Multiverse Computing SL, a startup with technology that reduces the hardware footprint of artificial intelligence models, is reportedly raising new capital.

Sources told Bloomberg today the Spanish company is seeking €500 million, or about $594.4 million. The round is expected to value Multiverse at €1.5 billion.

The report comes about eight months after the software maker raised $215 million from a consortium that included Toshiba Corp., HP Tech Ventures and others. At least some of those investors will likely participate in the round that Multiverse is raising. Additionally, Bloomberg’s sources said several new backers are expected to join.

Multiverse’s flagship product is a platform called CompactifAI that reduces the amount of infrastructure needed to run AI models. According to the company, the software can halve training times and speed up inference by 25%. Multiverse also promises to reduce models’ storage footprint in the process.

An AI model considers multiple data points when making a decision. Some of those data points influence the processing workflow more than others. A model tasked with predicting a retailer’s annual revenue, for example, might give more weight to its historical sales numbers than the final performance of rivals.

The model components that determine the extent to which different data point influence a decision are called weights. They’re stored in a mathematical structure called a weight matrix. According to Multiverse, its CompactifAI platform reduces AI models’ hardware requirements by turning their weight matrices into tensor networks. Those are mathematical objects that are most commonly used to study quantum mechanical phenomena.

The process of turning weight matrices into tensor networks introduces errors into AI models. CompactifAI mitigates those errors by retraining a neural network after compressing it, a process that Multiverse calls healing. The task can be carried out relatively quickly and requires only a handful of graphics cards.

In a May 2024 paper, Multiverse detailed that its researchers used CompactifAI to compress the Llama 2 7B language model. They combined the platform with a model optimization method called quantization. Multiverse says that it managed to reduce Llama 2 7B’s memory footprint by 93% at the expense of a 3% drop in AI output accuracy.

The company offers CompactifAI alongside a platform called Singularity. It provides access to AI models that automate industry-specific use cases across several major verticals. One module can detect factory equipment malfunctions, while another helps financial professionals make investment decisions. There’s also a cybersecurity engine that detects malicious network traffic.

Multiverse says that its products are used by more than 100 organizations. The company’s installed base includes Allianz SE, Moody’s Corp. and several other large enterprises. Customers can install CompactifAI on their infrastructure or use an application programming interface to access popular open-source AI models that have been compressed in advance. 

Multiverse reportedly plans to close its funding round in the first half of the year. 

Image: Unsplash