Alaska DOT&PF relies on drone imagery to understand landslides, avalanches, flooding, and other high-risk events across some of the harshest terrain in the country. But their data was trapped behind a slow, fragile workflow: drones filled memory cards, crews drove them into town, and terabytes were uploaded to distant cloud regions for processing.

This routine created 28-hour to multi-day delays between capturing imagery and producing usable terrain models. For seasonal planning this was painful; in an active disaster it was a liability. Crews were already in the field while decision-makers were still waiting for the data needed to keep people safe.

Intelligence Bottlenecks

• Physical movement of data
• Limited backhaul from remote sites
• Dependence on distant cloud compute

Atlas gave Alaska DOT&PF a unified way to manage connectivity across the state. Instead of relying on improvised links or driving data into town, the team could see every terminal and field system from one operational view, with Starlink fully integrated into their network. Crews were able to upload imagery directly from the field the moment a drone landed.

This change reduced processing time from twenty-eight days to a few hours. For the first time, Alaska DOT&PF could receive updated models the same day an event occurred, creating the foundation for faster, more coordinated decision making.

The remaining bottleneck was distance to compute. Even with faster uploads, pushing large datasets to distant cloud regions still introduced delays that were not acceptable during active incidents. Galleon solved this by bringing dedicated edge compute into Alaska. With a single modular data center in Fairbanks running AEP, the department could ingest imagery on site and produce terrain models, point clouds, and AI-assisted analyses immediately.

Workflows that once took days collapsed to real time. Teams could assess hazards, direct crews, and confirm conditions while still on location. For Alaska DOT&PF, Galleon turned geospatial processing from a remote cloud task into a local capability that operates at the same pace as the mission itself.