Meet Salience Labs: An Oxford startup that contributed $ 11.5 million to develop high-speed chips for AI applications

The computing speed of AI doubles every 3.4 months, ahead of what standard semiconductor technology has offered. Meanwhile, AI hardware is moving away from general-purpose applications in response to market demands that are increasingly verticalized by usage options. To accelerate the exponential progress in artificial intelligence in various fields, a new computing paradigm is now required – faster and more application-specific.

Bridging this widening gap between the demand for AI processing and the supply of the semiconductor industry, Oxford Salience Labs raised $ 11.5 million to develop a high-speed multi-chip processor that combines photonics and electronics to accelerate exponential progress in AI. Seed tour led by Cambridge Innovation Capital and Oxford Science Enterpriseswith Oxford Investment Advisersformer CEO of Dialog Semiconductor Jalal Bagherleyformer board member Temasek Yu Lin Goh and Arm-backed Deeptech Labs involved.

Salience Labs aims to commercialize an ultra-high-speed multi-chip processor that combines a photon chip with standard electronics.

The company uses a multi-chip design with photon processing display directly on top of static random access memory (SRAM). This “computing in memory” architecture is faster and adapted to the specific applications of different market verticals, making it ideal for implementing AI use options in communications, robotics, vision systems, healthcare and other data workloads.

Vaish Kewada, CEO and co-founder of Salience Labs said: “The world needs more and more fast chips to develop artificial intelligence capabilities, but the semiconductor industry can’t keep up with that demand. We address this supply and demand gap with our patented “in-memory computing” approach, which combines ultra-fast photonics, electronics flexibility and CMOS manufacturability to deliver ex-scale computing to Appendix AI. This will open a new era of processing when the AI ​​supercomputer becomes ubiquitous. ”

Ian Lane, partner, Cambridge Innovation Capital, said: “Salience Labs brings together deep experience in photonics, electronics and CMOS manufacturing. Their unique approach to photonics provides an extremely dense computing chip without the need to scale the photon chip to large sizes. ”