Frugal AI Hub launched at Cambridge Judge Business School
Frugal innovation, an area of research pioneered at Cambridge Judge, embodies doing more with less – and this applies to the computing, energy and complexity angles of artificial intelligence in order to make them more efficient and sustainable.
The new Frugal AI Hub seeks to boost accessibility for artificial intelligence in emerging markets and low-income countries.
“The principles of frugal AI apply to both high-income and lower-income countries,” says Serish Venkata Gandikota, Co-Director of the Frugal AI Hub. “For developed economies frugal AI can be practiced through lower-carbon infrastructure for AI, while in emerging markets and lower-income countries frugal innovation can help enable accessibility and inclusion for AI.”
A White Paper issued by the Frugal AI Hub, outlining its mission and principles, identified four interconnected pillars underlining its work: Resource Efficiency, Sustainability, Accessibility and Inclusion, and Impact and Scalability.
“Frugal innovation is not just about cost-cutting; it is about reimagining processes, products, and systems to create value under constraints,” says a Foreword to the White Paper by Jaideep Prabhu, Jawaharlal Nehru Professor of Indian Business and Enterprise at Cambridge Judge and Director of the Centre for India and Global Business, who has written two books on frugal innovation.
“He adds: Artificial Intelligence, a transformative force of our time, holds immense potential to amplify the impact of frugal innovation. However, as we integrate AI into our lives and systems, we must ensure that it aligns with the principles of frugality: accessibility, affordability, and sustainability. This is where frugal AI emerges as a groundbreaking concept – AI designed not just for efficiency but for equity.”
The hub and the United Nations International Computing Centre (UNICC) previously developed a three-level measurement framework that links total cost of ownership, return on investment and the UN Sustainable Development Goals, helping organisations evaluate AI portfolios across financial, technical and societal dimensions.
A follow-up White Paper was issued by the Hub in conjunction with the recent COP30 climate-change summit. Accelerating a frugal AI ecosystem says that the concept is a way to address an imbalance between AI’s technical progress and economic viability given the huge cost of developing AI infrastructure.
The document proposes a Frugal AI Technical Stack that combines compression with efficient hardware and open standards, aiming to achieve “substantial reductions in total cost of ownership (70-90 per cent) and energy use (up to 80 per cent) without compromising model utility”. It also defines a Socio-Economic Stack that links cost efficiency to social inclusion and equitable access to AI infrastructure.
“By lowering barriers to participation, Frugal AI extends innovation to communities, small enterprises and public institutions previously excluded by high compute costs,” says the White Paper.