NASA's Mars Rover Takes First AI-Plotted Drive, Marking a New Era for Exploration

In a quiet milestone for interplanetary travel, NASA’s Perseverance rover recently completed a drive across Mars guided not by human hands, but by an artificial intelligence model. Over three days in December, the rover traveled roughly 400 meters through the rocky terrain of Jezero Crater, following a path charted by Anthropic’s Claude. It’s the first time a large language model has been trusted to pilot a robotic mission on another world.

Plotting a safe course for the car-sized rover is typically a meticulous, human-led process. Operators combine orbital and rover-camera imagery to lay down a precise ‘breadcrumb trail’ of waypoints, avoiding hazards that could strand or damage the precious machine. For this test, engineers fed Claude Code—Anthropic’s programming agent—years of contextual mission data. The AI then constructed a route ten meters at a time, reviewing and refining each segment.

Engineers at NASA’s Jet Propulsion Laboratory rigorously validated the AI’s work, running the proposed path through their daily simulation tools. They made only minor adjustments, one based on ground-level images Claude hadn’t accessed. The result was a success. NASA estimates this approach could halve planning time, creating more consistent drives and freeing scientists to focus on analysis and discovery.

The efficiency boost comes at a critical moment for the agency. Following workforce reductions in 2025 under the Trump administration, NASA entered 2026 with a leaner team. Although Congress rejected a proposed deep cut to its science budget this January, funding remains tight. The agency now faces the monumental task of returning to the Moon with a fraction of the personnel it had during the Apollo era.

For Anthropic, the achievement marks a dramatic leap. Less than a year ago, its models were famously stumped by a classic video game. Now, one has successfully navigated the complex, unforgiving landscape of Mars. NASA officials express optimism, suggesting such autonomous systems could one day guide probes into the solar system’s farthest reaches.