Spanish boot MultiVerse Computing on Thursday said It has increased a huge round of B series of 189m euros (about $ 215 million) with the power of a technology called “compacttifai”.
Compactifai is a quantum -inspired compression technology capable of reducing the size of LLMS by up to 95% without affecting the performance of the model, the company said.
Specifically, MultiVerse offers compressed versions of known, LLMS open source-mainly small models-such as the Llama 4 Scout, the Llama 3.3 70B, the Llama 3.1 8B and the Mistral Small 3.1. The company plans to release a version of Deepseek R1 shortly and said it works on more open source and reasoning models. Professional models from Openai and others are not supported.
Slim models, as it calls them, as it calls them, are available through Web Amazon services, or may be licensed for use to the level. The company says its models are 4x to 12x faster than comparable non -compressed versions, which translate to a 50% to 80% reduction in costs. For example, multiiverse she says The Lama 4 Scout Scout costs 10 cents per million AWS brands compared to the 14 cents of Lama 4 Scout.
MultiVerse claims that some of its models can become so small and energy efficient that they can be run on computers, phones, cars, aircraft and even DIY-Enthusiast’s favorite PC, Raspberry Pi. (Suddenly we imagine these fantastic Raspberry Pi Christmas Houses upgraded with interactive Santas powered by LLM.)
MultiVerse has some technical power behind it. Co -founded by CTO Román Orús, a professor at the Donostia International Center in San Sebastián, Spain. Orús is known for his own Innovative work on Tensor networks (Not to be confused with all the projects associated with the AE called Tensor on Google).
Tensor networks are computational tools that mimic quantum computers, but run on normal computers. One of their primary uses these days is the compression of deep learning models.
The other co -founder of MultiVerse, the Managing Director of Enrique Lizaso Olmos, also holds multiple mathematical grades and is a college professor. He spent most of his career in banking and is best known as former UNNIM Bank chief executive.
Series B was led by Bullhound Capital, which supported companies such as Spotify, Revolut, Deliveryhero, Avito and Discord. HP Tech Ventures, Sett, ForgePoint Capital International, CDP Venture Capital, Santander Climate VC, Toshiba and Capital Riesgo de Euskadi – Grupe Spr participated in the round.
MultiVerse says it has 160 patents and 100 clients worldwide, including Iberdrola, Bosch and the Bank of Canada. With this funding, it has increased about $ 250 million to date.
