LONDON & SAN FRANCISCO – Addressing the critical “2% problem” in Earth observation, The Compression Company has closed a $3.4 million pre-seed round led by Long Journey Ventures. The funding, announced on Feb. 12, 2026, aims to commercialize an AI-driven, software-only compression platform that runs directly on orbital GPUs to reduce satellite file sizes by over 95%.
The Downlink Deficit
As the number of Earth observation (EO) satellites is projected to triple to over 5,400 in the next decade, the industry’s terrestrial infrastructure is struggling to keep pace.
- The Reality: Currently, only an estimated 2% of recorded satellite data reaches Earth.
- The Constraint: Satellites typically have windows of only 5 to 10 minutes per pass to transmit data to ground stations.
- The Waste: Millions of dollars in high-fidelity imagery are currently discarded or significantly delayed due to these “downlink bottlenecks.”
Long Journey, known for its early-stage bets on SpaceX and Anduril, is positioning the startup to solve the primary logistical hurdle of the modern EO boom: the massive disparity between data collection capacity and downlink bandwidth.
Technical Architecture: Selective Fidelity
Unlike standard lossless or lossy compression (like JPEG), The Compression Company’s platform uses context-aware AI to apply “tiered” compression within a single image:
| Target Area | Compression Level | Rationale |
| Cloud Cover | Ultra-High | Clouds account for ~60% of imagery and hold near-zero value. |
| Open Ocean | High | Low-frequency data with minimal tactical/commercial interest. |
| Identified Assets | Zero/Lossless | Ships, buildings, or vehicles are preserved at full fidelity. |
| Atmospheric Noise | High | Filtered out to reduce “noise” before transmission. |
Strategic Rationale: The Move to Edge Computing
The startup is capitalizing on the trend of Orbital Edge AI. By delivering the technology as a software-only solution, it can be uploaded to existing constellations already equipped with onboard compute (such as NVIDIA Jetson or similar space-hardened GPUs) without requiring hardware modifications.
For operators, this translates to a 20x increase in the amount of “usable” data delivered per ground station pass, significantly lowering the cost-per-byte for commercial and defense users.
Outlook: Enabling Real-Time Global Monitoring
The $3.4M pre-seed round will fund the company’s first on-orbit demonstration, scheduled for late 2026. If successful, this technology could fundamentally shift the economics of the “Golden Dome” defense shield and climate monitoring programs, where the speed of data delivery is often more critical than the volume of raw pixels.