OpenAI President Greg Brockman Clarifies Misunderstanding About Data Center Water Usage, States Actual Consumption Is Much Lower Than Perception
OpenAI President Greg Brockman clarified misunderstandings about data center water usage on The Knowledge Project podcast, stating that the actual consumption is much lower than public perception.
He explained that data centers use a closed-loop cooling system, with a fixed and small amount of water, equivalent to the water used in a swimming pool's circulation, rather than continuously consuming large amounts of external water resources.
Brockman pointed out that public fear of AI mainly stems from uncertainty about how AI directly benefits individual lives, rather than issues with the technology itself.
Source: Public Information
ABAB AI Insight
Greg Brockman has repeatedly explained infrastructure details in public since becoming president of OpenAI. This response continues the official clarification path in the 2025-2026 AI data center controversy, previously addressing water and energy criticisms alongside Meta, Google, and others.
On the capital front, OpenAI and Microsoft are jointly directing significant Capex towards closed-loop cooling, efficient chips, and future data center tracks, motivated by the desire to reduce water resource and public relations risks, while alleviating regulatory and community resistance through technological transparency, paving the way for the continued expansion of superclusters like Colossus.
Similar to the local protests triggered by Meta and Google's early data center site selections, and the gradual clarification of "water fear" in nuclear power promotion, the current AI infrastructure is in a transitional phase where public understanding lags behind actual deployment. Early movers are rebuilding trust through direct communication from founders.
Essentially, this reflects regulatory changes: closed-loop cooling technology shifts pricing power from the traditional "high water consumption" narrative to efficient, sustainable infrastructure. The mechanism involves significantly reducing external dependencies through fixed circulation water use, forcing public discourse to shift from resource consumption to actual benefits, accelerating the structural transition of AI data centers from "objects of resistance" to "community-accepted assets."
ABAB News · Law of Cognition
The more efficient the technology, the more fear relies on old narratives rather than real data. The less the public can see personal benefits, the easier misunderstandings are amplified into reasons for opposition. Clarifying facts is faster than debating positions; transparency is the true lever for calming panic.