Yet another startup wants to crack the LLM code, but this time with the help of light; optical pioneer Oriole Networks wants to train LLMs 100x faster with a fraction of the power
A few years ago, Israeli startup CogniFiber made headlines with Deeplight, a fiber optic cable that could “incorporate complex algorithms into the fiber itself before the signal reaches the terminal.” We warned at the time that this technology would not reach end users in the near future and that it was unlikely to appear on laptops or smartphones anytime soon.
However, eeNews embedded now reports on Oriole Networks, a British startup that uses light for a different purpose: to create efficient networks of AI chips.
The technology can reportedly train LLMs up to 100 times faster than conventional methods, while drastically reducing energy consumption. This research aims to reduce the growing energy consumption of data centers, driven by the rapid expansion of AI workloads and the increasing demand for high-performance computing. .
20 years of photonics research
“Our ambition is to create an ecosystem of photonic networks that can reshape this industry by solving current bottlenecks and enabling greater competition at the GPU layer. Building on decades of research, we are paving the way for faster, more efficient and more sustainable AI,” said James Regan, CEO of Oriole Networks.
The company’s roots lie in optical networking research at University College London (UCL) and Oriole’s unique IP is based on the work of founding scientists Professor George Zervas, Alessandro Ottino and Joshua Benjamin.
The startup has already attracted the attention of a number of investors eager to find solutions to AI’s increasing energy needs.
With plans to bring its early-stage products to market by 2025, Oriole Networks hopes to reshape the infrastructure of AI by making it faster, more energy efficient, and ultimately more sustainable.
eeNews embedded quotes Ian Hogarth, member of Oriole’s board of directors, and partner at Plural who led the latest funding round: “Applying 20 years of deep research and learning in photonics to create a better AI infrastructure, see how much more innovation will come to help capitalize on the benefits of this technology,” he said.
“The team behind Oriole Networks has proven experience in both building businesses and bringing deep science to commercialization, creating a fundamental shift in the design of next-generation network systems that will reduce latency and reduce the energy impact of data centers where we now depend on will decrease.”