Shield AI
Shield AI is an American defense technology company founded in 2015 by brothers Ryan Tseng and Brandon Tseng, along with Andrew Reiter, and headquartered in San Diego, California.[1][2] The firm specializes in developing artificial intelligence software for autonomous operations of aircraft and unmanned systems, with its core Hivemind platform enabling pilotless flight in GPS-denied and contested environments without reliance on constant human input or pre-programmed routes.[3][4] Shield AI's flagship products include the V-BAT, a vertical takeoff and landing unmanned aerial system designed for tactical intelligence, surveillance, and reconnaissance missions, capable of operating from diverse launch platforms such as unmanned surface vessels.[5] The company has achieved milestones such as autonomously piloting the U.S. Air Force's X-62 VISTA aircraft to perform tactical maneuvers against human-piloted jets and operating the MQ-20 Avenger drone using Hivemind.[3] In October 2025, Shield AI unveiled the X-BAT, an AI-piloted vertical takeoff and landing fighter jet intended for penetrating contested airspace, dynamically teaming with manned aircraft, and executing collaborative tactics over extended ranges.[6] The company has secured substantial venture funding, including a $240 million Series F round in March 2025 that valued it at $5.3 billion, supporting expansion of Hivemind as an enterprise autonomy platform for defense applications.[7][8] Shield AI serves clients including U.S. Special Operations Command, emphasizing systems that enhance warfighter safety by reducing exposure to high-risk scenarios through intelligent, edge-based AI decision-making.[3]
Founding and History
Origins and Early Development
Shield AI was founded on June 1, 2015, in San Diego, California, by brothers Brandon Tseng, a former U.S. Navy SEAL officer with combat experience in Afghanistan, Ryan Tseng, an engineer and serial entrepreneur who previously founded WiPower (acquired by Qualcomm), and Andrew Reiter, a specialist in computer vision.[9][10] The origins of the company trace to Brandon Tseng's firsthand observations of the perils faced by service members in urban warfare, including room-clearing operations in GPS- and communications-denied environments, where human-piloted systems proved inadequate for mitigating risks to troops.[9][11] The founders sought to address these challenges by leveraging artificial intelligence to enable autonomous operation of small unmanned systems, with an initial mission to protect soldiers and civilians through intelligent, edge-based autonomy rather than relying on large defense contractors' slower innovation cycles.[10][9] Early development commenced with a team of three, focusing on prototyping AI software for indoor navigation and decision-making independent of external signals. In 2015, the company built its first product prototype, Nova, a quadcopter drone equipped with onboard AI for fully autonomous reconnaissance in confined spaces.[9] Despite initial rejections from Silicon Valley investors skeptical of defense applications, Shield AI raised $800,000 in seed funding in 2016 from Founders Collective, which supported refinements to its core Hivemind autonomy engine.[9] By late 2017, the startup had expanded to under 30 employees while prioritizing technologies that allowed drones to interpret sensor data and execute missions without human intervention or GPS.[12][9]Key Milestones and Expansion
Shield AI secured its first major contract in 2016 from the U.S. Department of Defense's Defense Innovation Unit to develop AI autonomy capabilities.[13] This marked the company's initial entry into military applications, focusing on enabling unmanned systems to operate in GPS- and communication-denied environments.[14] Subsequent growth included significant funding infusions to scale operations and technology. In August 2021, the company raised $210 million in a Series D round to advance AI for autonomous piloting of military aircraft.[15] By December 2023, Shield AI expanded its Series F funding to $500 million, achieving a $2.8 billion valuation and accelerating AI deployment across platforms.[16] In March 2025, an additional $240 million in Series F-1 financing pushed the valuation to $5.3 billion, earmarked for enhancing Hivemind Enterprise, an AI developer platform.[7] Product advancements drove operational expansion, with V-BAT drones supporting U.S. Navy and Marine Corps missions since 2016 and achieving over 23 deployments by 2024, including drug interdictions and resilience testing in Ukraine.[17] The V-BAT Teams product launched in October 2023, enabling swarming autonomous missions deployable in 2024.[18] Hivemind integrations advanced in 2024, including multi-jet AI-piloted flights with the U.S. Air Force and dogfight maneuvers on Kratos platforms.[19] Geographic and capability expansion accelerated through acquisitions and partnerships. In 2024, Shield AI acquired Sentient Vision Systems to integrate ViDAR imaging for enhanced ISR, and partnered with India's JSW Group to bolster Indo-U.S. defense ties.[19] Contracts included U.S. Coast Guard V-BAT services, Japan's JMSDF selection for Group 3 UAS, and Naval Air Systems Command integration of Hivemind.[19] By 2025, European efforts yielded $100 million in contracts and a September autonomous flight demonstration with Airbus in Norway.[20][21] The company grew to over 1,000 employees, earned Inc. 5000 recognition, and reported $750 million in 2024 revenue with 360% year-over-year growth.[11][19][22]Technology and Products
Hivemind AI Autonomy Software
Hivemind is Shield AI's proprietary AI autonomy software stack designed to enable unmanned aerial vehicles (UAVs) and other robotic systems to perform complex missions autonomously in GPS- and communications-denied environments.[23] It functions as an "AI pilot," processing sensor data for perception, decision-making, and control without reliance on human intervention or external signals.[24] First deployed in operational settings in 2018, Hivemind has powered missions including intelligence, surveillance, and reconnaissance (ISR) on platforms like the V-BAT drone.[9] The software incorporates advanced capabilities such as state estimation, multi-agent teaming for coordinated operations, mapping and localization, object detection and tracking, task and behavior planning, and motion control.[25] These features allow systems equipped with Hivemind to navigate dynamic environments, avoid obstacles, and adapt to threats in real-time, even when operating in swarms.[23] Hivemind's edge computing architecture ensures low-latency processing on onboard hardware, prioritizing reliability for military applications where failure could be catastrophic.[26] To facilitate broader adoption, Shield AI offers Hivemind Enterprise, a developer platform including the Hivemind SDK with reference architectures, autonomy libraries, simulation tools, and testing frameworks.[27] This modular toolkit reduces development time for custom autonomy solutions and supports integration across air, ground, and surface domains.[28] In 2025, enhancements like Hivemind EdgeOS introduced middleware for predictable system configuration, efficient inter-process communication, and robust multi-agent orchestration via a specialized messaging language.[26] Demonstrations of Hivemind include its integration on General Atomics' MQ-20 Avenger for combat air patrol flights in July 2025, showcasing swarm autonomy and tactical maneuvering.[29] Partnerships, such as with HII in September 2025 for cross-domain unmanned systems and AWS for fleet-scale deployment via IoT Core and Elastic Kubernetes Service, underscore its expanding interoperability.[30][31] Platform-agnostic by design, Hivemind has been tested on diverse hardware, including vertical takeoff and landing (VTOL) systems like V-BAT and X-BAT, enabling missions ranging from single-unit operations to large-scale teaming.[24][32]V-BAT and Other Drone Platforms
The V-BAT is Shield AI's flagship Group 3 vertical takeoff and landing (VTOL) unmanned aerial system (UAS), optimized for intelligence, surveillance, and reconnaissance (ISR) in contested environments.[5] It employs a ducted fan configuration for transition from hover to fixed-wing flight in under 15 seconds, with up to 45 minutes of hover capability depending on payload.[33] Key specifications include a wingspan of 12.5 feet (3.8 meters), height of 9.6 feet (2.9 meters), and maximum gross takeoff weight of 161 pounds (73 kilograms), powered by a heavy-fuel engine compatible with JP-5.[5] The platform achieves a maximum speed of 56 miles per hour (90 kilometers per hour), endurance exceeding 10 hours (up to 13 hours in some configurations), and a service ceiling of 18,000 feet (5,500 meters).[34] V-BAT supports a payload capacity of up to 25 pounds across three open slots for modular sensors, enabling flexible mission profiles such as electro-optical/infrared imaging or electronic warfare payloads.[33][35] In April 2025, Shield AI introduced an upgraded V-BAT block integrating Hivemind AI autonomy software for enhanced edge computing, satellite communications (SATCOM), and fully autonomous operations without GPS reliance.[36] This version delivers Group 4 and 5 UAS capabilities—such as extended range and heavier payloads—within a compact Group 3 form factor, facilitating logistics simplicity and deployment from austere locations including unmanned surface vessels.[36] The U.S. Coast Guard awarded Shield AI a $198 million contract in July 2024 to produce V-BAT units for maritime ISR, following successful tests demonstrating 20,000-foot altitude operations and integration with naval assets.[37][38] Beyond V-BAT, Shield AI has developed the Nova, a small quadcopter UAS for confined-space operations, such as indoor navigation and swarm tactics, powered by Hivemind for autonomous flight in GPS-denied settings.[39] In October 2025, the company unveiled the X-BAT, an experimental jet-powered VTOL stealth drone designed as an AI-piloted loyal wingman or standalone fighter, featuring a 2,000-mile range, altitudes up to 50,000 feet, and supersonic capabilities without runways.[40][32] X-BAT prototypes are slated for VTOL testing by fall 2026, emphasizing integration with manned aircraft for collaborative combat roles.[41] These platforms collectively advance Shield AI's focus on scalable, AI-driven aerial autonomy for military applications.[3]Supporting Systems and Integrations
Shield AI's Hivemind autonomy software features modular architecture that enables integration with a wide array of hardware platforms, sensors, and command-and-control (C2) systems, allowing it to operate across diverse unmanned systems without requiring custom hardware redesigns.[24] The Hivemind SDK provides developers with libraries, simulation tools, and reference architectures to embed autonomy into existing platforms, supporting rapid prototyping and deployment on edge devices.[27] Additionally, Hivemind EdgeOS serves as middleware to ensure compatibility and reliability in resource-constrained environments, handling real-time processing for perception, planning, and control across heterogeneous fleets.[26] Key integrations include partnerships with major defense contractors to enhance specific platforms. In April 2025, Shield AI teamed with Airbus U.S. Space & Defense to integrate Hivemind onto the MQ-72C Logistics Connector unmanned aerial vehicle, enabling autonomous logistics missions with existing Airbus avionics and payloads.[42] RTX announced a collaboration in July 2025 to incorporate Hivemind into loitering munitions and sensor systems, aiming to deliver networked collaborative autonomy for U.S. and allied forces.[43] In September 2025, demonstrations on Kratos' BQM-177A target drone showcased Hivemind's compatibility with onboard compute, communications, and human-machine interfaces, supporting U.S. Navy manned-unmanned teaming objectives.[44] Further expansions involve cloud and edge computing synergies. A June 2025 integration with Amazon Web Services (AWS) leverages AWS IoT Core for scalable updates and fleet management of Hivemind-enabled systems, addressing challenges in over-the-air autonomy enhancements.[45] In October 2025, a strategic alliance with Parry Labs combined Hivemind with STRATIA software and edge-compute hardware for integrated autonomy at the tactical edge.[46] Hivemind Enterprise further supports seamless interfacing with any C2 or fleet management system via its Commander module, facilitating operator oversight without proprietary lock-in.[25] These integrations extend to counter-unmanned systems and communications. L3Harris integrated Shield AI's Tracker software with WESCAM MX-Series electro-optical/infrared sensors in September 2025 to detect and defeat drone threats passively.[47] Earlier, in September 2023, Shield AI adopted Persistent Systems' Wave Relay mobile ad hoc network as its preferred communications backbone for Hivemind operations, ensuring resilient data links in contested environments.[48] Such efforts underscore Hivemind's role in enabling interoperable, hardware-agnostic autonomy across air, land, and sea domains.[23]Military Applications and Deployments
U.S. Department of Defense Contracts
Shield AI has received numerous contracts from the U.S. Department of Defense branches, primarily for developing and integrating its Hivemind AI autonomy software into unmanned systems and supporting digital engineering for military applications.[49][50] These awards emphasize the company's role in advancing AI-driven solutions for aviation, missile defense, and joint operations in denied environments.[51] In September 2020, Shield AI was awarded a $7.2 million Phase II Small Business Innovation Research contract by the U.S. Army to develop AI capabilities for aviation units, enabling multi-agent unmanned systems to operate autonomously.[49] On October 18, 2021, the company secured a prime position on the Air Force's Eglin Wide Agile Acquisition Contract (EWAAC), a 10-year indefinite-delivery/indefinite-quantity vehicle with a $46 billion ceiling, focused on digital and model-based systems engineering services for rapid prototyping and acquisition support.[50] In December 2021, Shield AI received a $14.99 million firm-fixed-price modification to an existing Air Force contract (FA8649-20-C-0158) for related digital engineering efforts.[51] By July 2022, Shield AI was selected as one of multiple awardees under the U.S. Air Force's $950 million ceiling indefinite-delivery/indefinite-quantity contract for Joint All-Domain Command and Control (JADC2) initiatives, aimed at integrating AI for multi-domain operations.[52] In August 2024, the U.S. Navy's Naval Air Warfare Center Aircraft Division (NAWCAD) awarded Shield AI a contract to integrate Hivemind AI pilot software onto the Kratos BQM-177A target drone, enhancing autonomous flight behaviors for testing and training.[53] Most recently, on May 5, 2025, Shield AI won a U.S. Navy NAWCAD Weapons and Oceanographic Logistics (WOLF) Autonomous Systems Integration Division Broad Agency Announcement contract valued at $3.156 million total, with an initial obligation of $637,516, to advance AI/ML innovations for naval unmanned systems.[54]| Date | Agency | Value | Purpose |
|---|---|---|---|
| Sep 2020 | U.S. Army | $7.2M | Phase II SBIR for AI in aviation units[49] |
| Oct 2021 | U.S. Air Force | Prime on $46B IDIQ | EWAAC digital engineering services[50] |
| Dec 2021 | U.S. Air Force | $14.99M modification | Expansion of digital engineering contract FA8649-20-C-0158[51] |
| Jul 2022 | U.S. Air Force | Part of $950M IDIQ | JADC2 AI integration[52] |
| Aug 2024 | U.S. Navy (NAWCAD) | Undisclosed | Hivemind integration on BQM-177A[53] |
| May 2025 | U.S. Navy (NAWCAD WOLF) | $3.156M total | AI/ML for autonomous systems[54] |