Calls to Pause Data Center Construction Criticized

Experts argue that halting data center buildout would stall AI progress, not improve safety

Apr. 7, 2026 at 3:23am

Got story updates? Submit your updates here. ›

A highly detailed, glowing 3D illustration of a futuristic data center server rack, with neon cyan and magenta lights illuminating the intricate hardware components, conceptually representing the complex digital infrastructure that powers advanced AI systems.As calls to pause data center construction grow, the underlying digital infrastructure that enables AI progress remains a critical, yet often unseen, part of the technology landscape.Spring City Today

Congress is being urged to pause data center construction, with a bill introduced by Bernie Sanders and Alexandria Ocasio-Cortez to halt the infrastructure that enables advanced AI. However, experts argue that this would do more than just ban new data centers - it would effectively pause AI development altogether. They contend that computing power is the foundation of AI, and constraining it would not buy time for safety, but rather ration progress through politics.

Why it matters

This debate highlights the tension between the rapid pace of technological progress and the need for thoughtful governance and safety measures. Proponents of the pause argue that the risks of unchecked AI development are too great, while opponents say that halting progress is not the answer and that safety can only be improved through actual deployment and iteration.

The details

The calls for a pause on data center construction echo previous demands for a six-month halt on training systems more powerful than GPT-4. However, those predictions of catastrophic outcomes never materialized, as AI systems continued to advance with improved safety and accountability. Experts argue that pausing data centers would not lead to the development of a governance architecture, but rather just slow progress, as technology cannot be frozen while institutions catch up.

  • In March 2026, Bernie Sanders and Alexandria Ocasio-Cortez announced a bill to halt the buildout of data center infrastructure.
  • Almost three years ago, a group of over 31,000 technologists, academics, and public intellectuals signed an open letter demanding a six-month pause on training systems more powerful than GPT-4.

The players

Bernie Sanders

A U.S. Senator who co-introduced a bill to pause data center construction.

Alexandria Ocasio-Cortez

A U.S. Representative who co-introduced a bill to pause data center construction.

Chris Koopman

The CEO of the Abundance Institute and author of the commentary.

Keving Frazier

A Senior Fellow with the Abundance Institute and the lead of the AI Innovation and Law Program at the University of Texas School of Law.

Got photos? Submit your photos here. ›

What they’re saying

“Pausing data centers is pausing AI. Computing is not incidental to AI development but is the foundation. Constraining it doesn't buy time for safety. It just rations progress through politics, and hands that rationing power to whoever controls the permitting process.”

— Chris Koopman, CEO, Abundance Institute

“Nothing about that panicked call for a pause was tailored to the complexities of AI development and diffusion. We were told that continuing to scale advanced AI without a coordinated halt risked catastrophic outcomes: loss of control over systems, destabilization of institutions, large-scale economic shocks, and even existential risk. This wasn't a distant prediction. They claimed the moment was already here, and continuing past GPT-4 without interruption was reckless.”

— Chris Koopman, CEO, Abundance Institute

What’s next

The judge in the case will decide on Tuesday whether or not to allow the data center construction to proceed.

The takeaway

This debate highlights the need for a more nuanced approach to AI governance, one that balances the pace of technological progress with the need for thoughtful safety measures. Halting data center construction or AI development altogether is not the answer, as it would likely slow the very research and iteration that can lead to more reliable and trustworthy AI systems.