A year ago, I published an outlook for Earth observation (EO) in 2025, making 12 predictions about where the industry was headed. Time for an honest scorecard, and more importantly, what 2025 actually revealed about where EO is going.
A Year of Clarification
Twelve months ago, the EO sector was navigating familiar uncertainties: commercial adoption that always seemed one contract away from inflection, AI models that promised transformation but hadn't yet delivered it, and a funding environment caught between pandemic-era optimism and post-SPAC reality. The big questions - who is the real paying customer, what is the actual product, where does value accrue - remained pretty open.
2025 provided some answers. Not everyone will like them, but we now have clarity we didn't have before. Defense and national security became the organizing principle for much of the sector, not just its largest customer segment. AI crossed from research to deployment, with foundation models and AI weather models going operational. And the transatlantic consensus on civilian Earth observation might have fractured: the US proposed gutting NASA Earth science budgets and NOAA climate research while Europe approved a record budget explicitly framing EO as strategic infrastructure.
What follows is organized around the themes that defined the year: defense dominance, civilian divergence, commercial reality, the infrastructure nobody talks about, and AI's crossing from hype to deployment. The scorecard (hit/miss/mixed) matters less than what it reveals.
