10 Feb 2025
UK-German startup targets power-efficient connectivity for AI data center infrastructure.
Salience Labs, the startup company spun out of silicon photonics research at the University of Oxford in the UK and the University of Münster in Germany, says it has raised another $30 million in a series A round of venture finance.
The company, which was founded in 2021 and now has locations in Oxford, Heidelberg, and San Jose, is targeting artificial intelligence (AI) data centers with optical switch technology that promises dramatic savings in power consumption.
Leading the latest round of investment were Applied Ventures - the venture wing of semiconductor industry lynchpin Applied Materials - and the ICM HPQC Fund, a Singapore-based equity group. UK-based venture groups Braavos and existing investors Oxford Sciences Enterprises and Cambridge Innovation Capital also took part.
Salience says that the cash injection will go towards further development of its optical switches for large-scale AI connectivity, with CEO and co-founder Vaysh Kewada adding:
“What our customers want is a photonic switch to connect their AI clusters that is compatible with existing infrastructure while delivering high bandwidth, low latency and significant power savings.
“The completion of this round will further our development and help us bring our product to customers to enable not just the savings, but large cluster connectivity.”
‘Leveraging light’
Having previously raised $11.5 million in May 2022, Salience is one of a growing number of photonics startups targeting AI data center infrastructure.
While hundreds of billions of dollars are being invested in the build-out of data centers that use high-speed optical transceivers, the current approach remains relatively energy-inefficient and has even seen the likes of Amazon and Google turn to nuclear power as a potential solution.
Another approach to reducing those power requirements is to replace electrical connections with optical links deeper within the computing architecture, with Salience saying that its switch enables all-optical networking between compute nodes.
The company has previously outlined utilizing chip-based optical frequency combs, soliton microcombs at microwave line rates, ultra-low-loss silicon nitride waveguides, and high-speed on-chip detectors and modulators as some of the key elements of its technology.
Other silicon photonics startups working to achieve similar results include the Columbia University spin-out Xscape Photonics, which last year raised $44 million, Nvidia-backed Ayar Labs, Google-affiliated Lightmatter, and Celestial AI.
Commenting on the Salience funding round the Applied Ventures global chief Anand Kamannavar said: “Silicon photonics is a promising technology to deliver significant advancements in energy-efficient performance for AI data centers.
“Salience's optical switch solution has the potential to enable a new generation of interconnect network architectures for faster and more efficient AI systems.”
William Jeffrey, who chairs the ICM HPQC Fund’s technical advisory committee, added: “Salience Labs' optical switches will enable unprecedented bandwidth and scalability for the next generation of AI and high-performance data centers.
“By leveraging light, they will unlock the full capability of modern servers while decreasing power consumption and cost.”
© 2025 SPIE Europe |
|