Case Study: Telefónica achieves 75% energy savings with Multiverse Computing’s AI model compression

Telefónica and Multiverse Computing develop an AI-based model to support customer service agents with 75% energy savings

Telefónica partnered with Multiverse Computing to enhance its customer service operations. The challenge was to implement large language models to support customer service agents, but doing so in a way that would reduce the significant energy consumption and high computing requirements typically associated with such AI models.

Using Multiverse Computing's CompactifAI compression technology, Telefónica compressed two open-source LLMs. This solution achieved an 80% reduction in model size and, most significantly, reduced the associated energy consumption by 75% while maintaining the quality of the AI's responses. By enabling local deployment on its renewable-energy-powered infrastructure, Multiverse Computing's solution also allowed Telefónica to drastically cut operational costs and CO₂ emissions.

Telefónica