
Technology
|
Satellite Data Analytics
|
YC W26
|
Valuation:
Undisclosed

Last Updated:
March 24, 2026

Builds foundation models that translate raw SAR satellite data into analysis-ready optical imagery in real time. Uses deterministic one-step diffusion for 24/7 earth observation through clouds and darkness at 1/100th the cost of current tasking.
Foundation models for SAR-to-optical conversion. API and on-prem/edge deployment. Published DARN paper on foundation model adaptation for geospatial analysis.
Deep stealth. Core IP developed in-house. Founders building IP before scaling team.
AI-powered conversion of SAR radar satellite imagery into photorealistic optical images for all-weather, 24/7 earth observation.
It turns ugly, hard-to-read radar satellite images into clear, Google-Earth-style photos that anyone can understand, even when it's cloudy or dark outside.
It's like having a translator who can turn a doctor's illegible prescription into a perfectly typed paragraph—except the prescription is a radar image of Earth and the paragraph is a crystal-clear satellite photo.
No-code cloud platform enabling non-technical users to run planetary-scale satellite data analysis without writing code.
It lets anyone drag-and-drop their way through satellite data analysis the way Canva lets anyone design graphics without being a graphic designer.
It's like giving someone a fully stocked kitchen with a robot chef instead of expecting them to grow the ingredients, build the stove, and write the recipe from scratch.
AI-driven automated change detection and anomaly monitoring across satellite imagery time series for continuous situational awareness.
It automatically spots what changed on the ground between satellite photos—like a security camera for the entire planet that highlights only the important stuff.
It's like having a neighbor who watches your house 24/7 and only texts you when something actually important happens, not every time a squirrel crosses the yard.
Automated fusion of multi-sensor satellite data streams (radar, optical, elevation, vegetation) into unified analytical layers for comprehensive earth observation.
It stitches together different types of satellite data—like radar, photos, and elevation maps—into one unified view, the way your phone combines GPS, Wi-Fi, and cell signals to pinpoint your exact location.
It's like being a DJ who automatically mixes four different music tracks into one seamless song instead of making the audience listen to each instrument separately.
AI-assisted satellite mission planning and tasking management to optimize data collection scheduling and resource allocation.
It helps satellite operators figure out the smartest schedule for when and where to point their cameras in space, like a GPS route planner but for orbiting spacecraft.
It's like having an AI travel agent who plans the perfect multi-city trip for a fleet of planes, making sure every flight is full and no destination gets skipped—except the planes are satellites and the destinations are patches of Earth.
Atharva Peshkar (PhD CS CU Boulder, Harvard VCG researcher) brings PhD-level computer vision. Dhenenjay Yadav (ex-ISRO ML engineer, RL researcher at IIM Ahmedabad) brings ML engineering from India's space agency. Published research (DARN paper) demonstrates academic-grade rigor.