Fueled by the gold rush over artificial intelligence, big tech companies have launched an aggressive global expansion of data centers — massive physical structures packed with computer servers that store, process, and distribute the data and services we use every day.

But beneath the promise of technological progress lies a troubling reality: this breakneck expansion comes at a steep environmental and human cost.

A new generation of "hyperscale" data centers are rapidly proliferating across the globe, with capacity expected to triple by 2030 — driven almost entirely by AI's insatiable demand for computing power. Hyperscalers operated by tech giants often cover over a million square feet of land each.

These new-age data centers consume vast amounts of power and freshwater to run and cool the hardware inside. The energy and resource constraints emerging in traditional tech hubs are forcing companies to now look beyond the U.S. and China, pushing into countries with weaker environmental regulations and less public oversight, despite less favorable conditions (e.g. warmer countries in the global south).

While tech giants frame this expansion as a boon for local economic investment, job creation, and technological development, the reality is far more complicated. Most data centers employ relatively few local workers once constructed, and the promised infrastructure benefits often fail to materialize for surrounding communities. Meanwhile, environmentalists, public health advocates, and community leaders have raised alarm bells over the hidden environmental and human toll.

A sprawling global story

Over the past year, journalists and newsrooms around the world have produced incisive, in-depth reports and projects examining these impacts. These include The Earth Journalism Network's "Dark Side of the Boom" series, The New York Times' "The Global A.I. Divide", Rest of World's reporting on the global fallout, and Pulitzer Centre's AI Accountability Network.

Dirty Data, a collaborative project by the Environmental Reporting Collective (ERC), builds on this work by supporting nearly 40 journalists — reporters, photographers, editors, and designers — working on the issue across Asia, Latin America and Europe. Together, they are producing original reporting that contributes to a critical story still unfolding in communities around the world.

To produce Dirty Data, investigative and data journalists from Thailand, Singapore, Indonesia, India, China, Taiwan, Peru, Brazil, Spain and France venture into unreported regions and neighborhoods to deep dive into the human dimension of data center growth. We examine how governments are embracing data centers, pinning their hopes on AI as an engine of economic growth without considering the strain on the environment, resources, and communities.