Same goes for Multiverse Computing, a Spanish hard tech startup co-founded by Román Orús that has recently raised €189 million. That’s over $215 million of their Series B funding round! This latest round of investment pushes the company’s total funding to nearly $250 million. Multiverse Computing is located in San Sebastián, Spain. Through their pioneering technology, CompactifAI they’re changing the game on what AI can do for society.
Román Orús is an associated professor at the Donostia International Physics Center. That’s why he co-founded Multiverse Computing to get the most out of quantum computing and put it to work in real-world use cases. The startup’s CompactifAI, their signature product, leverages advanced quantum-computing inspired compression techniques. More importantly, it can drastically reduce the size of popular foundational large language models (LLMs) by up to 95%. Perhaps most surprisingly, this cut is made with no loss in model performance.
CompactifAI primarily targets the compression of deep learning models. As a result, they are much smaller and more efficient. This breakthrough enables the models to run effectively on personal computers, smartphones, and various devices including cars, drones, and Raspberry Pi systems. Madrid-based Quantum Computer multiverse computing optimizes these models suitable to push them to their speed. Consequently, they get performance that is four to twelve times faster than their uncompressed equivalents. This improvement in efficiency means that inference costs are lowered by 50% to 80%.
Along with the recent funding success, Multiverse Computing has been gearing up for the launch of the DeepSeek R1 model, coming soon. The company said it intends to release more open-source models in the near future. Their mission is to improve their factual AI models and grow their business in the rapidly expanding AI sector.
Multiverse Computing is developing breakthrough quantum algorithms that have the potential to revolutionize the AI industry. These innovations and improvements could lead to huge cost savings and increase accessibility. By enabling powerful AI models to operate on a wider range of devices, the startup aims to democratize access to advanced artificial intelligence capabilities.