Ukrainian military intelligence has discovered Nvidia's Jetson Orin NX modules, consumer-grade AI chips, inside Russian S-71 'Monochrome' cruise missiles, raising urgent questions about how widely available AI hardware is circumventing export restrictions designed to limit military applications. The finding reveals a critical vulnerability in how the US regulates artificial intelligence technology: while powerful chips used to train AI models face strict export controls, the smaller chips that actually run those models in weapons systems remain freely available worldwide.

What Are Jetson Orin Modules and Why Do They Matter?

Nvidia's Jetson Orin NX is a compact system-on-module designed for space and power-constrained applications like drones and mobile robots. According to Ukrainian intelligence, the specific chip found in the Russian missile was marked as SNVUP6.MOP TE980M-A1 and was packaged in March 2025. The module features eight Arm Cortex-A78AE processor cores, a GPU with Ampere architecture containing 1024 CUDA cores, and 32 Tensor cores that deliver up to 157 sparse INT8 TOPS of performance for AI tasks. In practical terms, this means the chip can process visual information in real time and make decisions locally without needing to communicate with a central command system.

In the S-71 'Monochrome' missile, the Jetson Orin NX serves as the computational engine behind an electro-optical perception system that recognizes images in real time and guides the weapon to its target. The missile itself can be fired from Su-57 fighters or S-70 Okhotnik drones and has a range of up to 300 kilometers, carrying a 250-kilogram high-explosive fragmentation warhead.

How Are Consumer Chips Reaching Russian Weapons?

Nvidia positions the Jetson Orin as a consumer product sold to students, developers, and startups for legitimate applications. The company states that these modules are not export-controlled and are not officially available in Russia. However, the chips have been shipping since 2023 and are widely available through retail channels and reseller networks worldwide. A German electronics retailer, for example, sells them for approximately 685 euros with value-added tax included.

The critical gap in US export policy is that while the American government strictly controls high-end AI accelerators like Nvidia's H100 and B200 chips, which are used to train large AI models, it does not restrict hardware designed to run those models locally. This creates a paradox: the models themselves are trained anyway, and the hardware to deploy them is readily available on the open market.

"Our Jetson Orin modules are consumer-grade products sold to students, developers, and startups for a wide range of beneficial applications. They are not available in Russia and are not designed for military purposes. Pre-owned Jetsons are available through many reseller channels. Although we cannot track products after they are sold, if we determine that any customer is violating U.S. export controls, we will take appropriate action," said an Nvidia spokesperson.

Nvidia Spokesperson, Nvidia

What Does This Pattern Reveal About Russian Military Innovation?

This is not the first time Ukrainian intelligence has documented Russia using Jetson Orin NX modules in weapons systems. Previously, the Main Directorate of Intelligence of Ukraine's Ministry of Defense reported that Russia deployed these same chips in its Shahed MS001 autonomous drones for local decision-making and terminal guidance. The repeated use suggests Russia has developed a reliable supply chain for acquiring consumer AI hardware and integrating it into military platforms.

The S-71 'Monochrome' missile is distinguished by reduced observability, meaning it is harder to detect on radar, and autonomous targeting capability, which allows it to adjust course and select targets without constant communication from operators. These features depend entirely on the kind of real-time visual processing that the Jetson Orin NX provides.

Steps to Understanding the Export Control Loophole

• Training vs. Deployment: The US restricts chips used to train AI models, like the H100 and B200, which require massive computing power and cost millions of dollars. Russia cannot easily obtain these chips for training. However, once a model is trained, much smaller and cheaper chips can run it, and these deployment chips face no export restrictions.
• Scale of Production: Nvidia sells hundreds of thousands, if not millions, of Jetson Orin NX modules every year to legitimate customers worldwide. The sheer volume makes it nearly impossible to prevent some units from reaching unauthorized buyers through secondary markets or reseller networks.
• Dual-Use Technology: Consumer AI chips are genuinely useful for civilian applications like robotics, autonomous vehicles, and edge computing. Restricting them would harm legitimate innovation, making policymakers reluctant to impose blanket export bans on this category of hardware.

The discovery of Nvidia chips in Russian weapons highlights a fundamental challenge in modern export control policy. As AI technology becomes more distributed and consumer-grade hardware becomes more capable, the traditional approach of restricting specific products becomes increasingly difficult to enforce. The chips themselves are not inherently military; their application depends entirely on how they are integrated into larger systems. Nvidia cannot track products after they leave its supply chain, and reseller markets operate across borders with minimal oversight.

This situation raises broader questions about the effectiveness of current US export controls on AI technology. While restrictions on training chips may slow China's development of large language models, they do nothing to prevent adversaries from acquiring the inference hardware needed to deploy AI-powered weapons. As Russia's use of Jetson Orin modules demonstrates, the real bottleneck in military AI adoption may not be hardware availability but rather the expertise to integrate it effectively into weapons systems.