Skip to content
IFAI

Industry

Meta pushes back on data center myths over water and energy

Published Reach 35

Meta just published a detailed breakdown countering growing backlash over AI data centers. The company claims its liquid-cooled AI facilities consume less water annually than an average U.S. golf course, and that it funds its own grid upgrades to keep local utility bills from rising. Why it matters: Big tech is fiercely defending its infrastructure footprint as public scrutiny around AI's environmental toll heats up.

Meta pushes back on data center myths over water and energy
Image: about.fb.com

Whenever a new data center gets announced, neighbors usually brace for the worst. The prevailing narrative is that these massive server farms drain municipal reservoirs, spike electricity bills, and sit completely empty once the ribbons are cut.

Meta decided it had heard enough. In a new piece addressing the biggest myths around data centers, the company pushed back hard on allegations surrounding water consumption, utility bills, and employment.

Less water than a golf course

The first myth Meta tackled is that server farms suck up all the drinking water.

Meta explained that its modern facilities do cooling very differently than people assume. Traditional air-cooled centers only consume water to chill intake air during the hottest months of the year.

For heavy AI workloads, Meta relies on closed-loop liquid cooling. That setup circulates a sealed mixture of water and glycol over hot chips, dumps the heat outside, and recirculates the fluid without continuous evaporation.

As for supply, Meta says it only draws water from permitted, metered sources. At one site tapping an aquifer, the water table was unsuitable for drinking in the first place. Put together, Meta claims many of its centers use less water annually than an average US golf course.

Subsidizing the grid and guaranteeing jobs

The second myth is that mega-scale facilities drive up electric bills for surrounding families.

Meta admitted data centers chew through huge amounts of power. But the company insisted it does not simply plug into the wall and leave locals with the bill.

In Louisiana, for instance, Meta structured its agreement with Entergy so the tech giant directly pays for the required new power generation and transmission lines. The contract explicitly contains guardrails ensuring current rate-payers do not shoulder capital upgrade costs.

Finally, Meta rejected the claim that data centers are job deserts. Construction requires thousands of skilled trades over several years, and ongoing operations still demand on-site IT technicians, facilities staff, and security teams.

To grease the skids, Meta launched America's Workforce Academy earlier this year, offering free skilled-trades training with guaranteed job placement at partner sites upon graduation.

Why it matters

For residents living near proposed AI corridors, this transparency gives community boards tangible benchmarks to hold tech giants accountable. Locals can now point-blank ask developers whether they are deploying closed-loop liquid cooling or agreeing to shoulder transmission grid costs without rate hikes.

Across the wider tech sector, the physical ceiling of AI has officially arrived. Compute power is no longer limited just by GPUs, but by community goodwill, megawatts, and water rights.

In our view, Meta played its hand very smartly here. Comparing a giant compute cluster to a suburban golf course is public relations genius because it replaces abstract megawatts with a familiar luxury.

That said, it looks like a curated reality. Not every tech operator can afford Entergy-style bespoke grid contracts, and closed-loop designs do not eliminate the broader environmental footprint of power generation behind the scenes. Meta may be an exceptional neighbor, but the industry as a whole still has a lot to prove.

Sources (1)

  1. [1]
    about.fb.com(opens in a new tab)Originalabout.fb.com

Best of AI, daily