Spanish company Multiverse Computing, specializing in developing neural network compression solutions, has closed a Series C funding round of up to $570 million. The startup's valuation has reached $1.7 billion, indicating high investor interest in artificial intelligence optimization technologies.

At the core of Multiverse Computing's product is an innovative technology for compressing AI models using tensor networks. The development allows reducing model size by up to 95% without significant loss of accuracy. This is critically important for deploying AI on devices with limited computing resources and for reducing infrastructure costs.

Key Investment Areas

The raised funds will be directed toward developing the CompactifAI platform, creating sovereign AI infrastructure for European clients, and scaling the business in the US, Asia, and the Middle East. The technology is expected to find applications in the financial sector, healthcare, and industry.

The tensor networks used by the company represent a mathematical framework that effectively approximates complex neural network architectures. Unlike traditional quantization or pruning methods, this approach maintains high accuracy even under extreme compression.

Expert Opinion

In my view, Multiverse Computing's success marks a paradigm shift in the AI industry. The market is gradually moving away from the race for model size in favor of efficiency. 95% compression without quality loss is not just an incremental improvement but a fundamental breakthrough that could make expensive computing resources unnecessary for many tasks. If the technology scales, we will see a sharp increase in the number of companies deploying AI on edge devices.