers

ALIA provides a rich ecosystem of datasets, methodologies, and high-performance computing resources, enabling AI researchers to develop state-of-the-art models for Spanish and co-official languages. The open nature of ALIA fosters collaboration and knowledge-sharing, accelerating breakthroughs in natural language processing (NLP), machine learning, and generative AI.

Developers

Developers can integrate ALIA’s language models, APIs, and multimodal AI tools into their applications, making it easier to build AI-powered solutions that understand and generate content in Spanish, Catalan, Basque, and Galician. Whether working on chatbots, voice recognition, or content automation, developers benefit from high-quality, publicly available AI resources without relying on proprietary models from major tech corporations.

The Power of Supercomputing

Training and deploying generative AI requires enormous computational power. The ALIA model family processes billions of words using MareNostrum 5, one of the most powerful supercomputers in the world, managed by the Barcelona Supercomputing Center.

MareNostrum 5, capable of 314,000 trillion calculations per second, is funded by the EU EuroHPC Joint Undertaking, along with contributions from Spain (through the Ministry of Science, Innovation, and Universities and the Government of Catalonia), Turkey, and Portugal.

Background

The ALIA project builds on Spain’s Plan for Language Technologies (2019), which laid the foundation for public AI infrastructure. Initiatives like AINA (led by the Government of Catalonia) and ILENIA (led by SEDIA) contributed to the development of AI resources in Spanish and co-official languages.

In 2024, the National AI Strategy identified ALIA as a key pillar for advancing AI in Spain, aligning with the EU’s Digital Decade Program, which aims to enhance Europe's digital transformation and technological sovereignty.

Leverage ALIA with Kobalt Languages

At Kobalt Languages, we help businesses, researchers, and developers unlock the full potential of ALIA’s AI models for Spanish and co-official languages. Whether you're looking to integrate NLP into your applications, optimize multilingual customer interactions, or explore cutting-edge AI use cases, our team can guide you through the possibilities.