AI startup Oriole Networks has raised €11.6m in funding
London-based startup Oriole Networks has raised €11.6 million in investment. The company has launched a seed round to fund its revolutionary solution. Investors included UCL Technology Fund, Clean Growth Fund and Dorilton Ventures. The funding round was one of the largest for the UK in recent years. The funds raised will capitalise on the solution and bring the company profitability.
Oriole Networks’ technology uses AI ‘superbrains’ to accelerate LLM learning. The company provides a solution to reduce the energy required to train language models. Energy consumption is one of the most pressing issues in the development of artificial intelligence. Oriole Networks believes its approach will help improve performance and reduce energy consumption.
Technology overview
George Zervas, Alessandro Ottino and Joshua Benjamin founded the company in 2023. They are all scientists at University College London involved in AI developments. James Regan holds the position of CEO. The Oriole team has developed a unique solution:
1. Light and its properties are the basis of the technology.
2. A single system integrates thousands of artificial intelligence chips.
3. The processors are linked together to form a ‘super-brain’ that can train language models.
4. This approach delivers results 100 times faster and with a thousandth of the latency.
At the same time, the system uses a small fraction of the power of a traditional LLM learning process.
Potential applications
In addition to energy efficiency, this technology opens up new possibilities for implementing AI. For example, it will enable:
– expand the use of AI solutions in different domains;
– improve approaches to algorithmic trading;
– accelerate the adoption of machine learning products to reduce carbon emissions.
Oriole Networks’ technology may interest companies working with AI. First, it will be relevant to data centres, which play an important role in the development of SaaS. At the same time, data centre operation is dependent on energy consumption and problems with power supply instability.
According to George Zervas, the computing requirements of AI are growing rapidly. It is growing ten times every 18 months, requiring more processors. Existing learning systems have some limitations that are slowing progress. For example, moving data between servers in data centres slows learning and increases the time needed to deploy processes. Oriole Networks offers a solution that addresses these issues and is available to many organisations. Trials of the technology show excellent results. It has accelerated LLM training by 100 times and reduced energy consumption by 40 times. Oriole’s product will enable the development of AI in the sustainable economy that the civilised world is striving for.
