Scout AI raised a $100M Series A on April 29, 2026, with investors including Align Ventures, Draper Associates, Booz Allen Ventures, Decisive Point, BVVC, Evolution VC Partners, Neman Ventures, Perot Jain, and FJ Labs.

Before the Series A, Scout AI raised a $15M seed round announced on April 16, 2025, when it emerged from stealth. That round included Booz Allen Ventures as a strategic participant alongside the other investors listed above.

Total disclosed funding across both rounds stands at $115M.

Scout AI builds an autonomy and coordination layer between a human commander and a mixed fleet of unmanned systems, ground vehicles, drones, and eventually maritime and other platforms. Its core product, Fury, is a defense-specific vision-language-action foundation model that connects what a robot perceives through its sensors, what a human operator specifies in natural language, and what action the robot or fleet should take next.

In use, a mission commander types or speaks a high-level objective, such as "search this route, identify the target vehicle, and keep the ground asset covered." Fury turns that into a structured mission plan, presents it for human approval, assigns subtasks to each vehicle in the fleet using natural language, monitors live telemetry and video feeds, and updates the plan as conditions change. Instead of joysticking individual robots, the operator supervises an AI system that handles decomposition and coordination.

Fury generates platform-native structured instructions for each vehicle's existing API without replacing or rewriting that vehicle's underlying flight controller or mobility stack. A customer using a ground vehicle from one vendor and drones from another does not need to standardize on new hardware, because Fury reads each platform's tool definitions and produces the right format for each one. Scout AI describes this as agentic interoperability.

The system is designed for field deployment. Fury runs on camera-only passive sensing, requires no lidar or radar, and is optimized for low-power edge inference. Its architecture pairs a larger model, over 100 billion parameters, that can run on a secure cloud or air-gapped on-site computer with smaller roughly 10-billion-parameter models that run directly on the vehicles. This hierarchy allows operation in GPS-denied and comms-degraded environments where cloud connectivity is unavailable.

Scout AI's reference hardware platforms, the G01 ground vehicle, A01 drone, and the NOMAD UGV built with Hendrick Motorsports Technical Solutions, are intended to show that Fury works across form factors rather than to compete with vehicle OEMs. NOMAD introduced a second-generation Fury hardware stack that is more than 90% smaller than prior versions, a relevant constraint because size, weight, and power often determine whether autonomy systems can be retrofitted into compact platforms. In a February 2026 Fury Autonomous Vehicle Orchestrator demo, a heterogeneous air-and-ground fleet executed a live mission workflow including battle-damage assessment and a continuously updated common operating picture.

Scout AI sells into defense as a B2B software and integration business, with the government as the end customer and a mix of direct contracts and partner-enabled channels as its go-to-market path. The company says it is not a prime contractor or vehicle manufacturer. It aims to own the AI reasoning and coordination layer, while hardware OEMs and integrators handle the physical platform.

Monetization is contract-driven at this stage: development contracts, prototype and evaluation awards, and OTA agreements with DoD program offices. The Army UxS award with Textron Systems as the vehicle integration partner is the clearest example of the B2B2C model, where Scout AI supplies the intelligence layer inside a partner platform that is ultimately delivered to a military end user. The Hendrick Motorsports Technical Solutions partnership on NOMAD follows the same template.

This model is designed to make Fury an embedded software layer across multiple platform families rather than a single-vehicle product. If a defense OEM or integrator already has procurement access but lacks a modern autonomy layer, Scout AI can license Fury into that relationship instead of competing for the hardware contract. The result is an "intel-inside" approach, where Scout AI monetizes the platform families that win budget share rather than betting on one vehicle class.

The cost structure is heavier than pure SaaS. Training and refining a defense-specific embodied AI stack requires significant compute spend, and Scout AI operates both R&D lab space and hundreds of acres of real-world proving grounds for hardware-in-the-loop testing. Gross margins could improve over time if Fury becomes a recurring software layer across many deployed platforms, but near-term the business carries substantial research, testing, and deployment-support costs alongside contract revenue. Prototype contracts and partner integrations also generate field data and operator feedback that improve model performance, attract additional contracts and partners, and fund larger-scale model training. Scout AI's jump from a $15M seed to a $100M Series A in roughly one year suggests that loop is accelerating faster than contract cash flow alone would support.

Scout AI is competing for the autonomy and orchestration layer for unmanned defense systems. That layer is also a target for major defense-tech platforms approaching from hardware upward or software downward.

Anduril is the clearest strategic threat because its Lattice stack already covers much of what Scout AI is proposing: cross-domain command and control, mission autonomy, mesh networking, and an SDK layer for tasking heterogeneous assets, while also fielding its own air and undersea systems. Anduril can pull autonomy adoption through hardware sales and existing program relationships, which makes a bundled deployment easier for customers to choose than Scout AI's separate intelligence layer.

Shield AI is the most direct software competitor. Its Hivemind stack is designed for edge autonomy without GPS or comms, has flight-test pedigree across multiple aircraft types, and is paired with owned platforms like V-BAT. Shield AI's reported $12.7B valuation in early 2026 suggests mission autonomy software is forming its own category, and Shield AI has a head start in airborne programs that Scout AI will need to displace or complement.

Saronic has built an integrated maritime autonomy stack, vessels, mission autonomy, simulation, and the Echelon C2 platform, and is scaling with a $9.25B valuation and active Navy contracts. For Scout AI's cross-domain ambitions, Saronic presents in maritime the same challenge Shield AI presents in air: a purpose-built competitor with deeper domain data and customer relationships than a general orchestration layer can easily replicate.

Overland AI is the most relevant ground-autonomy specialist, with its OverDrive stack and SPARK retrofit kit designed for off-road tactical maneuver in GPS-denied environments. Overland AI addresses a narrower problem than Scout AI but can win when buyers prefer a specialist optimized for a specific platform over a broader foundation-model vendor. Scout AI's clearest counter is to position Fury as the coordination layer above vehicle-native autonomy rather than trying to replace it.

Forterra is the closest analog to Scout AI's ground-autonomy kit economics, having built a recurring autonomy-kit model around retrofitting existing Army vehicles. Forterra's approach, selling a repeatable product into a specific urgent use case and then expanding, is the same playbook Scout AI is pursuing, and Forterra has a head start in Army ground programs.

A less visible competitive threat is that Anduril's Lattice APIs, Shield AI's autonomy management layer, and Saronic's Echelon each target the same middleware surface Scout AI is pursuing: the translation layer between commander intent and heterogeneous robot APIs. If one of these platforms becomes the default schema for unmanned fleet tasking, Scout AI risks ending up as a niche add-on rather than the operating layer.

Scout AI's strongest position is in programs where no single hardware vendor can dominate and where the government resists lock-in. Army UxS-style retrofit efforts and joint operations involving multiple unmanned vendors are the clearest examples.

Scout AI's expansion logic rests on a single bet: that the intelligence and coordination layer for unmanned systems will become more valuable than any individual robot platform, and that early adoption of that layer can compound across domains and customers.

The move from single-vehicle autonomy to multi-asset orchestration is Scout AI's most important near-term TAM expansion. The February 2026 Fury Orchestrator demo showed a heterogeneous air-and-ground fleet executing a live mission workflow from a single natural-language command, a materially different product from autonomy on one vehicle.

That distinction matters because the Pentagon's Replicator initiative is explicitly pushing toward thousands of lower-cost attritable autonomous systems operating together, which raises the value of software that can coordinate mixed fleets more than it raises the value of any single platform. The DoD's FY2026 budget request shows counter-unmanned systems funding rising from roughly $2.2B to $3.2B, and demand for integrated enablers that can manage large numbers of unmanned assets is the environment where an orchestration layer can expand its addressable market faster than a hardware OEM.

Scout AI's early traction is concentrated in Army-adjacent channels, but Fury's product language is deliberately joint, ground, air, sea, and space. The expansion path runs from Army UxS retrofit programs into Air Force drone teaming, Navy surface autonomy, and SOCOM small-unit robotics, each with a distinct procurement channel, budget, and program office.

The Army's FUZE model and xTechOverwatch pathway, which Scout AI has already entered as a winner, are designed to move startups from competition to SBIR to follow-on production contracts faster than traditional procurement. That transition funnel is the clearest near-term mechanism for Scout AI to convert prototype traction into program-of-record scale, and a successful Army UxS demonstration in May 2026 is the key near-term catalyst for expansion into other services.

Scout AI has described its mission as deploying AI agents for the DoD and its allies, making allied defense sales the clearest geographic expansion vector. Five Eyes and close Indo-Pacific and NATO-aligned countries face the same demand for lower-cost autonomous mass, operate mixed fleets from multiple vendors, and often buy U.S.-origin systems that need interoperable autonomy software.

The more scalable version of this expansion is the intel-inside licensing model. Rather than selling directly to allied governments, Scout AI embeds Fury into partner platforms, OEMs, primes, and integrators that already have procurement access in those markets. That is the same B2B2C structure Scout AI is using with Textron domestically, and it expands TAM without requiring a separate international sales organization. The constraint is that autonomous weapon system exports remain subject to technology-security and foreign-disclosure approvals, so allied expansion is real but gated by policy timelines.

Procurement conversion: Scout AI's $11M in booked contracts and its Army UxS prototype award are early signals of demand, but defense procurement has a well-documented gap between prototype success and program-of-record scale, and Scout AI's revenue trajectory depends largely on whether those evaluations convert into larger, recurring production contracts rather than remaining one-off development awards.

Platform absorption: Scout AI's hardware-agnostic, partner-led model supports its go-to-market motion, but it also means that primes and OEMs, including Anduril, Shield AI, and Saronic, could internalize similar orchestration capabilities or standardize their own API schemas in ways that reduce Fury to a replaceable subsystem rather than the coordination layer across mixed fleets.

Validation burden: DoD policy requires AI-enabled autonomous systems to be traceable, governable, and tested across their full lifecycle before operational deployment, which means Scout AI's bottleneck may be less model capability than the time and cost required to build safety cases, pass independent validation, and navigate senior-level approval processes for systems operating near lethal mission chains.