Sign in
Rate your bot: A new initiative to rate AI models based on their energy efficiency is being introduced by researchers amid rising concerns over the environmental impact of energy guzzling tech, according to a study published in the scientific journal Nature last week. Inspired by the US Environmental Protection Agency’s Energy Star program, the AI Energy Star rating aims to reduce AI’s carbon footprint and has already begun benchmarking a variety of AI models, starting with a focus on open-source models available on platforms like Hugging Face.
REMEMBER- Data centers are energy guzzlers: Electricity consumption for data centers — facilities composed of networked computers, computing infrastructure, and storage systems — is expected to double by 2026, data by the International Energy Agency showed earlier this year. The growing demand for electricity is also evident in global tech giants’ emissions, with Google’s emissions surging nearly 50% in five years due to AI energy demand. US tech company Microsoft also reported a 30% increase in carbon emissions since 2020 driven by its AI investments.
How the rating system works: The proposed AI Energy Star rating would assess AI models based on their energy consumption, providing a transparent measure for both developers and users, according to the researchers. The goal is to encourage the adoption of more energy-efficient AI models, thereby reducing the overall carbon footprint of the technology. This system would allow users to make informed decisions when selecting AI models in a way that is similar to how consumers choose energy-efficient appliances.
Elsewhere, green data centers are popping up: Digital DEWA subsidiary Moro Hub, a global digital hub focused on cyber security, and US-based computer tech company NVIDIA announced a collaboration in February to establish a green AI data center in the UAE. Egypt signed an MoU with Incom, SIC Investment, and Record Digital Asset Ventures in January to build a green data center powered by 200 MW of renewables. The project — which will be made up of both solar and wind plants — will be implemented in two 100 MW phases.
And key players are investing in them: Abu Dhabi sovereign wealth fund Mubadala is investing in global hyperscale data center developer Yondr Group. It is backing Yondr as global demand for cloud services surges on the back of advancements in artificial intelligence, 5G technology and the internet of things. US tech giant Microsoft and UAE AI outfit G42 are also partnering to build a USD 1 bn geothermal-powered data center in Kenya. The plan is part of a comprehensive suite of digital investments in Kenya to accelerate digitalization and promote cloud computing and AI services in East Africa.
The MENA data center market is poised for further growth: The MENA region is poised to develop into one of the fastest growing markets for data centers in the coming years, with ballooning demand and so-far limited supply, according to a recent JLL report. The region’s growing digital transformation, including 5G and IoT, along with increased investment in submarine cables and fiber connectivity, is driving a surge in data demand. There are currently 73 colocation facilities across the UAE, KSA, Qatar, Oman, and Kuwait, according to the report.
