close
close

UCL spinout raises another £17m to reduce AI data center energy use | UCL News

UCL spinout raises another £17m to reduce AI data center energy use | UCL News

A UCL spin-out developing technology that could make data centers significantly more energy efficient, helping to reduce artificial intelligence’s growing demand for electricity, has received a huge funding boost.

Oriole Networks, data centers, energy, photonics

Oriole Networks has raised £16.8 million from London-based venture capital firm Plural, with all existing investors reinvesting. This follows a first round of funding of £10 million earlier this year.

The company, founded at UCL last year, is trying to tackle a huge global technology problem: the vast and ever-increasing amount of energy used in data processing in the AI ​​era.

Conventional networking systems, although many transmit data via light, must convert it back to electrical pulses for processing and then back to light for transmission. Oriole Networks uses photonics, an optical switching technology that enables complete processing in the form of light pulses rather than electronic ones.

The advantages are threefold: it is much faster (information can be transmitted at the speed of light without the need to convert it into electricity for processing); it uses much less electricity (photonics uses much less electricity to process data than electronics and photonic signals do not produce heat, so they do not require huge amounts of energy to cool the equipment) ; etcCrucially, increasing processing speeds means that AI applications such as large language models – which are driving the huge demand for data processing centers and energy supplies – can be trained up to ‘100 times faster.

James Regan, CEO of Oriole Networks, said: “This funding is another significant milestone for Oriole following a year of rapid pace and growth.

“This is a growing market that is desperate for solutions and our ambition is to create a photonic networking ecosystem that can reshape this industry.

“Building on decades of research, we are paving the way for faster, more efficient and more sustainable AI. »

Oriole Networks builds on years of research within UCL’s optical networking laboratories, with technology licensed through UCL’s technology transfer company, UCL Business.

Robert Thompson, Deputy Dean (Business) of the Faculty of Engineering at UCL, said: “We champion the power of fundamental research to tackle major societal challenges.

“Oriole Networks is an example of this, with decades of optical networking research at UCL now providing innovative solutions to meet society’s growing data needs while minimizing environmental impact. »

Marina Santilli, Associate Director of Physical Sciences and Engineering, UCLB, said: “The Oriole team was already moving at pace in translating UCL research into commercial hardware following their seed round earlier this year , so this new funding provides this timeline. an extra boost.

“The sooner Oriole can realize its vision of an all-optical, energy-efficient data center, the better for the planet. »

Ian Hogarth, Partner at Plural, said: “Applying 20 years of extensive research and learning in photonics to create better AI infrastructure demonstrates how there is still much innovation ahead to help reap the benefits of this technology.

“The team behind Oriole Networks has a proven track record of both building businesses and bringing deep science to commercialization and is creating a fundamental shift in the design of next-generation networked systems that will reduce latency and reduce the energy impact of the data centers we now rely on. .”

Latency is the delay before a data transfer begins following an instruction for its transfer.

In conventional data centers, computers communicate with each other via Ethernet cables like those connected to Internet routers.

These cables cannot carry data at the speed that computers can produce it, creating a bottleneck. They are also energy inefficient, accounting for between a fifth and half of the total data center power consumption.

Mr Regan said their technology reduced the electricity demand linked to data transmission by 97%, allowing some data centers to reduce their overall electricity consumption by almost half.

Organizations that have reinvested in Oriole include UCL Technology Fund, XTX Ventures, Clean Growth Fund and Dorilton Ventures.

Links

Picture

  • Second from left: Professor George Zervas, Chief Technology Officer at Oriole Networks and Professor of Optical Networking Systems in the UCL Department of Electronic and Electrical Engineering; center: James Regan, CEO of Oriole Networks.

Media contact

Nick Hodgson

  • E: nick.hodgson (at) ucl.ac.uk
  • Mobile: 07769 240209