Multiverse. Credit: Instagram @joinmultiverse
Spanish startup Multiverse Computing has raised €189 million ($217 million) to scale its breakthrough AI compression technology, CompactifAI, which shrinks large language models (LLMs) by up to 95 per cent with no performance loss.
The San Sebastián-based firm announced the Series B funding on June 12, led by Bullhound Capital, with support from HP Inc, Forgepoint Capital, Toshiba, and several public-private European investors.
CEO Enrique Lizaso Olmos said:
“What started as a breakthrough in model compression quickly proved transformative – unlocking new efficiencies in AI deployment.” (Cited by MultiverseComputing.com).
The tool, inspired by quantum physics, can cut AI inference costs by up to 80 per cent and run models on smaller devices like phones, drones, and even Raspberry Pi.
Compressed versions of Llama, DeepSeek, and Mistral are already live, with more to follow. According to Bullhound’s Per Roman, Multiverse is helping drive “European sovereignty” in AI.
View all technology news.
