Salience Labs raises $125 million to manufacture optical switches for AI data centers

Salience Labs has secured $125 million to push its light-based networking chips toward volume production. The Oxford company announced the oversubscribed Series B on October 14, 2026, with Bullhound Capital leading. It plans to accelerate deployment of optical switches designed to connect thousands of GPUs across multiple racks in AI data centers.

Keeping data in light

Salience develops optical circuit switching, or OCS: technology that routes data through optical connections. Its design uses silicon photonics, which handles light on a chip. The company describes its switch architecture as all-optical and fully integrated, with the aim of linking large groups of AI processors without making the network a bottleneck.

The problem, as co-founder and CEO Vaysh Kewada explains in SiliconANGLE’s account, is the repeated conversion between optical and electrical signals in traditional networks. Kewada argues that connection bandwidth is falling behind computing capacity as AI clusters grow, creating delays and increasing power consumption. More processors do not resolve that constraint if their connections cannot keep pace.

Salience’s proposed answer is to route information optically on the chip rather than rely on those repeated conversions. The company’s case is that this can reduce latency—the delay in moving data—and power consumption. Those benefits are its technology pitch, not measured production results disclosed with this financing.

A demonstrated switch, a larger roadmap

The financing follows Salience’s live demonstration of a 64-port all-optical photonic switch at ECOC 2026. The company calls that demonstration an industry first for silicon-photonics OCS. Its next planned solutions have 128 and 256 ports, giving the development roadmap a concrete progression beyond the device it has already shown.

Salience says further development will extend networks within and across AI computing systems throughout data centers. The larger switch configurations remain development plans, not products released with the financing.

AI infrastructure is increasingly being designed at rack scale, where networking has a direct impact on system performance, efficiency and economics.

Patrick Rundell, Director, AMD Ventures

Building the production operation

The manufacturing plan includes continued engagement with Keysight Technologies and Tower Semiconductor to scale operations for high-volume output. Salience also intends to expand its Silicon Valley offices and open a second West Coast location. Its stated geographic strategy includes entering new markets and developing relationships with customers and partners.

The investor group includes AMD Ventures, Intel Capital, SBI Group, Hudson River Trading and the British Business Bank. Existing shareholders Applied Ventures, Oxford Science Enterprises, Cambridge Innovation Capital and Silicon Catalyst Ventures also participated. Bullhound Capital Principal Callum Stewart is joining Salience’s board as part of the round.

Salience was founded in 2021, drawing on more than a decade of research from the University of Oxford and the University of Münster.