Multiverse Computing Launches Free Compressed Generative AI Model

Large language models face a significant challenge: their immense size. Spanish startup Multiverse Computing is tackling this problem by creating compressed models designed to bridge the gap between the capabilities of cutting-edge AI and what businesses can practically afford to implement.

Their key innovation is CompactifAI, a compression technology inspired by quantum computing principles, which the Basque company has used to optimize models from OpenAI. Starting today, developers can freely access an enhanced version of Multiverse's HyperNova 60B model on Hugging Face. The company further plans to open-source additional compressed models in 2026 to cater to a broader spectrum of applications.

Multiverse states that its models are significantly smaller yet retain nearly equivalent power and accuracy. At 32GB, the HyperNova 60B is approximately half the size of its progenitor, OpenAI's gpt-oss-120b, while also offering reduced memory consumption and lower latency. The latest iteration, HyperNova 60B 2602, now provides improved support for tool calling and agentic coding, areas where inference expenses are typically high.

Multiverse claims its HyperNova 60B outperforms competitors like Mistral Large 3, a model from French AI decacorn Mistral. Beyond this technological competition, the two European AI firms share several similarities.

Similar to Mistral, Multiverse has established an international presence with offices in the United States, Canada, and throughout Europe. Both companies also serve enterprise clients; Multiverse counts Iberdrola, Bosch, and the Bank of Canada among its customers.

Although Multiverse has not officially achieved unicorn status, reports indicate it is currently raising a new €500 million funding round at a valuation exceeding €1.5 billion. In a statement to TechCrunch, the company confirmed ongoing discussions with potential investors for a new round but stated it is too early to comment on specific valuation or funding size. Multiverse also declined to comment on reports that its annual recurring revenue (ARR) hit €100 million in January.

If verified, this revenue would still be a fraction of OpenAI's reported $20 billion ARR, but it places Multiverse relatively close to Mistral, whose ARR reportedly surpassed $400 million—growth partly fueled by demand for alternatives to U.S. tech. Accordingly, Multiverse positions itself in recent communications as a provider capable of "delivering sovereign solutions across the entire AI stack."

Save Up to $300 (30% Off) TechCrunch Founder Summit Tickets

Join 1,000+ founders and investors at TechCrunch Founder Summit 2026 for a full day dedicated to growth, execution, and real-world scaling. Gain insights from founders and investors who have defined the industry. Network with peers facing similar growth challenges. Leave with actionable tactics you can implement right away.

This offer ends March 13.

Save Up to $300 (30% Off) TechCrunch Founder Summit Tickets

Join 1,000+ founders and investors at TechCrunch Founder Summit 2026 for a full day dedicated to growth, execution, and real-world scaling. Gain insights from founders and investors who have defined the industry. Network with peers facing similar growth challenges. Leave with actionable tactics you can implement right away.

This offer ends March 13.

These geopolitical dynamics recently aided Multiverse in securing a partnership with the regional government of Aragón in northeastern Spain. The Spanish Agency for Technological Transformation (SETT) also contributed to the AI startup's $215 million Series B round last year. Since its founding, Multiverse has consistently received support from the Basque region, which may very soon celebrate its first homegrown unicorn company.