At a recent AI summit in Mumbai, Sam Altman challenged prevailing narratives about artificial intelligence’s environmental impact, arguing that widely circulated claims about ChatGPT’s water and energy use lack context and, in some cases, factual basis. His remarks have reignited discussion about how society measures AI’s footprint and balances innovation with sustainability.

Debunking the Water Usage Claims

Altman directly addressed viral claims suggesting that a single ChatGPT query consumes roughly 17 gallons of water. He described such figures as inaccurate and based on outdated assumptions about data center cooling systems. Modern facilities, he noted, increasingly rely on advanced liquid cooling, outside air systems and optimized thermal management, significantly reducing water intensity compared with older evaporative cooling methods.

While rejecting exaggerated figures, Altman acknowledged that concerns about AI’s aggregate resource consumption remain legitimate. The environmental question, he suggested, should focus less on per-query sensationalism and more on total system-wide energy use.

The Energy Demands of AI Systems

AI infrastructure is undeniably energy intensive. Training large language models requires thousands of specialized processors operating continuously for extended periods. Inference, the process of generating responses in real time, adds another layer of ongoing computational demand across global server networks.

Estimates suggest AI-related workloads could account for 3 to 5 percent of global electricity consumption by 2030. However, efficiency gains have been significant. Modern AI chips deliver substantially more computational output per watt than previous generations, reflecting rapid hardware innovation.

The broader data center industry has also improved power usage effectiveness, reducing energy overhead associated with cooling and facility operations. Despite these gains, absolute electricity consumption continues to rise as AI adoption accelerates.

Read more: Toyota EV Production and Sustainability: Why Diversification Is Central to Its Strategy

Comparing AI and Human Energy Investment

Altman introduced a provocative comparison by framing AI efficiency relative to human development. He argued that human intelligence requires decades of biological growth, education and food consumption before reaching cognitive maturity. From this perspective, the energy investment in training AI models might appear competitive once spread across billions of uses.

This comparison is controversial, but it reflects a broader point: measuring AI’s efficiency depends heavily on the chosen benchmark. Evaluating a single query in isolation may obscure the cumulative amortization of training costs across global applications.

Renewable Energy as the Structural Solution

Regardless of measurement debates, Altman emphasized the need for cleaner energy systems. He called for faster adoption of nuclear, wind and solar power to support expanding digital infrastructure. Major technology firms are already committing to 100 percent renewable energy sourcing for cloud operations, and interest in next-generation nuclear technologies is growing as a stable, carbon-free power option for data centers.

The sustainability of AI, therefore, may depend less on limiting computational growth and more on transforming the underlying energy mix powering digital systems.

Transparency and Reporting Gaps

One of the persistent challenges in assessing AI’s environmental impact is limited disclosure. Technology companies are not legally required to publish detailed breakdowns of energy and water usage tied to specific AI workloads. As a result, independent researchers rely on estimates, creating room for both overstatement and underestimation.

Greater transparency could clarify debates and enable more evidence-based assessments of AI’s footprint.

Explore OneStop ESG Marketplace: AI (Artificial Intelligence)

Economic and Grid Implications

Beyond emissions, data center expansion affects electricity markets. Concentrated computing clusters can strain local grids and influence electricity pricing, prompting infrastructure upgrades that may impact broader communities. Policymakers are increasingly weighing how to balance economic development from digital infrastructure with grid resilience and affordability.

Toward Sustainable AI Development

The industry is actively pursuing strategies to reduce AI’s environmental intensity. These include improving algorithmic efficiency, designing specialized hardware for lower power consumption, situating data centers near renewable generation, and scheduling intensive tasks during periods of peak clean energy availability.

Lifecycle considerations are also gaining prominence, including emissions embedded in hardware manufacturing and end-of-life disposal.

Conclusion

The debate sparked by Sam Altman’s remarks illustrates the tension between rapid AI advancement and environmental accountability. While exaggerated claims can distort public understanding, legitimate concerns about aggregate energy use remain.

The long-term sustainability of AI will depend on continued efficiency improvements, cleaner energy infrastructure and greater transparency. As AI systems become more integrated into daily life and economic activity, aligning technological progress with ecological responsibility will remain central to the industry’s evolution.

Subscribe to our newsletter for more insights, case studies, and ESG intelligence.

Explore ESG Solutions on our marketplace - OneStop ESG Marketplace.

Keep abreast of the top ESG Events on OneStop ESG Events.

OneStop ESG Educate: Your go-to source for top ESG courses and training programs tailored to your needs.

Stay informed with the latest insights on OneStop ESG News.

Discover meaningful career opportunities on OneStop ESG Jobs.