• Source:JND
Add as a preferred

The obsession with AI technology in Silicon Valley is often accompanied by grandiose techno-optimism. This is perhaps most visible in the rhetoric of OpenAI’s Sam Altman, who continues to posture AI as a panacea for every crisis imaginable, including climate change, inequality, and even job displacement. He completely ignores the mounting environmental cost of running these systems.  

Deflecting With Deflated Numbers

In a blog published on Tuesday, Altman presented OpenAI’s energy and water consumption figures as drastically understated. He claims that responding to a single query on ChatGPT consumes: 

  • “0.34 watt-hours of energy — (the) what a high-efficiency lightbulb uses in a couple of minutes”

  • “0.000085 gallons of water — (one) roughly one-fifteenth of a teaspoon,” purportedly used for cooling the data center  

Altman offered no citations or methodology for these numbers. Requests for comment by Gizmodo went unanswered.

Let’s do some calculations. OpenAI claims that ChatGPT processes over 1 billion prompts daily from 300 million weekly active users. Even by Altman’s estimates, that translates to 85,000 gallons of water per day, which totals to over 31 million gallons per year. And this is an underestimation based on simplistic figures that ignore the complex image generation and multimodal tasks that require much more energy and cooling.

ALSO READ: Samsung Rolls Out One UI 8 Beta 2 For Galaxy S25 Series: New Features And Fixes Detailed

The Water Cost of AI: A Growing Concern

The impact of generative AI technology on the environment is not merely hypothetical; it has practical ramifications. After enhancing its partnership with OpenAI, Microsoft s ChatGPT on his data centers saw a considerable increase in water consumption. Microsoft’s ChatGPT water usage estimation shows that approximately 0.5 liters of water is used for every 10—50 queries. This is considered to be 8 million gallons per day. GPT-4 and GPT-4o are subsequent models that require far more energy and resources than their predecessors.

Although Microsoft's future-facing plans provide promise regarding the construction of “closed-loop” water-free data centers, such facilities are still in development and years away from scaling.

Cherry-Picking Data, Ignoring the Bigger Picture

Altman’s claims are notably vague on which version of ChatGPT his numbers reflect. Are these metrics from the lightweight free-tier model or the more advanced multimodal GPT-4o used by premium subscribers? There’s also no mention of AI-generated images or video—tasks that demand exponentially more processing power and energy.

He makes the further suggestion that future data centers for AI technologies will be “automated,” and their use will eventually be “the cost of electricity.” This is far-fetched. Even with automation, today’s systems still require enormous amounts of electricity and water, usually supplied by fossil fuel sources. 

False Promises, Real Consequences

Altman’s post ultimately reads like a manifesto of techno-solutionist deflection. He argues that the wealth generated by AI will make new policy ideas—like universal basic income (UBI)—suddenly viable. Yet OpenAI has never seriously advocated for UBI beyond vague gestures, and Altman himself has focused more energy lobbying against regulation than championing progressive solutions.

He casually acknowledges that “entire classes of jobs” will be destroyed by AI, but assures us the “world will be getting so much richer so quickly” that we won’t need to worry.

This isn’t innovation. It’s hubris.

ALSO READ: Midjourney In Legal Trouble As Disney, Universal File Copyright Lawsuit

Conclusion: Tech Can't Save Us From Tech

Altman seems to be convinced AI will eventually mitigate the issues it worsens—climate change, resource depletion, and economic inequality. Even if AI had the potential to assist us, the reality is that right now, it is not curbing environmental damage but rather accelerating it.

The future that Altman anticipates is underpinned by AI, but only if we ignore the stark reality that infrastructure powered by AI will decimate the planet's precious resources and further exacerbate the climate crisis. With unabated global warming and the need for increasingly thirsty AI infrastructure, spending and projections won't fix the problem. Urgent action in demanding transparency, regulation, and real accountability are imperative.

Until those goals are met, it seems that we should show more caution toward the ever-positive narratives from big tech, particularly when those slogans come couched in jargon and are supported by unproven claims.