Amazon reports 2.5 billion gallons of datacenter water use in 2025
The company says it is far below rivals and 75% to its 2030 goal, as US backlash heats up.

Amazon says its datacenters used about 2.5 billion gallons of water last year across its entire global footprint for 2025. For decision-makers, the disclosure sets the pace for how hyperscalers defend AI-era cooling demand amid growing siting and water-pressure politics.
Amazon is putting numbers on the table. In a blog post, the company says its global datacenter footprint for 2025 used about 2.5 billion gallons of water. It also claims that number is far less than rival hyperscalers and positions the business as “75 percent of the way” to a “water positive” goal by 2030.
The 2.5 billion gallons figure covers Amazon’s entire global datacenter footprint for 2025, and Amazon says it is also relatively efficient on a per-energy basis, reporting water usage of 0.12 liters per kilowatt-hour (L/kWh) at its data facilities. Amazon then tries to win the comparison fight with disclosed consumption metrics: it claims Microsoft used 0.27 L/kWh during 2025, Meta’s consumption stood at 0.19 L/kWh in 2024, and Google was the thirstiest at 1.15 L/kWh during the same year. The Register has asked Microsoft, Meta, and Google to comment.
Why this matters right now is that datacenter water use is no longer a background detail in infrastructure debates. The disclosure lands amid growing pushback against datacenter construction in the US, where an Ipsos survey found most Americans do not want facilities built nearby. The objections include worries over electricity prices, eyesore buildings, and water-hungry operations. The pressure is not theoretical. A 2022 report found Google datacenters were consuming more than a quarter of all the water used in The Dalles, Oregon. Even if you only focus on one metro or one state, the takeaway is the same: water becomes a gating factor for where expansion can happen.
Amazon’s strategy here is classic hyperscaler comms, with a tech dashboard twist. It downplays the headline gallons by comparing them to lawn and garden watering in the US, citing that Americans used that same water volume for lawns and gardens over the same period (the source notes a country of 350 million people). It also frames “water positive” not as a vanity target, but as an operational plan: facilities return more water to the environment than they consume through measures including rainwater capture or other treating waste water for reuse. In other words, it is not just about lowering intake. It is about changing what the facility does with the water once it touches the site.
But the bigger story is the system forces driving water use upward. The source notes datacenter water consumption has been climbing for years, driven by growth in the number of facilities and by AI servers. AI servers run hotter than traditional kit and demand more cooling. That cooling demand matters because it sits directly between rising compute demand and real-world constraints like water availability and drought exposure. For context grounded in the source, water consumption at Microsoft’s facilities surged 34 percent to 6.4 million cubic meters in 2022, with generative AI blamed. If you are a board member or CFO tracking risk, this is a reminder that “demand growth” in AI is also a resource footprint growth story.
Amazon’s technical defense is that it relies mostly on non-water cooling methods. It says its facilities use “free air cooling” about 90 percent of the time, pulling in outside air and flowing it past servers to absorb heat, with no water involved. It does resort to evaporative cooling during the hottest weather. That split is important because it signals that water use may be more climate-sensitive than traditional buyers expect. When heat waves hit, the system flips from lower-impact cooling to higher-impact cooling, and that is when local water conditions, permits, and public tolerance collide.
And even if the industry wanted to eliminate the water habit entirely, the source argues that would be nearly impossible regardless of operator claims. That line is doing a lot of work. It suggests that the conversation will shift from “Can we stop using water?” to “Can we use less, use it more efficiently, and get credit for return flows?” In the US pipeline, the timing is awkward because many datacenters slated for construction are in areas already experiencing drought, according to analysis by The Guardian newspaper. So even companies that claim progress toward 2030 are operating inside a narrowing window of geography, public approval, and regulatory tolerance.
For executives at peer hyperscalers, the strategic stake is straightforward: water disclosure is turning into a measurable performance metric that can shape siting decisions, permitting outcomes, and the reputational cost of expansion. Amazon’s numbers will not end the debate, but they will raise the baseline for what boards, regulators, and communities expect next: transparent reporting, credible pathways to “water positive,” and cooling strategies that hold up when the weather does not cooperate.
This story's Key Insights and Take-aways are locked.
Create a free account to unlock Executive Actions for one credit.
Register to UnlockAlways free for Executives Club members. Join the Club
More in Technology

US export order shuts off Anthropic Claude Fable 5 and Mythos 5 globally
Enterprises lose top-tier Claude access overnight, with fallback models auto-routing and an uncertain path to restoration.

Anthropic will disable Fable 5 and Mythos 5 for everyone after export-control letter
A US order bars foreign users, and Anthropic says it will comply by turning off its latest frontier models globally.

Gaussian splatting tips into mainstream as nearly every major engine adds a plugin
Indie creators are already pushing photo-real 3D with low-cost scans, fast playback, and smaller exports.
