Scaling AI Infrastructure: Why We Invested in Salience Labs
At Intel Capital, we’ve been investing in photonics for years, including optical links, interconnects, and the technologies that move data inside next-generation data centers. As AI infrastructure scales beyond single racks, networking is becoming a critical constraint on overall system performance.
We believe optical circuit switching (OCS) will play an increasingly important role in addressing this challenge. OCS can keep high-bandwidth, predictable traffic optical through the switching layer, reducing latency and power overhead, while complementing traditional packet switching.
That’s why we invested in Salience Labs.
Salience Labs has developed an all-optical, fully integrated silicon photonics switch architecture with proprietary optical amplification. Built on more than a decade of research at the University of Oxford and the University of Münster, the company’s technology is designed to provide high-speed, ultra-low-latency networking for AI data center infrastructure. This is critical as AI scales, ensuring it can power today's enterprise demands.
We have spent several years studying the OCS landscape, meeting with companies across the space and evaluating the different technical approaches. We have also known the Salience Labs team for five years and followed its progress from early technology development to increasingly capable systems. Most recently, the company demonstrated a live 64-port all-optical photonic switch at ECOC 2026, which Salience described as an industry first for silicon photonics OCS. The 64-port demonstration is an important proof point as AI systems scale to larger accelerator fabrics across multiple racks, increasing the need for networking that can deliver higher bandwidth with lower latency and power. The company plans to build on this milestone with future 128- and 256-port solutions.
The new funding will help Salience accelerate deployment of its OCS technology, scale volume manufacturing and expand its U.S. presence in Silicon Valley. Its architecture is purpose-built to connect thousands of GPUs across multiple racks and address the latency, performance and efficiency bottlenecks created by growing AI workloads.
This investment reflects a broader optical thesis we have been developing over several years. We are pleased to partner with the Salience Labs team as it scales manufacturing, expands its presence and works to bring high-performance optical connectivity to AI data center infrastructure.
