- AMD has announced a £2bn, five-year investment in UK AI infrastructure, revealed at London Tech Week, spanning supercomputing, research partnerships, and a first-of-its-kind photonic networking deployment.
- As part of the investment, AMD is collaborating with UCL spinout Oriole Networks to deploy the world’s first large-scale AI system powered by a pure photonic network, as part of the UK government’s £50M ARIA Scaling Inference Lab.
- The deployment marks Oriole’s first commercial rollout — with wider industry rollout across multiple accelerator platforms targeted for 2027.
AMD, the US chip designer, has committed up to £2 billion over five years to accelerate AI innovation in the UK, announced today by CEO Lisa Su at London Tech Week.
The package spans national supercomputing infrastructure, university research partnerships, and a first-of-its-kind deployment with a London startup that is about to do something no commercial operator has done before: run a large-scale AI system where the network connecting the chips carries no electrons at all.
The photonic deployment is the fastest piece
AMD is collaborating with Oriole Networks, a UCL spinout founded in 2023 by Professor George Zervas, Alessandro Ottino, Joshua Benjamin, and James Regan, to deploy the world’s first large-scale AI system. It is powered by a pure photonic network, as part of the UK government’s ARIA Scaling Inference Lab, a £50 million national testbed designed to address the infrastructure bottlenecks that constrain AI performance.
The system pairs Oriole’s PRISM photonic networking platform with AMD Instinct GPUs and EPYC CPUs, in a collaboration that has been running for more than a year and is now moving from validation to full-scale deployment.
It is Oriole’s first commercial deployment.
“A year ago, we were proving the physics; today, we’re proving the business. Our collaboration with AMD has moved from concept to deployment to a system an order of magnitude larger, and the data proves this is already driving performance increases at pace,” says Regan.
“This is what it looks like when photonic networking stops being a research curiosity and starts being the foundation of how serious AI infrastructure gets built,” he adds.
The problem with electrons
Data centre networks have run on electronic switches since the beginning of data centres. Each switch introduces latency, generates heat, consumes power, and adds another component to an already complex supply chain.
Oriole’s platform removes electronic switches from the network core entirely, replacing them with nanosecond-scale optical circuit switching. Data moves as photons directly from chip to chip.
The company says this cuts core network power consumption by 81% and drops GPU idle time from 60% to below 1%. Less hardware in the loop also reduces dependence on the supply chain that underpins today’s networking equipment and cuts cooling and water requirements.
“Oriole’s AI backend networking with nanosecond optical circuit switching represents a fundamentally different way to connect accelerators at scale,” says Madhu Rangarajan, corporate vice president of Compute and Enterprise AI at AMD.
What comes after the lab
The company has raised approximately $35 million across two rounds in 2024, from Plural, UCL Technology Fund, Clean Growth Fund, XTX Ventures, and Dorilton Ventures.
The photonic chip sector is attracting significant capital: OpenLight raised $34M for silicon photonics interconnects, while Optalysys closed a £23M round for photonic computing for encrypted AI workloads.
Oriole’s differentiation is the full removal of electronics from the network core.
What the AMD collaboration provides is something most deep-tech startups spend years trying to obtain: a credible, large-scale deployment with a Tier 1 hardware partner before going to market. By the time the commercial rollout arrives, Oriole will be selling something that has already run.
