Do What Honor Dictates: Shield AI’s Mission of Building the Most Advanced AI Pilots the World Has Ever Seen!

Former Navy SEAL Brandon Tseng was haunted by a recurring question in 2015: why didn’t his team have better ways to see inside buildings during raids? Every mission brought uncertainty—what lay behind closed doors could be deadly, and the margin for error was slim. These high-stakes operations made it clear that artificial intelligence (AI) and autonomous systems had the potential to revolutionize military missions. What if a small, self-piloted vehicle could scout ahead, entering dangerous buildings or exploring tunnels without risking human lives? What if fighter jets could be flown by AI, transforming them into powerful assets without needing human pilots?

Driven by these questions, Tseng reached out to his brother Ryan, a tech enthusiast who had already built a wireless cellphone charger in college. Together with Andrew Reiter, an expert in AI and software, they founded Shield AI. Their ambitious goal? To build intelligent systems that could protect both service members and civilians. From its inception, Shield AI looked toward the future of military operations, where AI would power a new generation of autonomous systems.

The Birth of Hivemind

The company’s breakthrough came with the development of Hivemind, an AI-powered pilot system designed to operate military aircraft autonomously. Brandon Tseng describes Hivemind as “self-driving technology for aircraft,” enabling mission execution with no need for a remote pilot, communications, or GPS. This game-changing technology allows aircraft to function independently, even in contested environments where traditional navigation tools like GPS are compromised.

Hivemind is built on three core principles: perception, cognition, and action. As Tseng explains, “Every autonomous system is about how it perceives, thinks, and acts.” The system relies on edge computing, performing calculations directly on the aircraft, reducing dependence on cloud resources and ensuring effective operation in high-risk environments.

One of the first products to showcase Hivemind was the Nova Quadcopter, designed to autonomously clear rooms in 60 seconds while creating real-time maps of building interiors. This capability was revolutionary, addressing the very challenge that sparked Tseng’s vision for the company. Nova drones provided operators with vital intelligence through a smartphone interface, allowing potential threats to be annotated before entering hostile environments.

Shield AI’s Role in Combat

In October 2023, following the Hamas attacks in Israel, Shield AI’s Nova 2 drone was deployed to assist Israeli forces. Named the “Most Mission Capable Indoor Drone” by joint US and Israeli forces, Nova 2 proved its worth in combat, helping search for barricaded shooters and trapped civilians. Equipped with 15 onboard cameras and advanced AI, it scouted dangerous environments, saving lives and demonstrating AI’s life-saving potential in modern warfare.

From Ground Combat to the Skies and Seas

While the Nova Quadcopter showcased Shield AI’s strengths in ground operations, Hivemind’s potential was fully realized when integrated into military aircraft, such as the MQ-35 V-BAT drone and fighter jets like the F-16. In 2022, Shield AI conducted the “first-ever autonomous dogfights of F-16s in the real world,” a milestone that displayed Hivemind’s ability to perform complex aerial maneuvers without human intervention. This technology opened the door to new possibilities in autonomous air combat.

The #MQ35 V-BAT is #aerospaceengineering at its best and a great example of first principles engineering.

Featuring a cutting-edge carbon fiber duct, the #VBAT not only achieves vertical takeoff and landing but also sets new standards in aerospace efficiency and design compared… pic.twitter.com/uyvCywRCT1

— Shield AI (@shieldaitech) June 26, 2024

Shield AI also began developing V-BAT Teams, swarms of autonomous drones designed for maritime awareness and enemy air defense suppression. These drone teams offer superior stability in challenging sea conditions, require minimal maintenance, and are extremely versatile. As Tseng highlighted, the V-BAT drone system is in its refinement phase, with deliveries expected in late 2024.

The Future of AI in Aerial Combat

At the heart of Shield AI’s success lies its flagship technology, Hivemind. With a 99% win rate in simulated F-16 dogfights against active-duty Air Force and Navy pilots, Hivemind demonstrated AI’s ability to outperform human pilots in complex aerial scenarios. Beyond dogfighting, Hivemind is being trained to execute a variety of missions, including Beyond Visual Range (BVR) strikes and Suppression of Enemy Air Defenses (SEAD), making it a versatile tool in combat.

Hivemind’s deployment on F-16s serves as a starting point, with the ultimate goal of integrating the system into next-generation aircraft, including F-22s, F-18s, and other fighter jets. Its ability to extend the operational relevance of older aircraft models offers a cost-effective alternative to full fleet replacements, making it a strategic asset for maintaining air dominance.

Commercial and Military Expansion

Though Shield AI’s origins are in military technology, its innovations are beginning to spill over into the commercial sector. The robust autonomy provided by Hivemind is seen as having immense potential in civil logistics, autonomous cargo transport, and infrastructure inspection. V-BAT’s low maintenance, minimal logistics, and expeditionary launch capabilities make it ideal for industries like oil and gas, disaster relief, and even autonomous cargo transportation.

In a significant milestone, Shield AI secured a $198 million contract with the U.S. Coast Guard in 2024, further cementing its status as a leader in autonomous systems. The company has raised over $1 billion in funding, backed by investors like Riot Ventures, and was recognized by Forbes’ AI 50 and Fast Company’s Most Innovative Companies lists.

Ethics and AI in Warfare

Despite the rapid advances in AI-driven military applications, ethical concerns persist, particularly around autonomous weapons. Shield AI has been clear in its stance: lethal decisions should always remain with humans. Tseng has emphasized that their systems are designed to save lives, not replace human judgment in critical situations. This aligns with U.S. military policies, ensuring human oversight in the deployment of autonomous systems.

Looking ahead, Tseng envisions a future where a single individual could command a million AI-powered drones, executing missions autonomously in ways that were once the stuff of science fiction. As AI technology advances, Shield AI is pushing the boundaries of what’s possible in both military and commercial applications.

Paving the Road to Full Autonomy

With plans to acquire Sentient Vision Systems, an Australian company specializing in AI-enabled real-time object detection, Shield AI is doubling down on its vision for the future. 

The World's Best AI Sensors now has the World's Best AI Pilot. We’re excited to announce @shieldaitech is acquiring Sentient* and the continued development of our ViDAR-enabled WAMI solution called “Sentient Observer." https://t.co/JmOicnS5Vn#AUKUS #AI
*pending gov approval

— Sentient Vision Systems (@SentientVision) April 4, 2024

This acquisition will strengthen the V-BAT drone platform, particularly in wide-area motion imagery. The company has ambitious growth plans, with projections of $750 million in annual sales by 2026, positioning itself as a key player in the next generation of AI-driven autonomy.

As Shield AI continues its journey, the company’s mission remains unchanged: to build the most advanced AI pilots the world has ever seen, protecting lives both on and off the battlefield. “You have to build systems that fundamentally do not rely on GPS,” Tseng says, pointing to a future where AI-powered aircraft operate independently, reshaping the very fabric of modern warfare.