By using 3D optical computing, we are enabling AI deployment at scale – with the fastest, most energy-efficient AI processor.

Watch our video

Our purpose

We are unlocking the full potential of next-generation AI with 3D optical computing that can compute at much faster speeds and at greater scales than ever before, all while consuming far less energy than traditional electronics.


Who are Lumai?

We are a spin-out from the University of Oxford, and our technology is based on many years of ground-breaking research. Our team of machine learning and optical computing experts are dedicated to meeting the demands of ever-increasing AI performance.

See leadership

Why optical computing for AI?

The demand for AI capacity in datacentres is growing at an incredible pace, however the amount of AI computing performance that can be delivered is being limited by both the amount of energy consumed and correspondingly the heat produced by silicon-based GPU AI accelerators. A paradigm shift is needed, and 3D optical AI computing is the answer.

Why optical computing?

Traditional computing methods simply cannot achieve the same level of performance scalability as 3D optical computing, as they are inherently constrained to the 2D geometry of silicon chips. In contrast, by using beams of light moving in the 3D volume of space, Lumai’s technology doesn’t suffer such limitations, with game-changing effects:

  • Ultra-low latency

    as quick as the speed of light

  • Revolutionary computing speed

    vs. digital electronics

  • Parallel processing

    delivering unprecedented scale

  • Far less energy consumed

    over traditional methods

Purpose-driven optical AI computing

Our innovative products will transform AI processing, empowering cloud computing providers and enterprises to deliver the next generation of AI at a far greater scale than ever before.

  • We are creating world-leading AI processing using 3D optics technology.

    Our team of experts are developing a revolution in AI processing by using the properties of light to perform matrix-vector multiplication – the operation at the heart of machine learning.

  • We are delivering computation speeds faster than anything else on the planet

    Our matrix-vector multiplier (MVM) is the world’s fastest, capable of a near-term speed limit of up to 1017 operations per second – which is 100x faster than the human brain and 1000x faster than traditional electronics.

  • We are developing the software and other tools needed to make developing on our technology fast and easy.

    By using open standards we are ensuring that AI developers can easily migrate models with just a few lines of code – putting ease of use at the heart of what we do.

Application in the real world

  • Data centres

    where increased speeds are vital.

  • Autonomous driving

    where environments are constantly changing and milliseconds count.

  • Robotics

    where sensors, information and processor speeds are paramount.

  • Logistics

    where real-time route planning can result in greater efficiencies.

  • Healthcare

    where complex AI and heavy computation are being used in drug discovery.

  • Cyber security

    where threats are ever-evolving and businesses need to be one step ahead.

  • Satellite imagery

    where accurate environmental and agricultural data is important.

  • Voice recognition

    where increased speed and accuracy will take it to another level.

  • Surveillance & defence

    where processing and decision-making are needed at scale.

Want to find out more?

Contact Lumai