At a recent White House meeting, OpenAI CEO Sam Altman presented a bold vision for building data centers with up to 5GW of power capacity, the equivalent energy usage of three million homes. Altman, alongside tech leaders like Nvidia’s Jensen Huang and CoreWeave’s Michael Intrator, joined discussions focused on the escalating power needs of AI data centers.

The Biden Administration responded by establishing a “Task Force on AI Datacenter Infrastructure” to address the logistical and regulatory challenges associated with such facilities. Altman’s report highlighted the potential economic and security benefits of investing in large-scale AI infrastructure, suggesting that projects of this magnitude could help maintain America’s lead in global innovation.

The report, which built upon a Goldman Sachs study, emphasized that U.S. data centers could require an additional 47GW of power capacity by 2030. OpenAI’s proposal envisions starting with one 5GW data center, potentially expanding to multiple facilities across various states. Constellation Energy CEO Joe Dominguez confirmed Altman’s ambitions to construct up to seven such data centers.

While the feasibility of finding U.S. locations capable of supporting these colossal data centers remains uncertain, companies like Microsoft, Google, and Amazon have also been discussing similar mega-scale projects. Microsoft, for instance, recently teamed up with Constellation to restart the Three Mile Island nuclear plant to meet its own growing data center needs.

As the demand for AI accelerates, the scale and energy requirements of future data centers are shaping up to be a pivotal issue in the technology landscape, fueling both excitement and concern over how to balance innovation with sustainability.