UK Startup Oriole Networks Raises £10m to Revolutionize LLM Training

Oriole Networks, a UK-based startup, has recently secured £10 million in seed funding to accelerate the training of Large Language Models (LLMs) using an innovative approach involving light technology. The funding round was co-led by prominent investors including University College London (UCL) Technology Fund, Clean Growth Fund, XTX Ventures, and Dorilton Ventures, with additional support from Innovate UK Investor Partnership.

The startup, founded in 2023 by a team of four visionaries comprising professor George Zervas, James Regan, Alessandro Ottino, and Joshua Benjamin, aims to address one of AI’s most pressing challenges - energy consumption.

A visualization of Oriole Networks’ innovative approach

Oriole Networks has developed a groundbreaking method that utilizes light to interconnect thousands of AI chips, forming what they term a ‘super-brain.’ This super-brain has the capability to train LLMs a hundred times faster than conventional methods while consuming only a fraction of the energy.

According to Oriole Networks, this technological advancement has the potential to revolutionize time-sensitive tasks, drive increased adoption of AI, and expedite progress in machine learning algorithms within a sustainable, low-carbon environment.

In a statement, George Zervas, the CTO of Oriole Networks, emphasized the critical role of data centers in the AI landscape, highlighting the challenges posed by escalating computational demands and unsustainable power usage.

‘AI computational needs are increasing by 10 times every 18 months. Collective data movement across the servers in the data center becomes a bottleneck which in turn limits the training and inference completion time. This requires a fundamental shift in the co-design of next-generation networked systems.’

The article continues to discuss the significant growth projected for the AI market, with GlobalData forecasting a value of $909 billion by 2030, driven by a compound annual growth rate of 35% from 2022 to 2030. Specifically, within the GenAI sector, revenues are expected to surge from $1.8 billion in 2022 to $33 billion in 2027, representing an impressive compound annual growth rate of 80%.

Latest Developments

The article also touches upon recent developments in the tech sphere, including the milestone achievement of 900 million users on the privacy-focused messaging platform, Telegram. Additionally, the platform successfully raised $330 million in bond sales, underscoring its growing influence in the digital communication landscape.

Furthermore, a study by the Institute for Public Policy Research (IPPR) warns of a potential ‘AI job apocalypse’ in the UK, estimating that up to eight million jobs could be displaced by AI under current government policies.

The National Grid CEO issued a cautionary statement regarding the surge in data center power usage over the next decade, driven by advancements in AI and quantum computing. The CEO predicts a six-fold increase in power consumption, necessitating strategic planning to mitigate the environmental impact.

Conclusion

The groundbreaking initiatives undertaken by Oriole Networks signify a pivotal step towards enhancing the efficiency and sustainability of AI training processes. As the tech industry continues to evolve rapidly, innovations like those pioneered by Oriole Networks are poised to shape the future of AI development and deployment.