Wheelwright Prize Winner Marina Otero Verzier on Data Centers and AI

Why It Matters

The analysis spotlights how data‑center energy use fuels climate change and geopolitical tension, urging firms and policymakers to redesign digital infrastructure for sustainability and ethical memory management.

Key Takeaways

• Data centers consume massive energy, intensifying climate warming.
• Heat expelled by servers adds to atmospheric emissions.
• Energy demand creates geopolitical conflicts over electricity grids.
• Proposes ecosocial architecture prioritizing planetary limits over profit.
• Advocates data mourning, selective storage, and memory decay.

Summary

Data centers are the invisible backbone of the cloud, but their thermal management systems consume vast amounts of electricity and expel heat into an already warming atmosphere. Marina Otero Verzier argues that this thermopolitical infrastructure reshapes climate dynamics and creates new geopolitical fault lines over energy access.

She highlights three interlocking problems: the relentless energy draw that pushes power grids to their limits, the heat and emissions released by cooling systems, and the cultural assumption that all digital memory must be endlessly archived. Drawing on fieldwork since 2022, she maps alternative models—sovereign and indigenous clouds, feminist servers, and low‑tech computation—that prioritize planetary limits over profit.

Notable lines such as “The cloud is not weightless” and “Data mourning is a practice of remembrance” illustrate her call for a speculative yet actionable shift: design data infrastructures for embodied memory decay, selective preservation, and collective mourning of obsolete information.

If businesses adopt these ecosocial principles, they could reduce carbon footprints, mitigate energy‑related conflicts, and reshape data governance toward ethical, climate‑aligned stewardship. The proposal urges regulators and tech firms to rethink storage strategies, favoring resilience and planetary health over endless accumulation.

Original Description

In a recent public lecture titled “Feral Clouds,” Marina Otero Verzier, lecturer in architecture at the GSD, examined the “architectures and politics of data processing and storage.” To cool servers running 24/7, data centers require vast amounts of energy and produce harmful emissions. “The cloud is not weightless,” Otero Verzier argued. “It is a vast thermo-political infrastructure reshaping the planet’s climate.” As the AI race drives investment in “hyperscale” data centers, the social and environmental impacts of computing power have come into focus and sparked resistance movements around the world.

In 2022, Otero Verzier received the Wheelwright Prize, which funded two years of research on the future of data storage. She travelled the world studying different models for data storage. In her talk, she contended with the limits of creating “greener” data centers to sustain ever-increasing demand. Instead, she challenged us to more fundamentally rethink our relationship with the data, questioning the need to preserve everything indefinitely and suggesting it may be necessary to let some of it go—a process she calls data mourning.

Please watch the full lecture on the GSD website or YouTube channel.

In a separate event, Otero Verzier was joined by scholar and artist Kate Crawford for a conversation about data centers and AI.

Video: Matt Smith