NASA and Microsoft build an Earth Copilot

NASA has over 100 petabytes of Earth science data sitting in its archives. The problem? Unless you have a PhD in geospatial data science, good luck actually using any of it.

Now, Microsoft and NASA are teaming up to fix that. On November 14, Microsoft announced Earth Copilot, a custom AI prototype powered by Azure OpenAI Service that lets users query complex Earth science datasets using plain English.

Chatting with satellite archives

NASA's satellites constantly track everything from hurricane damage and wildfire tracks to ocean temperatures and atmospheric shifts. But navigating that massive ocean of numbers usually requires wrestling with specialized interfaces and esoteric file formats.

Earth Copilot bridges that gap by connecting advanced natural language processing directly into NASA's Visualization, Exploration, and Data Analysis (VEDA) platform.

Instead of crafting complex geospatial queries, researchers can simply type natural questions like "What was the impact of Hurricane Ian on Sanibel Island?" or "How did COVID-19 lockdowns change air quality across the US?"

The AI then pulls the exact relevant satellite datasets in seconds, dramatically cutting down the tedious retrieval process.

Former NASA engineer and current Microsoft Azure specialist Juan Carlos López noted that the system is engineered to tackle complex queries across massive datasets without bogging users down in technical friction.

Scientists get first dibs

Before you run to try it out, keep in mind this is still a proof-of-concept prototype.

NASA is currently keeping Earth Copilot restricted to its own researchers and scientists for internal testing. The agency says it is running rigorous assessments to ensure accuracy and prevent potential misuse before considering a broader public rollout on the VEDA platform.

Why it matters

Space-based observation is one of humanity's most expensive undertakings, but its real-world utility has long been bottlenecked by terrible user interfaces.

If Earth Copilot delivers, it could democratize access for people outside elite academic labs: city planners tracking urban heat islands, farmers checking regional soil moisture, or journalists investigating deforestation patterns.

We think Microsoft's focus here is spot-on. AI shouldn't just write emails or generate pictures; its highest leverage use is serving as an intelligent translator between humans and massive, impenetrable public databases.

That said, the stakes are sky-high. In climate research and disaster response, a hallucinated trend or misquoted coordinate isn't just an embarrassing typo—it can misdirect actual resources. Keeping the tool behind closed doors for rigorous sanity checks first looks like the only sensible move.