📊 Full opportunity report: ALIA. The Spanish answer. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Spain’s ALIA project, with €240 million in public funding, has released a 40-billion-parameter multilingual language model. It aims to promote Spanish-language adoption and operational transparency, but benchmark results suggest it lags behind leading models like Llama 2.
Spain has officially launched ALIA, its largest public artificial intelligence project to date, featuring a 40-billion-parameter multilingual language model trained on over 9.37 trillion tokens. The $725 Billion Question: Hyperscaler Capex Q1 2026 and What the Earnings Don’t Answer The project aims to establish Spain as a leader in multilingual AI and promote the Spanish language within European AI ecosystems, marking a significant step in Europe’s sovereign AI efforts.
Developed by the Barcelona Supercomputing Center (BSC-CNS) and coordinated by the Spanish Secretary of State for Digitalisation and Artificial Intelligence (SEDIA), ALIA was released under the Apache License 2.0 on HuggingFace on April 22, 2025. It is trained on MareNostrum 5’s 4,480 NVIDIA H100 GPU partition, with a total public investment exceeding €240 million, including €90 million for hardware upgrades and €150 million dedicated to integrating ALIA into industry applications.
The model, named Salamandra-7B and Salamandra-2B, was trained from scratch on a combined total of approximately 12.875 trillion tokens across 35 European languages and 92 programming languages. It is designed to serve as Spain’s institutional answer to the European sovereign AI initiative, emphasizing multilingual coverage and transparency validated by AESIA. Despite its ambitious scope, benchmark testing indicates that ALIA’s performance—such as 51.77% accuracy on XNLI in English and 81.53% on SQuAD in English—lags behind models like Llama 2, which achieve 66% and 93-94%, respectively.
Official statements, including from Josep M. Martorell, highlight that the project’s strategic focus is on widespread adoption within the Spanish-speaking world rather than surpassing global performance leaders. This operational positioning aligns with the model’s structural emphasis on Spanish-language oversampling and multilingual coverage, aiming to foster regional AI sovereignty.
ALIA.
The Spanish
answer.
€240M+ Spanish public funding · ALIA-40B + Salamandra family · 9.37T tokens · 35 European languages + 92 programming languages · MareNostrum 5 · Apache 2.0 release. The largest publicly funded European national-AI project by cumulative scope — and the empirical test case for the Position 1 vs Position 3 strategic-positioning argument.
This is the tenth standalone essay in the European sovereign-LLM track and the third Tier 2 expansion piece. ALIA is Spain’s institutional answer — the largest EU member state by GDP not yet documented in the track. The project markets itself as Position 1 + Position 2 simultaneously — “Europe’s first public multilingual foundational model.” The benchmark evidence (ALIA-40B 51.77% XNLI_en vs Llama 2 66%) confirms the structural capability gap from Finding 1 of the synthesis essay. The Position 3 framing — Martorell’s “most widely adopted in the Spanish-speaking world” — is operationally honest. €90M MareNostrum 5 upgrade + €150M company integration = €240M+ cumulative scope. Apache 2.0 open-source release + AESIA validation + co-official languages oversampling. Both can be true at once. The Spanish public discourse would benefit from explicit Position 3 strategic positioning.
Six models. Apache 2.0.
The ALIA family operates as a tiered model portfolio. ALIA-40B is the flagship at 40 billion parameters; the Salamandra family scales down to 7B, 2B and instruct-tuned variants; mRoBERTa provides the foundational multilingual baseline. All released under Apache License 2.0 on April 22, 2025 at the HispanIA 2040 event — “Public Code, Public Money” approach.
multilingual
MN5 LLM
edge
target
instruct
encoder
Multilingual AI Translation Mastery: Building Accurate, Culturally Sensitive Language Tools and Global Communication Systems in 2026
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Four official. Oversampled by factor of 2.
ALIA’s distinctive multilingual coverage strategy. The four co-official Spanish languages are oversampled by factor of 2 in the training corpus — structurally distinct from Apertus’s broad 1,811-language coverage approach. The strategy targets deep coverage of Spanish co-official languages rather than maximum language breadth.
AI Translation Earbuds Real Time 164 Languages 80H Playtime Translator Ear Buds Audifonos Traductores Inglés Español Wireless Earphones Bluetooth AI Headphone for Travel Meeting Learning K08 Black
Supports 164 Languages Worldwide: Powered by cutting-edge AI translation technology, these translator earbuds real time support translation in…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
ALIA-40B vs Llama 2. 14-point gap.
The empirical evidence Finding 1 of the synthesis essay needed. ALIA-40B at 40 billion parameters with €240M+ public funding and 8+ months MareNostrum 5 training achieves performance below Llama 2 — a 2023 frontier model released approximately 18 months before ALIA-40B. The capability gap is real and consistent with six of seven prior national-project answers documented in the track.
AI model training hardware NVIDIA H100
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Two pilots. Public administration deployment.
The operational deployment targets that validate the Position 3 + Position 4 framing. Public administration deployment is the structurally credible Position 3 + Position 4 strategic positioning — captive demand from Spanish public institutions where Spanish-language specialization is operationally distinctive.
The work is real across the Spanish ALIA case. €240M+ public funding committed. 40B parameter from-scratch model trained on 9.37 trillion tokens. Salamandra family released under Apache 2.0. AESIA validation aligned with EU AI Act transparency standards. Two pilot applications shipped — Tax Agency chatbot and primary care medicine heart failure diagnosis. The Position 1 framing is operationally misleading. ALIA-40B performance below Llama 2 confirms the structural capability gap. The Position 3 framing is operationally honest — Spanish-speaking world adoption, co-official languages oversampling, public administration deployment. Both can be true at once. The Spanish public discourse would benefit from explicit Position 3 strategic positioning.
Beyond the Public Cloud: Architecting Private, Secure, and Sovereign AI for the European Enterprise
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications of ALIA for Europe’s AI Sovereignty
The ALIA project demonstrates Europe’s commitment to developing sovereign AI infrastructure, emphasizing multilingual capabilities and transparency. While its benchmark results indicate it is not yet competitive with leading commercial models like Llama 2, its strategic focus on Spanish and European languages positions it as a credible regional alternative. The project’s scale and public funding highlight Spain’s ambition to influence European AI policy and promote language diversity, although the performance gap raises questions about operational efficacy versus strategic branding. The emphasis on widespread adoption over top-tier performance signifies a shift towards regional AI sovereignty and operational transparency, which could influence future European AI initiatives.Background on Spain’s Public AI Investment and Strategic Goals
Spain’s ALIA initiative is part of a broader national and European effort to develop sovereign AI capabilities, with previous projects like Portugal’s AMÁLIA, Italy’s Minerva, and pan-European collaborations such as OpenEuroLLM and Mistral. The project is funded entirely through public sources, totaling over €240 million, making it the largest publicly funded European AI project by scope. Learn more about hyperscaler investments. It builds on Spain’s existing language technology programs, including AINA and ILENIA, and aims to position Spain as a regional leader in multilingual AI deployment.
Prior European initiatives have varied in scale and scope, with some focusing on enterprise applications and others on academic collaborations. ALIA’s training on MareNostrum 5’s advanced supercomputing infrastructure underscores Spain’s strategic intent to match European ambitions for sovereign AI, although benchmark data suggests a structural performance gap compared to commercial models like Llama 2. The project also aims to promote transparency and operational validation, aligning with AESIA standards.
“The goal is not to be the best-performing LLM in the world, but the most widely adopted in the Spanish-speaking world.”
— Josep M. Martorell
Operational Performance Versus Strategic Goals
While ALIA has been publicly released and benchmarked, its performance relative to leading models like Llama 2 indicates a structural gap. It remains unclear how the project will evolve to improve performance or whether its strategic focus on adoption and language coverage will offset current benchmark shortcomings. The long-term impact on Spain’s and Europe’s AI sovereignty efforts is still uncertain, as is the model’s actual adoption rate in industry and government sectors.
Next Steps for ALIA and European Sovereign AI Initiatives
Future developments will likely include ongoing benchmarking, performance improvements, and increased integration into Spanish industry and public sector applications. Monitoring the adoption rate and operational validation outcomes will be key to assessing ALIA’s success. Additionally, further European collaborations and funding allocations could expand the project’s scope, while its performance gap may drive efforts for technical enhancements or new models.
Key Questions
What is the main purpose of ALIA?
ALIA aims to serve as Spain’s institutional answer to European sovereign AI efforts, focusing on multilingual coverage, transparency, and regional adoption rather than outperforming global models in benchmark tests.
How does ALIA compare to models like Llama 2?
Benchmark results show ALIA’s performance is below Llama 2’s, with accuracy percentages around 51.77% on XNLI and 81.53% on SQuAD, indicating a structural performance gap.
What are the strategic implications of ALIA’s launch?
It signifies a regional focus on AI sovereignty, language diversity, and operational transparency, even if performance benchmarks lag behind leading commercial models.
What remains uncertain about ALIA’s future?
Its ability to improve performance, achieve widespread adoption, and influence European AI policy remains uncertain, as does its long-term operational impact.
Source: ThorstenMeyerAI.com
