The rapid progress of artificial intelligence has made it a focus of the American military and intelligence services. Historically, innovation in the defense industry has been led by hardware. However, the next generation of warfare will likely be driven by software, particularly artificial intelligence, working in concert with hardware. AI systems can provide operators with real-time intelligence and situational awareness by enabling rapid processing and analysis of vast amounts of data, enabling operators and troops to make more informed decisions that could give them tactical advantages over adversaries. Furthermore, autonomous software integrated into military systems such as aircraft, drones, and ground vehicles could improve reconnaissance, logistics, and combat missions without risking human lives. Driven by these dynamics, the artificial intelligence defense market is growing quickly. It was valued at $6.4 billion in 2021 and is expected to reach $13.2 billion by 2028.

AI integration with the military may also dramatically reduce human casualties and enhance the safety of military personnel. As AI applications continue to improve and become increasingly complex, the need for soldiers on the front lines may become reduced. Conflicts like the Russo-Ukrainian War have underscored the importance of AI technology on the battlefield, demonstrating a shift in the technological stack for warfare. Despite Russia's significantly larger defense budget, technologies such as drones, AI-based targeting, and imagery intelligence have enabled Ukraine to withstand Russian forces. In March 2023, the Biden-Harris Administration submitted a budget request to Congress of $842 billion for the Department of Defense, an increase of $26 billion over FY 2023 levels and $100 billion more than FY 2022. Over the past 30 years, the defense industry has consolidated, shrinking to 5 large companies: Lockheed Martin, Boeing, Raytheon, Northrop Grumman, and General Dynamics. This consolidation has not only likely inflated prices but has possibly stifled innovation. However, growing pressure may prompt the government to commit to investing more in AI and could foster a greater willingness to work with and invest in defense tech startups such as Shield AI.

Shield AI is a defense technology company that builds AI systems, starting with fully autonomous quadcopters that collect data in dangerous environments and protect military personnel and civilians. Shield AI’s Hivemind software is an AI pilot for military and commercial aircraft that enables aircraft to perform missions ranging from room clearance to penetrating air defense systems and dogfighting F-16s. Hivemind employs algorithms for planning, mapping, and state estimation to enable aircraft to execute dynamic flight maneuvers and uses reinforcement learning for the discovery, learning, and execution of winning strategies. The company has contracts with the Department of Defense and the Department of Homeland Security.

Shield AI was founded in 2015 by Brandon Tseng (President, CGO), Ryan Tseng (CEO), and Andrew Reiter (ex-CTO, Technical Fellow) to be the next-generation aerospace and defense technology company. As a Navy SEAL, Brandon Tseng experienced firsthand problems where intelligent autonomy could have made life-saving differences. He served two tours in Afghanistan and other deployments in Asia and the Pacific. He saw firsthand that soldiers were risking — and sometimes losing — life and limb when entering compounds and buildings with little information about what lay around the next corner or behind the next door.

After serving seven years in the Navy, Brandon Tseng left the service, went home, and began transitioning to civilian life. While overseas, Brandon Tseng spent time working with sensors and inexpensive computers. He identified a problem that he felt was solvable: the challenging task of physically searching buildings that had been a persistent problem for soldiers during the urban warfare of the post-9/11 era. In 2015, after receiving an MBA at Harvard Business School, Brandon co-founded Shield AI. The other co-founders were his brother Ryan Tseng, who had previously founded a wireless charging company acquired by Qualcomm, and Andrew Reiter, an engineer who had worked on computer vision applications for Draper Labs, a non-profit advanced engineering company. Together, the co-founders embarked on building a prototype for Nova, the company's flagship drone.

Hivemind

Hivemind, Shield AI’s flagship product, is an autonomous software stack that is a self-driving platform for military and commercial aircraft. Hivemind is an AI pilot that reads and reacts to the field, similar to a human pilot. With Hivemind, aircraft use sensors to create a map and then follow the instructions given by the "autopilot,” eliminating the need for GPS, waypoints, or communication systems. Shield AI calls this a “top gun for every aircraft.”

Hivemind employs planning, mapping, and state estimation algorithms to enable aircraft to execute dynamic flight maneuvers. It utilizes reinforcement learning for the discovery, learning, and execution of winning tactics and strategies. On an aircraft, Hivemind can 1) enable full autonomy and act as an AI pilot that can control the aircraft’s flight, navigation, sensors, weapons, and communication systems without relying on a GPS or human operations and 2) enable swarming capabilities, where multiple aircraft can coordinate their actions and share information autonomously.

The Hivemind ecosystem is made of three components that work together for simulation and autonomy:

1. Hivemind Edge consists of software that delivers intelligent autonomy within diverse data-denied and GPS-denied environments. It is platform agnostic and can integrate with third-party systems. Edge enables machines to execute complex, unscripted tasks without direct operator inputs. The application is edge-deployed, with all processing and computation occurring without relying on a central intelligence hub.

2. Hivemind Commander consists of hardware and software designed to support human interaction with the Hivemind ecosystem. It develops interfaces and methodologies for mission planning, command and control, training, and human factors considerations across Hivemind.

3. Hivemind Design consists of software intended for AI developers and designers to speed up capability development and testing for autonomous systems using highly customized and optimized local and cloud computing resources. It provides AI capability design, configuration, data management, analysis, testing, and reporting services.

According to Shield AI, Hivemind is the first & only autonomous AI pilot deployed since 2018. It is used to power aircraft and UAV (unmanned aerial vehicle) missions ranging from dogfighting to penetrating air defense systems. Shield AI’s long-term strategy is to integrate Hivemind with various platforms and missions across the military landscape, such as fixed-winged drones, combat jets, ships, satellites, and submarines.

Source: Shield AI

Nova 2

The Nova 2 is a small quadcopter drone that can be deployed by a digital scout. Powered by Hivemind, the Nova 2 can operate autonomously without needing a GPS or a human operator. Nova 2 can clear a building, detect threats and auto-populate them on a map while vibration notifications alert the user, providing real-time information for those entering dangerous environments. It can also build 3D multi-story maps, providing a common operating picture of the battlefield.

By using the Nova 2 to scope out a building before they go in, soldiers can better understand the layout of the building and identify potential threats, such as traps or enemy combatants. This can help them plan a more effective and safer approach, potentially reducing the risk of casualties and increasing the chances of a successful mission.

Nova 2 is not armed, but Tseng says that the company does not object to equipping it with weapons in the future. Shield AI’s Nova 2 drones have been used by the U.S. Special Operations Command since 2018. Its previous generation drone, Nova 1, was the first AI-powered drone deployed for defense in US military history. US and Israeli forces have also named it the most “mission-capable indoor drone” worldwide.

Source: Shield AI

V-Bat

V-Bat is Shield AI’s vertical take-off and landing (VTOL) unmanned aircraft system (UAS) that will soon be equipped with its Hivemind software. The V-Bat drone was first developed and manufactured by Martin UAV, a US-based aviation technology company; however, in July 2021, it was announced that Shield AI would acquire and continue to develop and produce the V-Bat.

The V-Bat drone has a flight ceiling of 15K feet and can be deployed for various missions, including force protection, aerial patrol, target search and detection, surveillance, and more. One key feature that sets the V-Bat apart from other Vertical Takeoff and Landing (VTOL) drones is its ducted fan design, which Shield AI says boosts the drone's thrust by more than 80% and allows for greater control authority in harsh conditions such as rocky terrain.

V-Bat can launch and recover from a hover, fly up to 10 hours in horizontal flight, and make mid-flight transitions to "hover and stare" anytime during a mission. Additionally, the V-Bat's has long-endurance capabilities. Standing roughly 10 feet tall with a 10-ft wingspan, V-Bats can be fully deployed by a team of 2 and is able to pack to fit in a minivan, the back of a pick truck, or a helicopter such as the UH-60 Blackhawk.

The V-Bat has been in operation since 2018 in support of the US Southern Command (SOUTHCOM), the US Coast Guard, the US Marine Corps, and multiple International customers. Shield AI’s long-term vision with V-Bat is to develop a swarm of drones and place them on every US and allied naval vessel in the world. A swarm of V-Bat drones could create an intelligence, surveillance, and reconnaissance network over battlefield areas that could transform situational awareness and data-sharing levels.

In an interview with Breaking Defense in 2022, Brandon Tseng outlined the company's plan to have three to five V-Bat drones ready for operations and production by the end of 2023 and to gradually increase the size of the swarm to about 10 V-Bats in 2024 and up to 30 drones in 2025. V-Bat’s modular nature enables it to be configured with various payloads giving its swarms flexibility to carry out multiple missions and to be rapidly figured.

Source: Shield AI

Customer

Shield AI was founded to serve the US military and its allies by providing AI technology to enhance the safety and effectiveness of troops in the field. Shield AI has worked with several departments of the US military, specifically the US Air Force, the US Army, the US Special Operations Command (SOCOM), and the Defense Innovation Unit (DIU). The Royal Canadian Navy and Brazil’s Armed Forces are also among its clients.

Shield AI also partners with major defense contractors like Boeing, Textron Systems, and Northrop Gruman to integrate its AI technologies into proven military platforms, allowing it to quickly penetrate the federal market and further scale its operations with the Department of Defense. These partnerships offer mutual benefits: Shield AI can leverage the resources and expertise of larger primes to expand its reach, while the primes can integrate innovative technologies without having to develop them in-house.

Market Size

The US is the largest military spender in the world by a large margin, and the US military budget makes up the largest share of the discretionary US federal budget spend. In 2022, the US spent $727 billion on the Department of Defense (DoD). Shield AI operates in two markets: autonomous software and UAS (unmanned aircraft system) platforms.

In 2020, the average total investment for autonomy and AI-related programs within the US Armed Services and DARPA (Defense Advanced Research Projects Agency) was $5.2 billion over three years. The FY2021 US defense budget request allocates $1.7 billion to autonomous technologies to enhance “speed of maneuver and lethality in contested environments” and the development of “human/machine teaming,” as well as $800 million to AI, the market served by Shield AI. AI research as a whole was allocated between $7.4 - $9.1 billion from FY2018 to FY2024. In addition, the Air Force is receiving new funding for a Collaborative Combat Aircraft (CCA) program for investment into autonomous software and drone research & development.

Competition

Anduril*: Anduril is a defense technology company founded in 2017. It builds technology for military agencies and border surveillance and has raised $2.2 billion in funding. It works with the US and its allies to develop autonomous defense solutions such as surveillance drones and the software needed to operate them. It is one of the most prominent defense-tech startups in the US, with a valuation of $8.5 billion. Shield AI competes with Anduril in both hardware and software. Like Shield AI, Anduril manufactures UAS, which includes Ghost, ALTIUS, and Anvil drones. However, the two companies take different approaches in terms of their software. Anduril Industries’ drones are powered by Lattice OS, a complex autonomous AI software that connects sensors and platforms to create a real-time situational awareness system for all of their military hardware. By contrast, Shield AI’s Hivemind software is a full autonomy stack designed specifically for aircraft. While Anduril's Lattice OS provides an all-encompassing operating system for the military, Shield AI's Hivemind offers a tailored solution for UAS autonomy.

Skydio: Skydio is a California-based company founded in 2014 that specializes in autonomous drones for government, commercial, and consumer use. It has raised $570 million in funding. Its drones are designed to be easy to use, safe, and able to fly without requiring a pilot to operate them. Skydio drones are equipped with advanced computer vision and artificial intelligence technology, Skydio Autonomy, that allows them to navigate and avoid obstacles. Drones are also used in construction, agriculture, and public safety industries. The company says its drones are used in every branch of the DoD, by half of all US state departments of transportation, and over 200 public safety agencies. In February 2023, Skydio raised a $230 million Series E at a $2.2 billion valuation. The company says it's seen 30x growth over the past 3 years and that it is the largest drone manufacturer in the US. Like Shield AI, Skydio manufactures drones used in the military for reconnaissance, powered by its autonomous software system Skydio Autonomous.

AeroVironment: AeroVironment is an American defense contractor that designs, develops, and manufactures a portfolio of intelligent, multi-domain robotic systems. It was founded in 1971. It is the leading manufacturer of unmanned aircraft systems, tactical missile systems, and unmanned ground vehicles used for the US military and the top supplier of small drones to the military. Like Shield AI’s drones, AeroVironment’s small drones are used for surveillance, tactic intelligence, and reconnaissance. Unlike Shield AI’s drones, these small drones are not fully autonomous and are hand-launched. The company’s Raven drone is the world's most widely deployed unmanned aircraft system. In 2022, AeroVironment reported full-year revenue of $446 million and an adjusted EBITDA of $62 million. As of March 2023, its market capitalization is $2.2 billion.

General Atomics: General Atomics, founded in 1955, is an American defense and energy company that specializes in research and technology development. Its affiliate, General Atomics Aeronautical Systems (GA-ASI), is the manufacturer of a series of unmanned aircraft and provides electro-optical, radar, and related mission systems solutions. GA-ASI is the developer of the well-known MQ-9A Reaper drone, an upgrade from their previous Predator drone, and is used for both surveillance and strike missions. The software that powers General Atomic’s drones is based on flight control, ground control, and data processing systems, while Shield AI's Hivemind software leverages AI and machine learning algorithms to enable autonomous navigation and decision-making capabilities.

Shield AI makes software and hardware for drones and other autonomous aircraft used by the US military and other government organizations. Its revenue streams consist of contracts with the government, drone sales, and licensing Hivemind software to third-party aircraft. The details of pricing are not disclosed publicly by the company.

Shield AI currently has contracts with the DoD (Department of Defense) and its allies and is actively supporting operations in the field. Its contracts include a $7.2 million Small Business Innovation Research (SBIR) contract to enable the development of AI capabilities for US Army Aviation units and a prime position on a multi-billion dollar Eglin Wide Agile Acquisition Contract (EWACC). While the details of this contract are undisclosed, the contract has a 10-year ordering period with a ceiling value of $46 billion.

Shield AI was also rewarded a maximum amount awarded through the Air Force’s Strategic Funding Increases (STRSTFI) program to accelerate the integration of Hivemind into various military and commercial platforms, and an indefinite-delivery/indefinite-quantity contract with a ceiling amount of $950 million from the US Air Force as part of their Joint All Domain Command and Control (JADC2) efforts. They were also recently selected to receive funding from the DoD’s APFIT Pilot Program, which aims to get new technologies into the hands of warfighters quickly. The funding will go into a project involving the initial purchase of Shield AI’s V-Bat. With these contracts, Shield AI can provide advanced AI capabilities to the military and its allies, helping them achieve their missions effectively and efficiently. Furthermore, Boeing and Shield AI have signed a memorandum of understanding to explore strategic collaboration in autonomous capabilities and artificial intelligence on current and future defense programs.

In December 2022, Shield AI raised $165 million in a Series E funding round, which brought the company's total funding to $573 million and increased its valuation to $2.3 billion, nearly doubling its value in less than a year. Previous investors include Snowpoint Ventures, Riot Ventures, Disruptive, Homebrew, Point72 Ventures, Andreessen Horowitz, Breyer Capital, and SVB Capital.

This fundraising success could be seen as a testament to Shield AI’s product successes and building a business with strong fundamentals. It is also indicative of a larger trend of increased investment in defense technology companies like Anduril Industries*, another market leader that recently raised a $1.5 billion Series E at a $7 billion pre-money valuation.

Software Backbone of the Military

Shield AI is aiming to become a leading defense contractor for the 21st century, focusing on providing artificial intelligence and autonomous solutions for defense applications. By actively partnering with leading aircraft manufacturers and integrating its AI pilot technology into established aircraft and unmanned systems, Shield AI has the opportunity to become the software backbone for unmanned assets — aircraft, ground vehicles, ships, boats, submarines, etc.— throughout the military. By successfully integrating Hivemind into an established platform like the F-16, Shield AI has demonstrated its ability to bring its AI pilot software stack to existing military systems, transforming its capabilities and enhancing its operational effectiveness.

Acquisitions & Partnerships

Shield AI’s goal is to power the next generation of aircraft with Hivemind, intending to expand its AI pilot to larger military platforms through organic growth and acquisitions. In 2021, Shield AI acquired Heron Systems, a Maryland-based defense contractor renowned for its use of reinforcement learning. Heron developed an AI pilot that was able to defeat an experienced Air Force F-16 pilot in five rounds of simulated air combat during the DARPA AlphaDogfight Trials. This strategic acquisition of Heron enables Shield AI to accelerate the deployment of its AI-pilot technology in both legacy and future aircraft.

Moreover, in 2021, Shield AI entered an agreement to acquire Martin UAV, an aerospace company best known for its industry-leading VOTL unmanned aircraft, the V-Bat. In its press release, Shield AI emphasized that the acquisition of Martin UAV is part of its deliberate strategy to integrate Hivemind into unmanned systems by building a portfolio of systems via vertical integration and strategic partnerships. By acquiring companies with complementary products and technologies, Shield AI can strengthen its position in the industry and extend its offerings to cater to a wider range of applications.

Expansion into Commercial

Beyond the military application of Shield’s AI pilot software, there is potential for success in the commercial space. Shield AI is focusing on establishing its product and brand within the US military and its allies. The commercial aviation sector is highly regulated; however, once Shield AI’s concepts are proven in the military, the capabilities can be translated into the commercial aviation sector.

Furthermore, Shield AI's Hivemind software has the potential to integrate not only with military assets but also with eVTOL (electric vertical takeoff and landing) to create a commercial market for short-distance air transport. The economics of flying taxis is challenging, as employing two or even one pilot per aircraft would be economically unfeasible. Flying taxis can only become a reality with highly reliable autonomous systems. Brandon Tseng has stated that Shield AI has been in talks with several companies in the eVTOL space. By venturing into the commercial sector, Shield AI could expand its market. The US commercial drone market was estimated at $29.9 billion in 2022 and has an expected compound annual growth rate of ~39% from 2023-2030. This expansion could position Shield AI as a key player in both military and commercial applications of AI pilot technology.

Ethical Concerns of AI

Ethical concerns around the development and use of AI in military operations and autonomous drones have been on the rise. As AI technology advances, so do ethical concerns and public debate, particularly regarding the potential for autonomous weapons and the loss of human control over life and death decisions. One such example is the "Stop Killer Robots" campaign, a global coalition of NGOs that seeks to ban fully autonomous weapons, highlighting the dangers of delegating critical decisions to machines without human intervention. Ethical concerns around AI could negatively impact Shield AI’s reputation, funding, and partnerships as public opinion and regulatory pressures could lead to restrictions on the development and deployment of AI-powered military technology.

Competition

The market for AI and autonomy in aviation is becoming increasingly competitive, with numerous startups and established companies entering and operating in the space. Shield AI faces the challenge of staying ahead of competitors in terms of technological innovation, performance, and market share.

Reliant on government contracts and partnerships

Shield AI's growth strategy relies heavily on partnerships with other defense contractors, aircraft manufacturers, and acquisitions of complementary companies. Failure to forge successful partnerships or integrate acquisitions effectively could hinder the company's growth. Additionally, Shield AI has relied on government contracts, which have historically moved very slowly. Militarily vetting process moves extremely slowly compared to the breakneck speeds that we are seeing in AI development.

Shield AI is a defense-tech company that focuses on developing advanced artificial intelligence solutions for military applications. The company's goal is to create superior military technology that can save lives by integrating AI and autonomous software into military systems, such as aircraft and drones. Its AI pilot software, Hivemind, is currently powering several aircraft and drones.

*Contrary is an investor in Anduril Industries through one or more affiliates.