For quantum software aimed at LLM leviathans, Multiverse raises $27 million

The most recent news comes from Multiverse Computing, a San Sebastian, Spain-based startup that says it has raised €25 million, or $27 million, in an equity funding round headed by Columbus Venture Partners.

The startup is valued at €100 million ($108 million), and the investment will be utilized primarily for two purposes. The startup intends to establish new initiatives to collaborate more closely with AI companies who are developing and running large language models, in addition to continuing to expand its current business with startups in sectors like manufacturing and finance.

CEO Enrique Lizaso-Olmos makes the same argument in both situations: "optimization."

Put another way, as computing gets more sophisticated, it can also get more expensive and occasionally be too complex to use regularly. In its pitch, Multiverse claims that its software platform, Singularity, can be used to execute and optimize complex modeling and prediction applications more effectively. Singularity is intended to be applicable across a wide range of industries, including banking, manufacturing, energy, cybersecurity, and defense.

In AI, the emphasis is primarily on using the platform to compress Large Language Models. A new product called CompactifAI focuses on the calculations that are made continuously when creating and querying LLMs in order to reduce noise and speed up work, which in turn increases reliability when producing results.

The program, according to the business, can condense LLMs "with quantum-inspired tensor networks" by more than 80% while maintaining accurate results. If verified, that might have significant effects on the way businesses purchase and utilize processors, resolving one of the major bottlenecks facing the sector at the moment.