Public Sector
August 24, 2026 | 7 min read
The Department of the Navy’s new data and AI strategy defines technological superiority over adversaries through a metric called Mean Time to Effect (MTTE). MTTE tracks how quickly naval forces move through the Bits2Effects Cycle, measuring the time it takes to convert raw sensor data to direct operational impact. To maintain an advantage and defend U.S. interests, the Navy’s defense AI ambitions require it to outpace its adversaries technologically and operationalize AI at speed and scale. Simply deploying AI models is no longer enough as those adversaries advance their own technical capabilities. The advantage now goes to whichever military organization turns data into action fastest - and with the best judgment. Acting Secretary of the Navy Hung Cao recently highlighted this urgency for military decision making, calling the current operational climate the most compressed period of learning and adaptation in recent memory.
Domino’s work on Project AMMO (Accelerated Machine Learning for Maritime Operations) shows what successfully lowering MTTE looks like in practice. Running since 2022, this undersea mine countermeasures program has helped the Navy cut model retraining cycles from 12 months to 6 days. It also reduced deployment times to unmanned underwater vehicles from 6 months to 6 days. These gains dramatically shortened the path to operational effect for naval operations while meeting strict accreditation and classification requirements.
Project AMMO began when the Navy's Expeditionary Missions team partnered with the Defense Innovation Unit (DIU) to keep automatic target recognition (ATR) models current for mine countermeasures. DIU evaluated 30 commercial proposals and selected Domino to help solve a major technical bottleneck. At the time, ATR software updates were directly tied to hardware maintenance cycles. Improving what an unmanned vehicle could recognize required waiting for the vehicle to physically return to the surface before new software could be deployed.
The Project AMMO team decoupled those software and hardware cycles. Serving as the software integration layer across four commercial technologies and three separate contractor teams, Domino provided a governed IL5 environment, allowing each team to collaborate within clear access boundaries without duplicating infrastructure or exposing data across contract lines.
This setup delivered trusted, audit-ready sonar and imagery intelligence that naval leaders could immediately act on. Commodore Shaun Lieb, then Commander of Task Force 75, noted that Project AMMO proved the Navy can deploy target recognition models at the speed of operational relevance. Recognizing this success, the Navy awarded Domino a $16.5 million APFIT award in 2025 and a follow-on contract worth up to $99.7 million in 2026 to expand the technology beyond mine countermeasures.
Project AMMO aligns directly with the Navy’s mandate for a Modular Open Systems Approach, a standard for AI infrastructure that avoids proprietary lock-in. At the same time, it maintains strict data protection across multi-vendor teams and far-edge, disconnected platforms.
The strategy also emphasizes software reuse through the DON AI and Analytics inventory. Shortening MTTE requires avoiding rebuilds of capabilities that already exist. A unified platform allows naval teams to carry proven models, workflows, and governance controls into new multi-domain programs without starting from scratch.
Counting models in production is no longer an accurate benchmark for AI maturity. Instead, success is measured by how fast a force adapts to changing conditions on the battlefield. Project AMMO’s 6 day cycle proves that rapid adaptation is possible in complex defense programs.
Explore how Domino accelerates mission outcomes for the United States Navy through Project AMMO.
The Navy AI Strategy is a Department of the Navy (DON) policy document signed by Acting Secretary of the Navy Hung Cao in July 2026. It details how the Navy will use data and artificial intelligence to make faster operational decisions. It introduces Mean Time to Effect (MTTE) as the key metric for measuring that speed. MTTE tracks how long it takes to turn raw data into an operational effect, and the strategy's goal is for the Navy's MTTE to stay lower than its adversaries'. The document organizes that goal into six broader objectives, covering areas like AI infrastructure, governance, and workforce readiness, and it sets FY27 as the target for most of its early milestones.
The strategy sets six goals: accelerate operational AI, improve data readiness, optimize data and AI infrastructure, streamline data and AI governance, build a data and AI ready workforce, and partner and collaborate with industry, academia, and allies. Each goal comes with specific objectives and FY27 deadlines. Two goals in particular, AI infrastructure and AI governance, set conditions that commercial technology or partners have to meet because they cover how systems perform at the tactical edge and how quickly risk decisions can be made without slowing delivery.
The strategy's approach runs through the Bits2Effects Cycle, a five-stage framework that tracks data from collection to delivering an operational effect. Those results are then fed back into the next cycle for iteration. Project AMMO is a successful example of this cycle because MTTE was lowered significantly. The undersea mine countermeasures program cut model retraining time from 12 months to 6 days and deployment time to unmanned underwater vehicles from 6 months to 6 days, by decoupling software updates from hardware maintenance cycles. The Navy has since expanded the program through an APFIT award and a follow-on contract to extend the approach beyond mine countermeasures.
The strategy signals a shift in how defense AI gets evaluated. Instead of counting how many models are in production, the Navy is measuring how fast it can turn data into action relative to an adversary, which raises the bar for what counts as a working AI capability. It also points toward more reuse across programs: initiatives like the DON AI and Analytics Inventory and the CATTLE DRIVE consolidation are meant to keep new missions from rebuilding infrastructure that already exists elsewhere in the fleet.
Domino Data Lab empowers the largest AI-driven enterprises to build and operate AI at scale. Domino’s Enterprise AI Platform provides an integrated experience encompassing model development, MLOps, collaboration, and governance. With Domino, global enterprises can develop better medicines, grow more productive crops, develop more competitive products, and more. Founded in 2013, Domino is backed by Sequoia Capital, Coatue Management, NVIDIA, Snowflake, and other leading investors.