AI model compression startup Multiverse raises $570M at $1.7B valuation
Multiverse Computing SL, a startup working on technology that compresses artificial intelligence models so that they run more efficiently on less hardware, announced Monday it raised $570 million in Series C funding.
The round was co-led by Forgepoint Capital International, BNPP SIVF, and Bullhound Capital. Santander Alternative Investments, Tikehau Capital, HP Inc., Orange Ventures, Scania Invest, NAventures and half a dozen other investors also committed capital to the funding round. The raise values the company at a $1.7 billion pre-money valuation, an almost five-fold step up from its $215 million Series B raise in June 2025.
Multiverse’s flagship product is CompactifAI, a compression technology that uses tensor networks, a mathematical framework from quantum physics, to efficiently shrink the size of AI models. The company claims that this breakthrough can reduce the hardware footprint of large language models up to 80% to 95% with minimal accuracy loss.
Smaller AI models run faster, consume less energy and can be deployed on less costly hardware than extremely large models. This opens up the possibility of running extremely powerful models on devices that otherwise would have to run tiny models, for example, consumer PCs, laptops, smartphones, and edge devices.
“The AI industry has accepted a false constraint for years — that powerful models require expensive infrastructure,” said Chief Executive Enrique Lizaso. “That constraint is gone. We have proven that AI can run at full performance on a smartphone, inside a sovereign data center, on a factory floor with no cloud connection.”
In many cases, edge devices, such as smartphones and computers attached to sensors, must offload high-powered AI computing into the cloud when the local AI is insufficient. Having a model run locally reduces latency, the time it takes a question to be sent out to the cloud and an answer returned, and it also keeps sensitive information on the device so that it is never seen by the outside world.
The company recently released details about its platform showing that its CompactifAI-compressed version of Meta Platform Inc.’s Llama 3.3 70B can run on an Intel Xeon 6 processor with extremely minimal accuracy loss. The company reduced the model size on disk by around 50%, dropping the size from 130 gigabytes to 65 gigabytes. Although it still requires significant amounts of RAM, at around 1 terabyte, it can run the LLM efficiently entirely on the central processing unit, without the need for an ultra-powerful graphics processing unit.
Multiverse’s platform covers the full spectrum of efficient AI deployment, including compressed models that can be run on devices and models optimized for the cloud and on-premises for higher efficiency. CompactifAI Router decides in real time if an AI workload can be run locally or must be sent to the cloud.
The company said it is currently also applying its technology to a software layer for next-generation AI factories that supplies model compression, graphics processing unit orchestration, AI deployment, compute and governance controls into a unified solution. According to the company, this will allow enterprise customers to isolate their entire AI stack within their own firewall without replacing existing infrastructure.
Multiverse will allow the entire AI and data flow to remain within enterprise hardware, including local and on-premises AI model inference, without the need to route through a hyperscaler.
The company’s customers and partners include Allianz, Bank of Canada, Bosch, Iberdrola, Indra, PwC, and Telefónica, covering industries from manufacturing to finance, energy to aerospace, cybersecurity, defense and life sciences.
The company said the new fundraise would fund the expansion of its stable of high-efficiency, smaller-footprint models. It will push forward with research and development on proprietary, cutting-edge algorithms for model compression. The company said it also intends to focus on strategic investments in AI gigafactory infrastructure and the software stack to support deployments, and establishing a stronger regional presence in key markets, including East Asia, Southeast Asia, the Middle East, Canada and the United States.
Image: SiliconANGLE/Microsoft Designer
Support our mission to keep content open and free by engaging with theCUBE community. Join theCUBE’s Alumni Trust Network, where technology leaders connect, share intelligence and create opportunities.
15M+ viewers of theCUBE videos, powering conversations across AI, cloud, cybersecurity and more** 11.4k+ theCUBE alumni**— Connect with more than 11,400 tech and business leaders shaping the future through a unique trusted-based network.
About SiliconANGLE Media
theCUBE AIand theCUBE SuperStudios — with flagship locations in Silicon Valley and the New York Stock Exchange — SiliconANGLE Media operates at the intersection of media, technology and AI.
Founded by tech visionaries John Furrier and Dave Vellante, SiliconANGLE Media has built a dynamic ecosystem of industry-leading digital media brands that reach 15+ million elite tech professionals. Our new proprietary theCUBE AI Video Cloud is breaking ground in audience interaction, leveraging theCUBEai.com neural network to help technology companies make data-driven decisions and stay at the forefront of industry conversations.