The Startup That Launched Its Own Satellites to Make Weather Intelligent

Tomorrow.io operates the first commercial constellation of weather radar satellites, has raised more than US$175M at a valuation above US$1B, and has secured work with the US Department of Defense and the Air Force. This is not a weather app story. It is a new infrastructure story.

The point is not simply that a startup uses AI. It is that the company changed the data layer that AI depends on.

Tomorrow.io began as an API business delivering weather intelligence to operational users such as logistics firms, airlines, and construction businesses. But founder Shimon Elkabetz saw that the upstream data itself was inadequate: ageing satellites, sparse coverage, and refresh rates too slow for real operational decisions.

So the company built its own satellites with proprietary microwave radar instruments. That decision moved the company from software interface to data owner. Once that happened, the AI story became much more defensible.

Government users saw the value quickly. Air Force and Department of Defense contracts helped fund the satellite expansion. The Palantir partnership then operationalised weather intelligence inside secure environments and pushed the company toward agentic AI for weather - systems that do not just forecast conditions but recommend what to do next.

DeepSky takes that logic further by processing data on the satellites themselves rather than waiting for ground-station download, turning the network into an AI-native sensing layer.

Tomorrow.io understood that the AI revolution often requires new data, not simply better algorithms. While competitors tuned models on the same upstream sources, Tomorrow.io built proprietary sensors that generated data no one else had.

The lesson is that if your data is the same as everyone else's, your AI will converge toward the same insights. Their moat is not the model. It is the fact that the data layer is literally in orbit.