A recent white paper from NTT DATA, a global leader in AI and digital technology services, emphasizes the urgent need to integrate sustainability into AI development and deployment to mitigate the growing environmental impact of the technology. Titled Sustainable AI for a Greener Tomorrow, the paper highlights how the increasing computational demands of AI—particularly in training large language models, running inference pipelines, and maintaining always-on services—are driving substantial energy consumption, water usage for data center cooling, e-waste generation, and rare earth mineral extraction for hardware. Experts predict that AI workloads could account for over 50% of global data center power usage by 2028. NTT DATA stresses that while these environmental consequences are significant, AI also offers opportunities to address them, such as optimizing energy grids, reducing emissions, modeling environmental risks, and enhancing water conservation.
The paper outlines key strategies for sustainable AI, urging organizations to prioritize efficiency alongside traditional performance metrics like speed and accuracy. It calls for measurable environmental impact metrics, including energy consumption, carbon emissions, and water footprint, and advocates for lifecycle-centric approaches that consider hardware production, deployment, maintenance, and disposal. Best practices recommended include using green software engineering, aligning AI workloads with renewable energy availability, leveraging modular and upgradable components to extend hardware lifespan, and employing responsible recycling and refurbishment methods. The report also underscores the importance of shared accountability, requiring collaboration among hardware manufacturers, data center operators, cloud providers, software developers, policymakers, investors, and consumers.
NTT DATA’s white paper warns that fragmented assessments and inconsistent metrics remain a barrier to meaningful progress, with many organizations focusing narrowly on energy or emissions without accounting for other critical factors. By implementing a comprehensive, end-to-end redesign of AI systems, organizations can reduce resource consumption while harnessing AI’s transformative potential. The report ultimately positions sustainable AI not just as a corporate responsibility, but as a strategic opportunity to create lasting value, strengthen organizational capabilities, and safeguard environmental resources for the future. The full white paper, along with insights on sustainability services, is available on NTT DATA’s website.