To address the escalating energy demands triggered by the proliferation of AI data centers, Google and Westinghouse have collaborated to optimize the development and management of 10 proposed nuclear reactors scheduled for construction in 2030.
The two companies have devised a strategy over several months to expedite the typically sluggish timeline for nuclear reactor construction using a specialized AI-driven optimization platform on Google Cloud. Early trials of the platform have demonstrated substantial savings in both cost and time.
Westinghouse unveiled an $80 billion initiative earlier this year, backed by the US government, to kick off the construction of 10 AP1000 nuclear reactors by 2030. These 1,150 MW reactors have the capacity to power 7.5 million households and multiple data centers, addressing the soaring energy demands associated with AI technology.
This endeavor aims to narrow the widening energy deficit brought about by the escalating energy requirements. However, with massive data center projects already underway nationwide, the industry is under pressure to expedite the deployment of new energy sources, including renewable and non-renewable options.
Nuclear-Powered Data Center Wave
The attractiveness of nuclear power as an option has grown due to improved safety, efficiency, and global sustainability concerns. The International Atomic Energy Agency (IAEA) foresees a doubling of nuclear operational capacity by 2050, with more than 20 countries committed to tripling nuclear capacity by the same year.
Nevertheless, historical challenges have often plagued the construction of nuclear capacity, with the IAEA reporting that reactor construction can span 10-12 years. Google Cloud and Westinghouse are optimistic about significantly reducing this timeframe through AI integration.
“Our ultimate objective is to leverage our cutting-edge capabilities and AI portfolio to assist the industry in establishing a safer, more reliable, and sustainable system for all clients,” stated Raiford Smith, Google Cloud’s global director for the power and energy sector. “This partnership marks a pivotal moment in our industry, specifically focusing on the utilization of AI to revolutionize the construction of advanced nuclear reactors and optimize the operations of existing facilities.”
Reactors Fast-tracked
Westinghouse stated that the new AI platform will expedite the commissioning of the reactors within 5-6 years of commencing construction. In contrast, the Vogtle project in Georgia, powered by Westinghouse’s AP1000 reactors, required 15 years to finalize due to delays and budget overruns.
Lou Martinez Sancho, Westinghouse’s Chief Technology Officer and Executive Vice President of Research and Development and Innovation, expressed optimism that the new AI innovations would significantly reduce project timelines.
“We are confident that through AI integration, we can construct and operate at a faster and safer pace than ever before,” remarked Martinez Sancho during a press briefing, highlighting that a typical reactor project entails coordination among 10,000 employees at the peak of construction. “Addressing all these challenges necessitated a fresh approach and methodology that AI now provides us.”
During the presentation, Scott Sidener, Westinghouse’s Chief Data Scientist and Chief Digital Engineer, demonstrated the new tool’s interface, which analyzes millions of construction tasks to devise the optimal daily schedule and minimize delays. Sidener showcased how the tool reduced building costs for a task by 25%, equating to $1 million, with a simple click.
“This can be accomplished in seconds rather than a week,” Sidener emphasized.
