Sparxity, a leader in sparse-native computing, announced a new funding round that brings total funding to date to $90 million.
The funding will accelerate the development of the Dresden, Germany-based company’s specialized chips for fine-grained dynamic sparsity, targeting up to 44 times faster AI inference in ultra-sparse models, bringing greater energy efficiency by computing only what matters.
The round was led by ICM HPQC Fund and joined by investors such as Lockheed Martin Ventures, GlobalFoundries, GS Futures, Cloudberry Ventures, TGFS, and the EIC Fund. The financing also includes angel investments from AI leaders Richard Socher (Recursive Superintelligence) and Nancy Xu (Salesforce-acquired Moonhub), alongside continued support from Atlantic and Silicon Catalyst.
“Our work with PEER demonstrated the potential of dynamic sparsity, activating only a fraction of the compute across millions of fine-grained experts,” said Xu Owen He, former Google DeepMind researcher and Sparxity strategic advisor, in a statement. “Yet today’s hardware limits how far we can take these approaches. Co-designing AI models, distillation pipelines, and hardware around sparsity unlock a new level of computing efficiency.”
Sparxity builds full-stack systems around a simple principle: compute only what matters. Its approach addresses rising AI infrastructure costs and energy consumption. The funding follows the sparse-native startup quadrupling its revenue between 2024 and 2025 with continued momentum in 2026.
Sparxity said it has created a robust value proposition around sparse-native computing with tremendous traction. To date, the company’s foundation include three tapeouts, more than 70,000 chips produced and 12 racks deployed.
Source link
