Participated · Series A · Feb 2025

Salience Labs develops silicon photonics optical switches to connect AI datacenter clusters, targeting high bandwidth, low latency and significant power savings. Its switches are designed to be compatible with existing infrastructure and enable scalable, energy-efficient interconnect fabrics for large AI clusters. The company draws on over a decade of research from the University of Oxford and the University of Münster and was founded in 2021 in Oxford, England. Salience plans to further develop its optical switches, bring products to customers, and expand to the U.S. to serve key customers. It recently closed a $30 million Series A to accelerate development and go-to-market; no revenue or user metrics were disclosed in the article. Salience Labs develops an ultra-high-speed multi-chip processor that packages a photonics chip with standard electronics to accelerate AI computation. Its proprietary “on-memory compute” architecture combines the speed of photonics, the flexibility of electronics, and the manufacturability of CMOS. The company says its approach uses a broad bandwidth of light to execute operations and can stack up to 64 vectors into a beam, enabling highly scalable, densely parallel processing. Salience positions the technology as delivering far more calculation for the same power requirement, targeting more efficient AI systems. The company spun out of Oxford University and the University of Münster in 2021 and is based in Oxford. It recently raised funding to advance development and scaling of its photonics-enabled processors.