Multiverse Computing, a deep-tech company known for bridging quantum computing and AI, has raised $21.5 million to scale its groundbreaking LLM compression technology. This Series A funding round is co-led by Columbus Venture Partners and Quantonation, with additional support from European Innovation Council (EIC) Fund and new investors Redstone Quantum Fund and NSV Wolf. The company, which operates across offices in Spain, France, Germany, Canada, and the U.S., plans to use the funding to accelerate go-to-market activities and expand its client base in manufacturing, finance, and defense. Multiverse’s core innovation lies in its ability to compress large language models (LLMs) using advanced tensor networks. These mathematical structures enable a drastic reduction in memory requirements without sacrificing accuracy or performance. In benchmark tests, the company’s tech compressed Meta’s LLaMA models by 100 to 1000 times while maintaining high accuracy. This could lead to significant advances in deploying powerful models in resource-constrained environments, such as on-device AI and edge computing. CEO Enrique Lizaso emphasized that the funding will be used to commercialize the company’s products further and support international expansion. He added, “We’re delivering massive reductions in inference time and memory needs—improvements that can make frontier AI available to broader markets.” Founded in 2019, Multiverse is a spin-off from the Donostia International Physics Center and has since grown to a team of over 100, including 25 PhDs. With this new round, Multiverse positions itself at the forefront of scalable AI innovation. For more details, visit Multiverse Computing.