Valar Partners with Nvidia to Build Water-Efficient AI Data Center Powered by Nuclear Energy

Nuclear energy startup Valar is teaming with Nvidia on an next-gen AI data center that’s intended to consume substantially less water than traditional models and still generate enormous compute power. This alliance merges Valar’s state-of-the-art nuclear technology and Nvidia’s AI computing expertise to create an AI data center designed to support rapidly intensifying AI workloads and do it with a minimal impact on the environment. AI applications are becoming increasingly larger and complex, and the massive amounts of energy and water needed to run these applications in the data centers are an increasing concern for both tech companies and governments.

AI data centers typically use enormous amounts of electricity to run countless GPUs-high-powered graphics chips- and require enormous volumes of water in cooling systems, particularly in hotter regions.

Valar asserts it has an answer for both issues. Valar is using small, innovative reactors to create consistent carbon-free electricity in conjunction with new technologies that significantly decrease water consumption compared to existing data centers and is hoping the new facility can operate in a wider geographic region. For Nvidia, this project represents a rising prioritization in building more sustainable AI data centers. Demand for Nvidia AI chips is booming as the technology world rapidly adopts applications of generative AI but with rapid deployment comes increased examination of the energy use and the water required in large data centers.

Energy consultants estimate global demand for AI power could soar in the next decade forcing tech giants to explore alternate energy sources, like nuclear.

Energy from nuclear plants can provide round-the-clock stable and carbon-free electricity, unlike many renewable energy sources like solar and wind. Water use in AI data centers has emerged as another issue. Many modern data centers rely on millions of gallons of water for evaporation for their cooling systems, which the industry hopes to remedy with better liquid and closed-loop cooling systems. This Valar-Nvidia initiative aligns with the wider trends of seeking to minimize overall use. Instead of dealing with energy use and water use as disparate challenges, the companies will optimize for both. This partnership also points to a merging of the energy and tech sectors.

The immense demand for computing power from AI has fueled a growing number of partnerships between the nuclear industry, power utility companies, and semiconductor manufacturers as companies grapple to create reliable energy supplies with sustainability in mind. Enterprises stand to see more cost efficiencies and better availability of high-performance computers with a move toward smarter, greener infrastructure. If this plan works, the nuclear power company could serve as a model for other data centers.

As more companies harness AI's growing power, innovative solutions for electricity supply and cooling are becoming as critical as new AI chips.