Nuclear Power Artificial Intelligence: Is the "ultimate" solution for large technology companies a reality or a pipe dream?

Wallstreetcn
2024.10.22 13:52
portai
I'm PortAI, I can summarize articles.

In the short term, nuclear power generation still faces significant financial and technical risks. Almost without exception, nuclear projects in the United States in the past few decades have faced cost overruns and delays; the commitment to sustainable development is also becoming increasingly difficult to fulfill

In the era of AI explosion, the technology industry is facing unprecedented energy challenges.

To meet the huge electricity demand of AI, countries around the world are turning their attention to nuclear energy and investing billions of dollars in developing next-generation nuclear technologies, hoping that nuclear energy can help curb carbon emissions associated with data centers.

However, nuclear power generation still faces significant financial and technological risks in the short term. Almost without exception, nuclear projects in the United States in the past few decades have faced cost overruns and delays.

Is nuclear-powered artificial intelligence really the "ultimate" solution to the energy problem, or is it just wishful thinking?

Nuclear Energy, the "Lifesaver" of AI?

In recent years, the emergence of large language models such as ChatGPT has led to a sharp increase in electricity consumption in data centers, prompting tech companies to turn to nuclear energy.

Last week, Google announced its first commercial transaction to support the construction of small modular nuclear reactors (SMRs) in the United States. SMR is a developing nuclear technology aimed at avoiding the high construction costs and long construction periods of traditional nuclear power plants.

Under the terms of the transaction, Google has pledged to support the startup Kairos Power to build approximately 7 small nuclear reactors in the coming years, expected to provide 500 megawatts of electricity.

Meanwhile, Amazon is not falling behind. The company has signed agreements with utility companies in Virginia and Washington to support the development of SMR projects and plans to invest in SMR technology developer X-energy.

Matt Garman, CEO of Amazon's cloud computing division, told The Wall Street Journal on Monday, "Some small modular reactors won't solve any problems in the 2020s, but they could be great energy sources in the 2030s and beyond."

The Battle for Future Energy

Despite the high hopes of tech companies for nuclear energy, the industry still faces significant financial and technological risks in the short term.

Almost without exception, nuclear projects in the United States in the past few decades have faced cost overruns and delays. For example, the two new reactors at the Vogtle nuclear power plant in Georgia have exceeded budget by billions of dollars and experienced significant construction delays.

Moreover, the contradiction between the sustainability goals of the tech industry and its energy needs is becoming increasingly apparent. Companies like Microsoft and Google originally set ambitious carbon neutrality goals, but with the development of AI technology, these commitments are becoming increasingly difficult to fulfill.

According to a McKinsey report, the power supply needed for data centers has become one of the biggest challenges tech companies face in the AI race.

Natural gas plays an important role as a transitional energy source in the short term. J.P. Morgan stated that tech companies are increasingly willing to build data centers powered by natural gas plants.

Dominion Energy, Virginia's largest power supplier, also plans to meet growing demand by developing natural gas

"No single energy source, grid solution, or energy efficiency program can reliably meet the growing needs of our customers," said Ed Baine, President of Dominion Energy Virginia