Salience Labs raises $125 million Series B for optical AI switching

Salience Labs is betting that optical circuit switching can solve a growing bottleneck in AI data centers: connecting thousands of GPUs across multiple racks without sacrificing performance. The company's all-optical, fully integrated switch architecture is purpose-built for that scale-up job, and the fresh capital is meant to move it from development into high-volume manufacturing.

The investor list carries weight. AMD Ventures and Intel Capital both participated, and their representatives framed the investment around rack-scale AI infrastructure where networking directly affects system performance and economics. That suggests the technology is being watched by chipmakers whose roadmaps depend on efficient interconnects. Existing backers including Applied Ventures LLC, Oxford Science Enterprises, Cambridge Innovation Capital and Silicon Catalyst Ventures also followed on.

CEO Vaysh Kewada said the funding, alongside support from AMD Ventures and Intel Capital, will accelerate growth into high-volume production. Bullhound's Callum Stewart pointed to the leadership team and customer relationships as evidence of capability. What remains to be seen is how quickly the company can scale manufacturing and convert those relationships into deployed systems; no timeline is given.