Multiverse Computing Secures $217 Million for AI Model Compression

Multiverse Computing has raised $217 million to further develop its AI model compression technology, CompactifAI, reports Reuters. The funding round was led by Bullhound Capital, with participation from HP Inc, Toshiba, and other investors.

CompactifAI is designed to reduce the size of large language models (LLMs) by up to 95% while maintaining their performance. This technology leverages ideas from quantum physics and machine learning, allowing for significant cost reductions and increased efficiency. The compressed models can run on various platforms, including cloud services and edge devices like smartphones and drones.

The funding will enable Multiverse to expand the adoption of its technology, addressing the high costs associated with deploying LLMs. The company has already launched compressed versions of popular models such as Meta's Llama and plans to release more in the future.