Spanish Startup Multiverse Computing Raises €189 Million in Series B Funding
Spanish startup Multiverse Computing is making waves in the tech world, announcing a hefty €189 million (approximately $215 million) in a Series B funding round. This substantial investment underscores the growing interest in advanced technology, particularly their innovative approach dubbed “CompactifAI.”
What is CompactifAI?
CompactifAI is a groundbreaking technology inspired by quantum computing. It focuses on model compression, achieving an impressive reduction in the size of large language models (LLMs) by up to 95% without sacrificing performance. This capability positions Multiverse as a key player in the competitive AI landscape, serving multiple industries with its efficient solutions.
Compressed Large Language Models
The core of Multiverse’s offering lies in the compressed versions of various well-known open-source LLMs. Currently, their lineup includes popular models such as Llama 4 Scout, Llama 3.3 70B, Llama 3.1 8B, and Mistral Small 3.1. The company plans to expand its offerings by releasing a version of DeepSeek R1 soon and is continuously working on new open-source and reasoning models. It’s important to note that proprietary models from giants like OpenAI are not included in their portfolio.
Performance and Cost Efficiency
Multiverse’s “slim” models provide a significant boost in performance, being available through Amazon Web Services (AWS) or via licensing for on-premise use. The startup claims their models are 4x to 12x faster than their non-compressed counterparts, resulting in a remarkable 50% to 80% reduction in inference costs. For example, the Llama 4 Scout Slim runs at just 10 cents per million tokens on AWS, compared to 14 cents for the standard Llama 4 Scout model, a clear indicator of the cost-effectiveness Multiverse brings to the table.
Innovative Applications
One of the standout features of Multiverse’s models is their small size and energy efficiency, enabling them to run on a range of devices, from PCs and smartphones to cars and drones. Imagine the potential applications of such technology: DIY enthusiasts could even integrate LLM-powered functions into Raspberry Pi projects. This versatility opens up new horizons for personal and professional use, making advanced AI accessible in everyday technology.
More Read
Founders and Technical Expertise
Multiverse’s journey is driven by its adept co-founders. CTO Román Orús is a respected professor at the Donostia International Physics Center in San Sebastián, Spain, noted for his pioneering innovations in tensor networks. These computational tools emulate quantum computing principles, primarily serving to compress deep learning models.
Meanwhile, CEO Enrique Lizaso Olmos has a rich academic background in mathematics and extensive experience in the banking sector. His expertise helps steer the company towards growth and innovation.
Strong Backing and Future Prospects
The Series B funding round was led by Bullhound Capital, known for supporting prominent names like Spotify and Discord. Other notable investors include HP Tech Ventures, SETT, and Toshiba, among others. With this fresh influx of capital, Multiverse has now raised about $250 million to date, signaling strong investor confidence in its technology and vision.
Multiverse Computing has already secured 160 patents and counts over 100 customers worldwide, including reputable organizations such as Iberdrola, Bosch, and the Bank of Canada. The combination of innovative technology and strong partnerships sets the stage for Multiverse to expand its footprint in the AI sector further.
Industry Impact
As the demand for more efficient AI solutions continues to rise, Multiverse Computing’s advancements in model compression could lead to significant shifts in how businesses utilize AI technologies. By reducing the costs and enhancing the efficiency of LLMs, Multiverse is not just keeping up with trends—it’s shaping the future of artificial intelligence.
Inspired by: Source
