Spanish startup Multiverse Computing, specializing in developing technologies for compressing artificial intelligence, has completed a Series C funding round of up to $570 million. The company's valuation has reached $1.7 billion, highlighting strong investor interest in solutions that optimize the resource intensity of AI.
At the core of Multiverse Computing's technology are tensor networks — a mathematical framework that allows for radically reducing the size of large neural networks. According to the company, their solutions can reduce model size by up to 95% without significant loss of accuracy. This is critically important for deploying AI in environments with limited computing power, especially in the corporate sector and on edge devices.
CompactifAI Platform and Global Expansion
The raised funds will be directed toward developing the CompactifAI platform, which is already positioned as a key product for "sovereign AI infrastructure." Additionally, the startup plans to expand operations in the US, Asia, and the Middle East — regions with active demand for energy-efficient AI solutions.
Particular attention should be paid to the focus on "sovereign infrastructure": this indicates the growing need of states and large corporations for computing power independent of Western giants. Multiverse Computing essentially offers an alternative to expensive cloud solutions from Big Tech, making their technology strategically important.
Expert opinion: Reducing model size by 95% without loss of quality is a breakthrough that could change the economics of AI. If the technology scales, we will see a sharp increase in AI adoption in industry, fintech, and even crypto projects, where every watt counts. However, the question remains: how resistant are such models to attacks, and do they retain interpretability under heavy compression? Answers to this will only emerge after mass deployment.