Spanish ‘Soonicorn’ Multiverse Computing Releases Free Compressed AI Model

Compressed AI Model: Multiverse's Free HyperNova 60B Launch

What if you could run a powerful AI model without the massive infrastructure costs? Multiverse Computing's newly released compressed AI model makes that possible. The Spanish startup has launched HyperNova 60B 2602, a free, open-weight model that delivers frontier-level performance at roughly half the size of its source. Developers and enterprises seeking efficient, sovereign AI solutions now have a compelling new option. This release signals a major shift toward accessible, high-performance AI that doesn't sacrifice capability for efficiency.

Credit: Multiverse Computing

What Is a Compressed AI Model and Why Does It Matter?

A compressed AI model retains the intelligence of a large language model while significantly reducing its computational footprint. This isn't about removing features—it's about smarter architecture. For businesses, this means lower cloud costs, faster inference times, and the ability to deploy advanced AI on more modest hardware. As AI adoption accelerates across industries, the demand for models that balance power with practicality has never been higher. Multiverse Computing's approach directly addresses this pain point, offering a path forward for teams constrained by budget or infrastructure.

How CompactifAI Technology Shrinks Large Language Models

At the heart of this breakthrough is CompactifAI, a proprietary compression method inspired by quantum computing principles. Rather than simply pruning weights, the technology restructures how information flows through the model. This allows HyperNova 60B to maintain high accuracy while using less memory and generating responses with lower latency. The result is a 32GB model derived from OpenAI's larger architecture, optimized for real-world deployment. Developers gain access to sophisticated capabilities without the typical overhead, accelerating experimentation and production rollout.

HyperNova 60B 2602: Performance Meets Efficiency

The latest iteration, HyperNova 60B 2602, introduces meaningful upgrades for modern development workflows. It now better supports tool calling and agentic coding—tasks where models interact with external systems or write and debug code autonomously. These features often drive up inference costs, making efficiency gains even more valuable. Early benchmarks suggest the compressed model performs competitively against larger counterparts, including notable European alternatives. For engineering teams, this translates to faster iteration cycles and more responsive AI assistants, all while keeping operational expenses in check.

Enterprise Adoption and Real-World Use Cases

Multiverse Computing isn't just targeting developers—it's building for the enterprise. The company already counts major organizations like Iberdrola, Bosch, and the Bank of Canada among its clients. These partnerships highlight a growing appetite for AI solutions that prioritize data sovereignty, cost control, and customization. A compressed model like HyperNova 60B can be fine-tuned on proprietary data and deployed within secure environments, addressing key concerns for regulated industries. As businesses move beyond AI experimentation, the ability to scale efficiently becomes a decisive competitive advantage.

Funding Buzz and the Road to Unicorn Status

Behind the technology, Multiverse Computing is making strategic moves to fuel its growth. Industry sources indicate the company is in active discussions for a new funding round potentially valuing it above €1.5 billion. While the startup hasn't confirmed specific figures, it acknowledges ongoing investor interest. This momentum reflects broader confidence in European AI innovation and the commercial viability of efficient model architectures. Achieving unicorn status would place Multiverse among a select group of AI startups reshaping the global landscape from outside Silicon Valley.

Why European AI Innovation Is Gaining Global Attention

Multiverse Computing's rise parallels a broader trend: European AI firms are carving out distinct niches focused on efficiency, transparency, and sovereign infrastructure. With regulations like the EU AI Act setting new standards, companies building compliant, high-performance tools are well-positioned for long-term success. HyperNova 60B exemplifies this strategy—delivering cutting-edge capability while aligning with regional values around data governance and sustainability. As global demand for responsible AI grows, these homegrown innovators are proving that scale isn't the only path to impact.
The release of a free, compressed AI model marks more than a technical milestone—it represents a philosophical shift in how advanced intelligence can be distributed and deployed. By prioritizing accessibility without compromising performance, Multiverse Computing invites a wider community of builders to participate in the next wave of AI innovation. For developers weighing cost against capability, HyperNova 60B 2602 offers a compelling proof point: powerful AI doesn't have to mean prohibitive overhead. As the ecosystem evolves, expect efficiency to become as celebrated as raw scale.
This development also underscores a critical truth for enterprise leaders: the future of AI adoption hinges on practicality. Models must integrate seamlessly into existing workflows, respect budgetary constraints, and deliver measurable value. Compression technology like CompactifAI isn't just a clever engineering feat—it's an enabler of real-world transformation. Organizations that recognize this early will be better equipped to harness AI's potential without getting bogged down by its traditional costs.
Looking ahead, Multiverse plans to open source additional compressed models throughout 2026, expanding support for diverse use cases and domains. This commitment to openness could accelerate community-driven improvements and foster greater trust in the technology. In an era where AI development often feels concentrated in a few hands, such transparency offers a refreshing counterbalance. It also aligns with growing developer preference for flexible, auditable tools that can be adapted to specific needs.
For now, the availability of HyperNova 60B 2602 gives teams a powerful new option to explore. Whether building intelligent agents, optimizing customer service workflows, or prototyping next-generation applications, developers can test frontier-grade performance with minimal setup. As adoption grows, the feedback loop between users and creators will likely drive further refinements, creating a virtuous cycle of improvement. In the race to make AI both powerful and practical, efficiency may well be the ultimate advantage.