Salience Labs wins $11.5m for chip that will power supercompute AI
Business Weekly article | By Tony Quested
A startup set to transform compute with a superchip that will fuel exponential advances in Artificial Intelligence has won $11.5 million seed cash from Oxbridge investors.
Salience Labs was spun-out of the University of Oxford and the University of Münster in 2021 to commercialise an ultra high-speed multi-chip processor that packages a photonics chip together with standard electronics.
The technology is highly scalable – capable of stacking up to 64 vectors into a beam of light.
By using a broad bandwidth of light to execute operations, Salience Labs says it delivers massively parallel processing performance within a given power envelope.
Salience Labs uses a proprietary amplitude-based approach to photonics, resulting in dense computing chips clocking at 10’s of GHz. This, combined with massively parallel performance, will enable exascale compute in a wide array of new and existing AI processes and applications.
The company leverages multi-chip design, with the photonic processing mapping directly on top of the Static Random Access Memory. This novel ‘on-memory compute’ architecture is inherently faster and can be adapted to the application-specific requirements of different market verticals, making it ideal for realising AI use-cases in communications, robotics, vision systems, healthcare and other data workloads.
The technology has been designed from first principles for volume manufacture and is currently fabricating with production-level foundries using standard CMOS processes.
The seed round was led by Cambridge Innovation Capital and Oxford Science Enterprises. Oxford Investment Consultants, former CEO of Dialog Semiconductor Jalal Bagherli, Yew Lin Goh and Arm-backed Deeptech Labs also participated.
The rationale is clear: The speed of AI computation doubles every 3.4 months, outpacing what standard semiconductor technologies have left to offer.
At the same time, AI hardware is moving away from general purpose applications in response to market demands which are increasingly verticalised by use-case.
To accelerate exponential advances in AI across industries, a new paradigm for compute – one that is both faster and highly application-specific – is required. Step forward Salience Labs.
CEO and co-founder Vaysh Kewada said:
The world needs ever faster chips to grow AI capability but the semiconductor industry cannot keep pace with this demand. We’re solving this with our proprietary ‘on-memory compute’ architecture which combines the ultra-fast speed of photonics, the flexibility of electronics and the manufacturability of CMOS. This will usher in a new era of processing, where supercompute AI becomes ubiquitous.
Ian Lane, Partner, Cambridge Innovation Capital added:
Salience Labs brings together deep domain expertise in photonics, electronics and CMOS manufacture. Its unique approach to photonics delivers an exceedingly dense computing chip without having to scale the photonics chip to large sizes.
Miles Kirby, CEO at Deeptech Labs in Cambridge, also hailed the company as a global gamechanger and added that companies from the length and breadth of Europe were applying for a place on the organisation’s accelerator and a slice of the tech action in the cluster.
He said: “Salience Labs is exactly the kind of DeepTech company Deeptech Labs was set up to help accelerate and fund. Vaysh and the team are commercialising leading-edge research out of Oxford and Münster universities and their photonic multi-chip processors have the potential to solve the compute bottleneck currently hampering AI growth.
“It’s testament to the global power of Cambridge’s network that DeepTech companies of the calibre of Salience Labs from all over Europe are applying to Deeptech Labs’ accelerator.”