Applied Intuition wins U.S. Navy contract to scale its Data Engine across the fleet
Applied Intuition's Data Engine is built to fix the U.S. Navy's scarcity of high-quality training data by collecting, curating and pushing improved perception models back to ships over the air.
- millions
- months
What Happened
For decades the U.S. Navy has collected vast amounts of data from its ships, sensors and missions, but it lacked an organized infrastructure to log and aggregate that data, relying on ad-hoc methods instead. Even today, individual programs collect data for different purposes, each with its own logging standards and host environments. As a result, countless hours of valuable sensor data from operationally relevant environments never contribute to the insertion of new autonomous capabilities. Military systems also must learn to operate in environments that are difficult to replicate and constantly changing, unlike commercial autonomous vehicles that can accumulate millions of miles with geofenced test fleets.
- Can accumulate millions of miles with geofenced test fleets
- Must operate in environments that are difficult to replicate and constantly changing
Applied Intuition's Data Engine is the infrastructure layer that helps collect operational data, curates it and trains perception models. It then evaluates those models and deploys them back to vessels over the air, so what one ship learns reaches the entire fleet. The system integrates shipboard perception capabilities, scalable data collection solutions and shore-based infrastructure to support continuous improvement of AI/ML models across the maritime domain. Applied Intuition was recently awarded a contract with the U.S. Navy to scale Data Engine across the fleet.
- Vessels operate with thin, intermittent satellite connectivity.
- Ships must minimize electromagnetic emissions during operations.
- Crews face constant decisions about which clips are worth the bandwidth, and useful data means footage of other vessels rather than hours of empty water.
- Anything not sent over satellite during deployment must be manually collected when the ship returns, creating delays that can stretch for months.
- That data is often archived or discarded rather than leveraged for training.
- Even good data decays: a perception model loses usefulness as conditions change, mission requirements evolve and adversaries adapt.
“Autonomous systems live or die on training data, and the U.S. Navy has a severe scarcity problem—Data Engine is the infrastructure that solves it. It's the same continuous learning approach that made ground autonomy possible, now applied to an environment where the stakes are far higher. Every deployment teaches the next generation of systems.”
Together these capabilities establish a continuous learning cycle, allowing maritime AI technologies to evolve through real-world operational experience while laying the foundation for increasingly capable autonomous systems. Applied Intuition frames the stakes in terms of speed, arguing that future conflicts will be won by nations that can field autonomous capabilities faster and more effectively than their adversaries. The company says it is building toward a future where each military service operates its own Data Engine, tailored to its domain and feeding into a common infrastructure.
Previously from Applied Intuition, Inc.
Applied Intuition has previously addressed the testing side of military autonomy. As of September 2026, it was pitching its Digital Proving Grounds — always-on, vendor-neutral simulation — as the fix for a testing bottleneck that delays multi-vendor autonomous military systems.
Background drawn from MotorClaw's earlier coverage of Applied Intuition, Inc.'s official releases.
Why this matters
The U.S. Navy has awarded Applied Intuition a contract to scale Data Engine across the fleet, aiming to turn a threat spotted by one ship into a model every ship can detect. It targets what the company calls a severe training-data scarcity, a bottleneck that affects naval crews whose sensor data is often archived or discarded instead of used.
Terms in This Story
- Perception model
- A machine-learning model that interprets sensor data so a vehicle or vessel can detect and identify objects around it.
- Over the air
- Sending software or model updates to a vehicle or vessel remotely over a wireless connection rather than in person.
- Geofenced
- Confined to a defined geographic area, so a test fleet only operates within mapped boundaries.
Summarised from the linked release; details can be imperfect — always verify against the original source.