Claude AI Powers NASA's First AI-Planned Mars Rover Drive

What changed

Anthropic's Claude AI model was utilized by engineers at NASA's Jet Propulsion Laboratory (JPL) to plot a 400-meter route for the Perseverance rover on Mars. This marks the first instance of an AI planning a rover drive on another planet. The AI generated commands in Rover Markup Language, an XML-based language, by analyzing overhead imagery and leveraging data from years of rover operations provided by JPL engineers. Claude iterated on its waypoint suggestions, refining the path. The AI-generated plans were then subjected to a rigorous simulation, modeling over 500,000 variables to predict rover positions and identify hazards. While minor adjustments were made based on ground-level imagery not available to Claude, the AI's plan was largely accurate and led to the rover successfully traversing the intended path on December 8 and 10, 2025. This process is estimated to cut route-planning time in half and improve consistency.

Why it matters for builders

This successful application of Claude demonstrates the growing capability of AI models to handle complex, real-time decision-making in environments with significant communication latency. For AI builders, this signifies a leap in AI's utility for autonomous systems, particularly in high-stakes, remote operations. It validates the use of AI for tasks requiring intricate planning, code generation, and adaptation to novel situations, paving the way for more sophisticated AI applications in robotics and exploration.

Practical impact

Builders can look to this as a case study for deploying AI in scenarios where direct human control is impractical due to signal delays. The use of Claude Code for generating specific commands in a bespoke language (Rover Markup Language) suggests potential for AI assistants to aid in developing and executing control systems for specialized hardware. The iterative refinement process used by Claude also offers insights into building more robust AI planning tools. NASA anticipates this will allow for more frequent drives and increased scientific data collection on Mars, accelerating our understanding of the planet. The success also sets a precedent for future missions, including NASA's Artemis campaign to the Moon and beyond.

Caveats and source limits

The source details the AI's role in planning a specific 400-meter drive and the estimated time savings. It does not provide specific benchmark results comparing Claude's planning accuracy against human operators beyond the qualitative assessment that "only minor changes were needed." The exact version of Claude used or specific technical details of the Claude Code environment are not elaborated upon. The source also does not detail the computational resources required for Claude to perform this task or provide information on the cost-effectiveness of using AI for such planning.