Nvidia Says AI Data Centers Use Just 0.2% of US Daily Water

Share:
Audio Loading voice…
Nvidia Says AI Data Centers Use Just 0.2% of US Daily Water

Synopsis

Nvidia cited Manhattan Institute findings on June 22, 2026, claiming AI data centers consume just 0.2 percent of daily US water, and that the figure has fallen sharply due to new technology — a pushback against growing environmental scrutiny of the sector.

Key Takeaways

Chip giant Nvidia shared research from the Manhattan Institute on June 22, 2026 , via its official X account.
The research claims AI data centers use only 0.2 percent of daily water consumption in the United States .
Nvidia stated the figure has 'dramatically decreased' in recent years due to a new, unspecified technology.
The claim comes amid intense regulatory and public scrutiny of data center resource use, particularly in drought-prone US states.
The Manhattan Institute is a New York -based think tank; the precise methodology behind the statistic has not been independently verified.
State-level permitting decisions and industry water recycling disclosures will be key indicators to watch going forward.

Chip giant Nvidia on Monday, June 22, 2026, shared findings from the Manhattan Institute citing that AI data centers account for just 0.2 percent of daily water usage in the United States, pushing back on widespread concerns about the sector's environmental footprint.

Context

Nvidia's post states: 'Water usage has been a hot topic in the AI data center world, but the numbers may surprise you.' The company cited research from the Manhattan Institute — a New York-based public policy think tank — which puts data center water consumption at 0.2 percent of total daily water use in the United States. Nvidia added that this figure 'has dramatically decreased in the past few years due to a new' technology, though the specific innovation was not named in the visible text of the post.

The post included a video and a link, suggesting the full detail is contained in accompanying media. The corporate account did not attribute the claim to any individual executive.

Policy Backdrop

The rapid expansion of AI infrastructure since 2022 has placed data centers at the centre of environmental scrutiny across the United States. Regulators, environmental groups, and state governments — particularly in drought-prone regions — have raised questions about how much water large-scale computing facilities consume for cooling purposes.

Independent research organisations, including the Manhattan Institute, have produced analyses comparing data center resource use to agriculture, power generation, and municipal consumption. The broader debate has influenced state-level permitting decisions for new facilities.

Stakeholders and Impact

Data center operators — including hyperscale cloud providers that rely heavily on Nvidia GPUs for AI training and inference — stand to benefit from a narrative that frames their water footprint as smaller than popularly assumed. The findings, if validated, could ease regulatory pressure in states where water scarcity is a live political issue.

Environmental regulators and advocacy groups are likely to scrutinise the methodology behind the 0.2 percent figure, particularly given the pace of new data center construction. The identity of the 'new' cooling or water-recycling technology referenced in the post has not been independently confirmed.

What's Next

State-level permitting bodies are expected to weigh industry-backed research alongside independent environmental assessments as they evaluate proposals for new AI data centers. Forthcoming disclosures on water recycling metrics from major operators will be closely watched by both regulators and investors.

As Nvidia's hardware remains central to global AI infrastructure buildout, the company's public engagement on resource consumption signals a broader industry effort to shape the environmental policy conversation ahead of potential federal or state-level mandates.

Point of View

A figure like 0.2 percent, if it gains traction, could blunt the strongest environmental objections. However, the omission of the specific technology driving the reduction, and the unverified nature of the statistic, leaves the claim open to challenge from regulators and environmental groups. This episode reflects a broader pattern in which AI hardware leaders are increasingly entering the resource-policy debate directly, not just through lobbying but through public communications strategy.
NationPress
23 Jun 2026

Frequently Asked Questions

How much water do AI data centers use in the US?
According to research cited by Nvidia from the Manhattan Institute, AI data centers account for just 0.2 percent of daily water usage in the United States, and that share has fallen in recent years due to new technology.
Why is Nvidia talking about water usage?
Nvidia posted on X on June 22, 2026, to highlight that data center water consumption is lower than widely perceived, pushing back against environmental concerns tied to rapid AI infrastructure expansion.
What is the Manhattan Institute and why does it matter here?
The Manhattan Institute is a New York-based public policy think tank. Nvidia cited its research to argue that AI data centers have a smaller water footprint than the public debate suggests.
Are AI data centers bad for the environment?
The environmental impact of AI data centers — covering water, electricity, and land use — is actively debated. Nvidia's cited figure of 0.2 percent daily US water use is one data point, but independent verification and broader metrics are still being assessed by regulators and researchers.
What technology has reduced data center water usage?
Nvidia's post referenced a 'new' technology that has driven down data center water consumption but did not name it explicitly; the detail appears to be contained in the linked video and article.
Nation Press
The Trail

Connected Dots

Tracing the thread behind this story — newest first.

8 Dots
  1. Latest 1 week ago
  2. 1 week ago
  3. 2 weeks ago
  4. 2 weeks ago
  5. 3 weeks ago
  6. 3 weeks ago
  7. 6 months ago
  8. 9 months ago
Google Prefer NP
On Google