Particle.news

Download on the App Store

Multiverse Computing Secures $215M to Shrink AI Model Sizes by 95%

The new funds will accelerate roll-out of its quantum-inspired compression service on cloud platforms alongside customer data centers.

AI (Artificial Intelligence) letters and robot hand are placed on computer motherboard in this illustration taken, June 23, 2023. REUTERS/Dado Ruvic/Illustration/File Photo
Businessman and computer technology network connection on futuristic virtual interface screen artificial intelligence
fundraising Spain ai research quantum physics artifical intelligence

Overview

  • The €189 million Series B was led by Bullhound Capital with backing from HP Tech Ventures, SETT, Forgepoint Capital International, CDP Venture Capital, Santander Climate VC, Toshiba and Capital Riesgo de Euskadi – Grupo SPR.
  • CompactifAI leverages quantum-inspired tensor networks to reduce large language model sizes by up to 95% without sacrificing accuracy.
  • Compressed models run 4×–12× faster than their uncompressed counterparts, translating into 50%–80% savings in inference costs.
  • Slimmed versions of open-source LLMs such as Meta’s Llama, DeepSeek and Mistral are available via AWS Marketplace alongside on-premises licensing options.
  • Multiverse holds 160 patents, serves more than 100 customers globally including Iberdrola, Bosch and the Bank of Canada, and was co-founded by physicist Román Orús.