General Caine casually says the air campaign in Iran is shifting to "dynamic targets". What does this mean and why should we care?

Dynamic targets are unplanned or unanticipated objectives that have become known in near-real time to the intelligence surveillance and reconnaissance (ISR) kill chain. Given the kill chain finds, tracks, targets (includes selecting ordnance and collateral damage estimation), engagement and battle damage assessment while the attack platform is whizzing around at hundreds of knots strongly implies automation and almost certainly the use of AI.

We don't know the capacity and granularity of this system but I suspect an outraged media will soon be aware of it.

The Maven Combat System, more formally known as the Maven Smart System (MSS) or simply Maven, is a U.S. Department of Defense (DoD) AI-powered command-and-control platform. It integrates vast amounts of battlefield data from sensors, enabling faster intelligence analysis, targeting, mission planning, logistics, and decision-making.

Origins in Project Maven
\
It evolved from Project Maven (officially the Algorithmic Warfare Cross-Functional Team), launched by the Pentagon in April 2017. The project aimed to apply machine learning (especially computer vision) to process overwhelming volumes of surveillance data—such as drone video, satellite imagery, and other intelligence feeds—that human analysts could not handle alone.

Early efforts focused on automatic target recognition (ATR) to detect and classify objects like vehicles, personnel, or infrastructure in imagery. The program initially partnered with tech companies (including a controversial early involvement with Google, which later withdrew due to employee protests). It is now primarily associated with Palantir Technologies, which developed the core data-fusion and user interface platform. The effort operates under the National Geospatial-Intelligence Agency (NGA) and has expanded across the DoD.

Key Capabilities

Maven functions as a "god’s-eye view" of the battlespace by fusing data from 150+ sources, including:

Satellites and geospatial intelligence

Drones and airborne ISR (intelligence, surveillance, reconnaissance)

Ground-based radar, signals intelligence (SIGINT)

Other inputs like social media or logistics data

AI and machine learning components automatically detect, classify, and prioritize military-relevant objects or changes (e.g., new enemy facilities, rocket launchers, or ships). It supports:

Targeting: Nominating targets, pairing them with appropriate weapons/systems, sequencing strikes, and assessing battle damage. Reports indicate it can generate up to 1,000 targeting recommendations per hour in some setups.

Command and Control (C2): Creating a common operating picture for real-time situational awareness, mission planning, and rapid "sensor-to-shooter" loops.

Logistics and sustainment: Predicting supply needs and supporting contracting/operational planning.

Other uses: Disaster relief, medical data integration (e.g., point-of-injury trauma info), and multi-domain operations (air, land, sea, space, cyber).

The interface allows commanders or analysts to visualize the battlefield on a digital map, query data (sometimes via natural language with embedded AI agents), approve/reject AI suggestions, and execute actions with minimal clicks. Humans retain final decision authority for lethal actions—AI assists but does not autonomously engage targets.

This dramatically compresses timelines: processes that once required hundreds of personnel over hours or days can now occur in minutes.

Development and Deployment

Prototyping: Refined through U.S. Army exercises like Scarlet Dragon (led by the XVIII Airborne Corps).

Fielding: Deployed across all major U.S. combatant commands by early 2026. It has supported operations in Iraq, Syria, Yemen, the Red Sea, and reportedly combat operations against Iran in 2026 (e.g., aiding target identification and prioritization amid strikes).

Allies: NATO acquired a version in 2025 via Task Force Maven, with demonstrations of rapid integration with third-party tools (e.g., from European firms).

U.S. branches: Used by Army, Air Force (e.g., in human-machine teaming tests), Marines, and others. It has grown to over 20,000 users.

Status: In 2026, the Pentagon moved to formalize it as a "program of record" for stable multi-year funding (investment reportedly scaling significantly).

Contributors include Palantir (core platform), plus companies like Amazon Web Services, Microsoft, Maxar, and others for data, sensors, or models.

Significance and Context

Maven represents a shift toward AI-augmented warfare, emphasizing speed and data dominance in multi-domain operations. Proponents highlight how it reduces cognitive overload for operators and provides an edge against peer adversaries. Critics raise ethical questions about AI in targeting, potential errors in complex environments, and reliance on commercial tech.It is often described as one of the DoD's most visible and operational AI tools, moving from experimental computer vision to a broader warfighting decision-support system.

In short, "Maven Combat System" is informal shorthand for this AI-driven platform (Maven Smart System) rooted in Project Maven, designed to turn overwhelming sensor data into actionable combat advantages while keeping humans in the loop for critical decisions.