Spanish startup Multiverse Computing has successfully raised €189 million in a Series B funding round aimed at enhancing its AI model compression technology, CompactifAI, which can reduce large language models (LLMs) by up to 95%.

This €189 million funding round is particularly noteworthy as European quantum startups typically attract a smaller share of global funding, marking a significant milestone for the region.

Multiverse's 'slim' models are available on Amazon Web Services, achieving speeds 4x-12x faster than their uncompressed counterparts, and can also be licensed for on-premises use.

CEO Enrique Lizaso Olmos emphasized that their focus is on compressing widely used open-source LLMs, particularly those adopted by corporations, such as the Llama family.

Multiverse claims its models can achieve 95% compression while retaining 97-98% accuracy, significantly outperforming traditional methods that often incur a 20-30% accuracy loss.

In the previous funding round at the end of 2024, Multiverse received an investment from CDP Venture Capital, which involved participation from prominent Italian corporations.

Key investors in this funding round include Bullhound Capital, Hewlett Packard Enterprise, Forgepoint Capital, and Toshiba, reflecting strong confidence in Multiverse's innovative approach.

Co-founded in 2019 by CEO Enrique Lizaso Olmos and CTO Román Orús, Multiverse seeks to bridge advanced quantum research with practical AI applications on classical hardware.

The company's compression technology integrates principles from quantum physics and machine learning, allowing it to simulate quantum systems without the need for a quantum computer.

As the AI inference market is projected to grow significantly from $106 billion in 2025 to $255 billion by 2030, Multiverse aims to capture substantial value by targeting key cost components in AI operations.

CompactifAI addresses the high operational costs associated with running LLMs, which typically require specialized cloud infrastructure, thus making AI more accessible.

The efficiency improvements offered by CompactifAI not only enhance accessibility for smaller organizations but also contribute to environmental sustainability by reducing AI's energy consumption.