Valar Atomics Demonstrates Nuclear Microreactor Powering Nvidia RTX Spark, Plans 30MW AI Data Center

Valar Atomics successfully powered an Nvidia RTX Spark desktop PC using its Ward 250 nuclear microreactor, marking a first in live nuclear microreactor demonstration. The startup plans a 30MW closed-loop AI data center with Nvidia, eliminating local water use.

David Park David Park
2 min read
Valar Atomics Demonstrates Nuclear Microreactor Powering Nvidia RTX Spark, Plans 30MW AI Data Center

Valar Atomics has taken a significant step in integrating advanced nuclear microreactor technology with AI hardware by powering an Nvidia RTX Spark desktop PC live on stage using its Ward 250 microreactor. This demonstration is notable as a first-of-its-kind real-time showcase of nuclear microreactor power applied directly to high-performance computing equipment.

The Ward 250 microreactor is a compact nuclear fission device designed for safe, reliable power generation with minimal environmental footprint. Its ability to sustain power for sensitive electronics like the RTX Spark highlights progress in nuclear microreactor control and stability, critical for future commercial deployments.

Valar Atomics and Nvidia are collaborating on a 30MW closed-loop AI factory powered by these microreactors. This facility aims to operate without relying on local water sources for cooling, a major innovation addressing the growing concern of water scarcity in data center operations, especially in drought-prone regions.

This development indicates a potential paradigm shift in data center infrastructure, where nuclear microreactors could provide dense, carbon-free power directly at or near AI computing sites. It alleviates pressure on electrical grids and water resources, enabling more sustainable scaling of AI workloads that demand increasing power density and reliability.

The technology's impact will be closely watched by chipmakers, data center operators, and infrastructure engineers as it challenges conventional reliance on fossil fuels and large-scale grid power. If scaled, this could influence semiconductor process node deployment strategies by ensuring stable, local power supply for next-generation AI silicon fabrication and training environments.

Key considerations ahead include regulatory approvals, integration complexity, and economic viability compared to existing power solutions. However, the live demonstration marks a credible proof point that nuclear microreactors can interface effectively with sensitive, high-performance AI silicon hardware.

Sources

  1. 01 Startup activates nuclear microreactor live on stage to power an Nvidia RTX Spark desktop PC — firm working with Nvidia to build a 30MW closed loop AI factory that doesn’t use local water — Tom's Hardware