A new generation of researchers is tackling one of artificial intelligence's most pressing environmental challenges: making the hardware that powers AI systems dramatically more energy-efficient and sustainable. At the University of Essex, PhD students supported through a scholarship program are exploring ways to reduce the carbon footprint of AI inference hardware, design adaptive computing systems for net-zero operations, and advance embodied biological computing.

Why Does AI Hardware Efficiency Matter Right Now?

As artificial intelligence becomes embedded in everyday life through chatbots, smart assistants, and digital tools, the energy demands of AI systems continue to grow. The hardware that runs these systems consumes significant power and generates electronic waste when devices need frequent replacement. Researchers at Essex are investigating how to extend the lifespan of AI hardware while reducing the emissions produced during manufacturing, potentially cutting both carbon output and electronic waste simultaneously.

The challenge is particularly urgent because AI deployment is accelerating globally. Data centers and computing infrastructure must become more efficient to support this growth without straining electrical grids or increasing carbon emissions. This is why universities and technology companies are investing heavily in research that bridges hardware design, energy efficiency, and environmental responsibility.

What Are Students Actually Researching?

The Dr. Jalal Bagherli PhD programme at Essex supports pioneering research across three interconnected areas designed to address major technological and environmental challenges:

• Embedded Hardware Systems for Net Zero: Research exploring how computing systems can become more energy-efficient and sustainable as demand for AI and digital technologies continues to grow, with a focus on adaptive hardware that adjusts power consumption based on workload demands.
• Embodied Biological Computing: Investigation into alternative computing paradigms that could potentially use biological processes to perform computation, offering fundamentally different approaches to energy efficiency compared to traditional silicon-based systems.
• Responsible AI: Work ensuring that artificial intelligence systems are developed and deployed in ways that consider ethical implications, environmental impact, and societal benefit alongside technical performance.

One of the five scholars in the program, Nelson Kimaro, is specifically focused on reducing the carbon footprint of hardware systems used in AI inference. His research explores ways to reduce emissions while extending the lifespan of AI hardware, which could help cut manufacturing emissions and electronic waste by reducing the need for frequent hardware replacement.

"I am very grateful to be part of this programme and for the support of Dr. Bagherli and the University of Essex in enabling this research. It has provided me with a valuable platform to develop my research skills and build professional connections. In the future, I hope to continue working towards solutions that make advanced technologies more environmentally responsible while still meeting the needs of society," said Nelson Kimaro.

Nelson Kimaro, PhD Student, University of Essex

Another scholar, Shengyang Huang, is investigating adaptive embedded hardware systems for net-zero operations. This research examines how computing systems can dynamically adjust their energy consumption and efficiency as demand for AI and digital technologies continues to grow.

"Being part of this programme has been a fantastic opportunity to develop my research and connect with others working towards net zero. I hope my work can contribute to creating smarter and greener hardware systems in the future," noted Shengyang Huang.

Shengyang Huang, PhD Student, University of Essex

How Can Universities Drive Green AI Innovation?

The University of Essex model demonstrates several practical approaches that academic institutions can use to accelerate research in sustainable computing:

• Dedicated Funding Programs: Establishing scholarship programs specifically focused on environmental and technological challenges creates sustained support for researchers working on long-term problems like hardware efficiency and net-zero computing systems.
• Industry Partnerships and Mentorship: Connecting PhD students with experienced technology entrepreneurs and industry leaders provides real-world context, professional networks, and practical insights that bridge academic research and commercial application.
• Interdisciplinary Research Focus: Supporting research across embedded systems, biological computing, and responsible AI ensures that sustainability considerations are integrated into multiple aspects of technology development rather than treated as an afterthought.
• Visiting Scholar Programs: Hosting accomplished alumni and technology pioneers on campus creates opportunities for knowledge exchange and helps students understand how their research might translate into real-world impact.

The program was established by Dr. Jalal Bagherli, a technology entrepreneur and University of Essex alumnus who built an internationally successful career developing computer chips used in smartphones, cars, and smart technology devices. During a recent campus visit, Bagherli met with the PhD students to discuss their progress and research directions.

"Visiting the University of Essex is always a special day in my annual calendar and this year's visit was no different. I enjoyed immensely spending time with the PhD students and the academic staff and learnt about tremendous progress they have made in a year in their learning and research for moving Embedded Hardware Systems towards Net Zero, progressing the state of the art in Embodied Biological computing and further embarking on the use of AI to aid humanitarian causes," said Dr. Jalal Bagherli.

Dr. Jalal Bagherli, Technology Entrepreneur and University of Essex Alumnus

What's the Broader Context for Green AI Research?

The work at Essex is part of a larger European movement toward sustainable computing and green technology innovation. Across the continent, research institutions and technology companies are exploring ways to make artificial intelligence and high-performance computing more environmentally responsible. This includes advances in liquid cooling systems for data centers, energy-efficient server architectures, and hardware designs that reduce power consumption without sacrificing performance.

The research being conducted by Essex's PhD students addresses a critical gap: while much attention focuses on the energy consumption of running AI systems, less emphasis has been placed on the environmental impact of manufacturing and maintaining the hardware itself. By extending hardware lifespan and reducing manufacturing emissions, these researchers are tackling a part of the AI sustainability puzzle that is often overlooked but potentially significant.

As AI continues to become more integrated into everyday applications, from healthcare diagnostics to climate modeling, the importance of developing energy-efficient and sustainable hardware systems will only increase. The work of students like Nelson Kimaro and Shengyang Huang represents an important investment in ensuring that the technological advances of the next decade do not come at an unacceptable environmental cost.