NTT DATA, a global leader in AI, digital business, and technology services, has released a new white paper underscoring the urgent need to embed sustainability into every stage of artificial intelligence development and deployment. As AI adoption accelerates worldwide, its environmental footprint is expanding rapidly — making sustainable AI not only an ethical imperative but also a strategic opportunity to build long-term value, resilience, and efficiency.
Titled Sustainable AI for a Greener Tomorrow, the paper outlines how surging computational requirements are driving steep increases in electricity use, particularly for training large language models, powering inference engines, and running always-on AI services. Researchers estimate that AI-related workloads could account for more than half of all data-center power consumption by 2028. Beyond energy use, AI contributes to water consumption for cooling infrastructure, rare-earth mineral extraction for hardware, and growing e-waste volumes.
“The resource consequences of AI’s rapid growth and adoption are daunting, but the technology also can empower innovative solutions to the environmental problems it creates,” said David Costa, Head of Sustainability Innovation Headquarters at NTT DATA. “AI can help optimize energy grids, model climate risks, improve water management, and drive emissions reductions. The key is to integrate sustainability principles into AI systems from the outset.”
Key Insights
Prioritize Green Outcomes Alongside Performance
Traditional metrics such as accuracy and latency are no longer sufficient. NTT DATA advocates for efficiency as a core design principle, ensuring sustainability goals are aligned with AI performance objectives from the beginning.
Measure and Verify Environmental Impact
Standardized metrics are critical. Frameworks like the AI Energy Score and Software Carbon Intensity (SCI) for AI offer structured ways to track energy use, carbon output, and water consumption — improving governance, compliance, and procurement decisions.
Adopt a Full Lifecycle Perspective
Sustainable AI must account for each phase of the technology lifecycle — from raw-material extraction and chip manufacturing to deployment, maintenance, and eventual disposal. Extending hardware life, optimizing cooling systems, and adopting circular-economy models are essential steps.
Enable Shared Responsibility Across the Ecosystem
Sustainable AI requires collaboration among hardware makers, data-center operators, cloud providers, policymakers, investors, and end-users. Systemic change depends on coordinated action across this value chain.
Many organizations still lack consistent metrics and actionable tactics, the report notes. To address this gap, it recommends strategies such as green software engineering practices, workload scheduling based on renewable-energy availability, use of remote GPU services, and designing for modular upgrades to reduce e-waste.
NTT DATA emphasizes that building sustainable AI is a complex journey, but one that will ultimately ensure the powerful technology delivers benefits while protecting the planet. To explore the full findings and download the white paper, visit NTT DATA’s website.
//Staff writer
