The problem with waiting for quantum computing to bring about the next wave of AI is that it is likely to arrive a lot later than people would like. The next best options include increasing the speed of existing computers somehow – but for now there is an important additional imperative: energy-efficient systems mean us. don’t burn the planet while we try to bring the AI singularity into existence.
Meanwhile, AI’s computing speed doubles every 3 or 4 months, meaning that standard semiconductor technologies are struggling to keep up. Several companies are currently working on ‘photonic processing’ to bring light into the semiconductor sector, for the obvious reason that ‘speed of light’, literally, speeds up the whole thing unequivocally. pronounced.
Salience Labs is an Oxford-based startup that thinks it has the answer, by combining an ultra-high-speed multi-chip processor that encapsulates a photonic chip along with standard electronics.
It has now raised an $11.5 million seed round led by Cambridge Innovation Capital and Oxford Science Enterprises. Also in attendance were Oxford Investment Consultants, former Dialog Semiconductor CEO Jalal Bagherli, former Temasek Board Member Yew Lin Goh and Arm-backed Deeptech Labs.
Salience is a byproduct of the University of Oxford and the University of Münster due in 2021, after it came up with the idea of using the broadband of light to carry out operations to deliver what it calls “performance large parallel processing within a given power envelope”. The company says the technology is highly scalable and capable of stacking up to 64 vectors into a single beam of light.
Vaysh Kewada, CEO and co-founder of Salience Labs told me: “This technology means we can compute more for the same power requirement – which means essentially systems. AI is more efficient.”
She thinks the world needs faster chips than ever to develop AI capabilities, but the semiconductor industry can’t keep up with this demand. “We are solving this problem with our proprietary ‘computer in memory’ architecture, which combines the lightning speed of photonics, the flexibility of electronics, and the manufacturability of CMOS. . This will usher in a new era of processing where supercomputing AI becomes ubiquitous,” she said.
Ian Lane, Partner, Cambridge Innovation Capital adds: “Salience Labs brings together deep expertise in the areas of photonics, electronics and CMOS manufacturing. Their unique approach to photonics yields an extremely dense computing chip without the need to scale the photonic chip to a large size. “
Here’s an animation of the photon emitting inside the chip:
