Donostia, Spain and San Francisco – January 16, 2025 – Multiverse Computing and Kinesis Network Inc. have joined to provide what the companies said is an effort to optimize AI performance while reducing resource consumption. The partnership comes as demand for AI capabilities grows and organizations face challenges related to power demands for AI models.
Quantum and HPC company Multiverse has developed solutions for optimizing large language models to improve performance and efficiency. By using advanced quantum-inspired algorithms, Multiverse enhances the training and inference processes of LLMs, reducing computation time and energy requirements without compromising accuracy. The company’s compression software, CompactifAI, uses quantum-inspired tensor networks to improve the efficiency and performance of AI models. Shrinking these models reduces the compute power required to run them, which in turn reduces operating costs.
“We have found the perfect partner for our first significant U.S. partnership,” said Enrique Lizaso Olmos, CEO and co-founder of Multiverse Computing. “This collaboration will power new use cases for AI companies with models currently limited by power and compute requirements.”
Kinesis specializes in compute optimization, pooling distributed unused computing resources to maximize efficiency and minimize waste. Kinesis’ platform intelligently allocates workloads, ensuring that computational resources are used to their fullest potential. This dynamic resource optimization reduces infrastructure costs and aligns with global sustainability goals by decreasing carbon footprints.
Kinesis’s innovative compute platform has also captured the attention of leading-edge organizations such as nCorium, Rare Compute, and Liminal Capital. By leveraging Kinesis’s optimization technology, these forward-thinking organizations aim to advance their sustainability journey while enhancing computational efficiency and reducing costs—without compromising performance. This growing interest from diverse industries highlights the versatility and transformative potential of Kinesis’s approach to compute access and optimization.
Together, Multiverse Computing and Kinesis are solving a pressing industry challenge: harnessing the power of AI while mitigating the environmental and financial costs associated with significant computational demands. By integrating their respective expertise in LLM and compute optimization, the two companies offer a comprehensive solution for customers seeking to drive innovation efficiently and sustainably.
“AI workloads are pushing the limits of what current infrastructure can handle,” said Victor Gasper, Chief Sales Officer of Multiverse Computing. “Our partnership with Kinesis allows us to offer a comprehensive approach to optimization, addressing both the performance of AI models and the utilization of compute resources. This collaboration ensures our customers can innovate without compromising on cost or sustainability.”
Multiverse Computing recently completed the AWS Gen AI Accelerator program which included AWS credits, mentorship and resources to expand AI and machine learning (ML) technologies. Kinesis is an official AWS partner.
“The synergy between Kinesis and Multiverse Computing is a game-changer for the AI industry,” said Bina Khimani, Chief Product and Revenue Officer of Kinesis. “By tackling inefficiencies in both model performance and computational resource allocation, we’re enabling enterprises to achieve more with less. It’s a powerful step forward for both business and the planet.”
This partnership comes at a critical time when industries face increasing scrutiny over the environmental impact of AI technologies. By reducing energy consumption and optimizing resource usage, Multiverse Computing and Kinesis are driving operational efficiencies and setting a benchmark for sustainable AI practices.
Together, Multiverse Computing and Kinesis are charting a course toward a more efficient and sustainable AI future, proving that innovation doesn’t have to come at the cost of the planet.