Multiverse Computing Raises $215M for Quantum-Inspired AI Compression
Spanish AI company, Multiverse Computing, announced a significant Series B funding round totaling €189 million (approximately $215 million). This substantial investment fuels the development and deployment of their innovative CompactifAI technology.
CompactifAI: Shrinking LLMs Without Sacrificing Performance
CompactifAI is a groundbreaking technology inspired by quantum computing. It drastically reduces the size of large language models (LLMs) – up to 95% – without compromising performance. Multiverse focuses primarily on compressing open-source LLMs like Llama 2 and Mistral, offering "slim" versions accessible via AWS or on-premise licenses.
Significant Cost and Speed Improvements
The compressed models boast impressive speed improvements, ranging from 4x to 12x faster than their uncompressed counterparts. This translates to a substantial reduction in inference costs, estimated at 50% to 80%. For example, Multiverse claims its compressed Llama 4 Scout costs significantly less per million tokens on AWS compared to the original.
Expanding Accessibility and Applications
The efficiency of Multiverse's technology makes it possible to run these models on a wider range of devices, including PCs, smartphones, and even Raspberry Pis. This expansion opens up new possibilities for integrating LLM capabilities into various applications and industries.
Strong Leadership and Investor Backing
Multiverse's success is built upon a strong foundation of expertise. Co-founded by a leading expert in tensor networks and a former banking executive, the company boasts a remarkable team and a notable investor base including Bullhound Capital and other prominent venture capital firms. With a portfolio of over 100 clients, Multiverse is positioned for significant growth and impact.
This funding brings Multiverse's total funding to approximately $250 million, showcasing strong investor confidence in their vision.
Source: TechCrunch