Home Innovation Networking Nuclear Power Plans for AI Dat...
Networking
Business Honor
08 November, 2024
Meta and Amazon encounter roadblocks in nuclear power ambitions for US-based AI data centers.
Toward meeting all the energy requirements for its AI-powered data centres, Meta and Amazon have had a lot of hurdles in trying to power facilities based in the US using nuclear energy. That should be counted among the significant challenges companies face towards the development of sustainable solutions in powering fast-growing data infrastructure.
This Meta AI data center near a nuclear power plant ran into an unexpected hurdle when some surveyors found a rare species of bee at the site. Due to this, environmental regulators stayed the project, and Meta ultimately put the development on the shelf. According to some reports, CEO Mark Zuckerberg told employees that the bees had complicated the project so much that it was not practical to move forward with the project. However, Meta doesn't give up and continues to work towards reducing carbon footprint. In this regard, it is working hard to find other alternative sustainable energy options for its data centers.
Meanwhile, Amazon faced a separate challenge involving its Cumulus data center in Susquehanna, Pennsylvania. The company sought approval to increase the power supplied to the facility by the adjacent nuclear plant, from 300MW to 480MW. However, the US Federal Energy Regulatory Commission (FERC) rejected the application, citing insufficient justification for the power increase. FERC raised concerns about the potential strain on the region’s transmission grid, which may not have the capacity to handle the increased demand from co-located facilities. The decision underscores the delicate balance between expanding data center energy needs and ensuring grid reliability.
Both companies are focused on transitioning to carbon-free energy sources, but these regulatory challenges demonstrate the complexities of integrating nuclear power into the energy mix for next-generation data centers. As AI continues to drive demand for computing power, finding viable, sustainable energy solutions will remain a critical issue for the industry.