Oriole Networks

What is the mission?

Oriole Networks solves AI’s biggest challenges – speed, latency and sustainability. It uses light, via advanced photonics technology, to create networks of AI chips and combine their processing power. This means LLMs can be trained up to a 100x faster, whilst consuming a fraction of power and energy, allowing algorithms to run with much lower latency.

Who are the founders?

Founded in 2023 out of University College London (UCL), Oriole Networks brings together pioneering science in optical networks, with first-hand founder experience from a tech entrepreneur who has built major businesses in the sector. CEO James Regan, who built EFFECT Photonics, has put together an extremely experienced commercial team to bring the technology and unique IP developed over two decades at UCL by founding scientists Professor George Zervas, Alessandro Ottino, and Joshua Benjamin to a sector that is crying out for a solution to its sustainability problems.

Professor George Zervas is one of the world’s leading experts in optical networked systems, and he has spent his career pushing the limits of how we connect large networks with photonics. Their mission is to make a wholesale shift from a datacentre networked via electronic packet switching to one that is fully optical.

Why did Ian get involved?

I’m deeply motivated by the potential climate impact of Oriole. If society continues to scale these AI systems, we urgently need a way to reduce their carbon footprint.

As companies continue to train larger and larger AI models, the amount of data that needs to be moved between GPUs is growing extremely rapidly. This places huge strain on the networking technology that connects thousands or even hundreds of thousands of GPUs together, the ‘interconnect’. As much as 90% of the time training a large AI model is spent moving data around, which is an inefficient use of hugely expensive GPUs as well as a contributor to the substantial energy footprint of AI clusters.

With Oriole’s network, it will be possible to unlock up to a 10x gain in AI cluster performance – so you could train a model like GPT-4 10 times faster, or with fewer GPUs. As importantly, the proposed optical network consumes a fraction of the power compared to current AI networks.

I’m also keen to see a UK start-up play a real leadership role in the future of the AI compute stack.