Defense Tech Signals

Issue #8: Shield AI

Editor’s Brief1

Last week we covered Saronic Technologies and their new approach to naval power and distributed maritime operations. The number of companies in the autonomous space is growing fast—with a new player, Blue Water Autonomy, emerging from stealth just days ago.

As the DoD accelerates its push toward unmanned systems, a deeper challenge is coming into focus:
How do you integrate autonomy into the heart of military operations—without breaking trust or command?

Shield AI is one of the companies at the center of that question.

As always, your feedback shapes our coverage—reply directly with insights or questions.

Signal Brief: Shield AI - Scaling Autonomy Where GPS and Comms Fail

Shield AI is transforming military aviation with AI-driven autonomy that dramatically reduces dependence on communication links and human operators. Its flagship V-BAT drone—a tail-sitting vertical takeoff and landing (VTOL) platform—paired with the Hivemind AI pilot, can maintain effective ISR even when GPS and comms are jammed.

Origins & Vision

Founded in 2015, Shield AI emerged from battlefield necessity. Brandon Tseng, a former Navy SEAL, teamed up with his brother Ryan, a tech entrepreneur, and Andrew Reiter, an autonomy expert, after seeing how losing drone support in denied environments put ground forces at risk.

From the start, the company has been mission-driven to protect service members and civilians with intelligent, autonomous systems. Its vision is to build “the world’s best AI pilot”—a software-defined autonomy layer that can fly any aircraft and operate in denied environments.

After launching the Nova quadcopter for indoor ISR, Shield AI acquired Martin UAV in 2021 to integrate Hivemind into the V-BAT platform—a unique VTOL design that doesn’t require runways, launchers, or recovery nets.

This experience shaped Shield AI’s core philosophy: autonomy must work at the tactical edge, without relying on GPS or comms. By developing systems that operate independently, they’ve shrunk the sensor-to-shooter timeline from hours to minutes—while staying effective in contested electromagnetic environments.

Key Takeaways

  • Enable Persistent Operation – Shield AI platforms remain effective in GPS-denied, comms-contested environments.

  • Outperform Human Pilots – Hivemind has beaten human F-16 pilots in dogfighting simulations.

  • Expanded Indo-Pacific Footprint – V-BAT adoption by Japan and India reflects growing demand for persistent maritime ISR.

  • Ukrainian Integration – In March 2025, Shield AI began training with Ukraine’s Unmanned Systems Forces and established a presence in Kyiv.

  • Scaling Path Validation – Growth from small innovation contracts to program-of-record status shows the viability of non-traditional defense companies scaling.

Tech Radar: 

V-BAT (MQ-35) – Expeditionary ISR Without Compromise

Shield AI’s flagship Group 3 UAS delivers capabilities that previously required larger platforms with significant ground support.

Key Capabilities

  • Vertical Takeoff and Landing – Operates from confined 20x20 ft spaces, enabling deployment from small vessels or austere locations.

  • Extended Endurance – Heavy-fuel variants offer 13+ hours of flight time with a 40 lb payload capacity.

  • Autonomous Navigation – Operates effectively without GPS using onboard sensor fusion and computer vision.

  • Comms-Independent Operation – Executes full missions without constant communication links.

  • Modular Payload Integration – Supports EO/IR sensors, wide-area motion imagery, SAR, and electronic warfare packages.

In early 2025, Shield AI announced a major V-BAT block upgrade, integrating SATCOM for beyond-line-of-sight operations, a new engine for greater endurance, and further autonomy enhancements via Hivemind.

Hivemind AI Pilot – True Autonomy for Aerial Systems

Hivemind is Shield AI’s core autonomy software, enabling fully autonomous execution without GPS or remote control. Trained through millions of simulated flight hours, it actively perceives, decides, and acts like an experienced aviator.

It’s been integrated into systems beyond V-BAT, including the X-62A VISTA (a modified F-16). In 2024, Hivemind outperformed human pilots in beyond-visual-range engagements—building on its 2020 DARPA AlphaDogfight victory.

A single software stack controls platforms ranging from quadcopters to jets. Recent tests showed multiple Hivemind-controlled Firejet drones coordinating autonomously in flight—no human input required for basic tactical decisions.

ViDAR Pod – Passive, Wide-Area Surveillance

Shield AI’s ViDAR Pod offers stealthy maritime and ground ISR using onboard AI to detect and track targets passively—ideal for contested environments where active sensors would give away position.

Nova – Autonomous Indoor Reconnaissance Platform

The Nova quadcopter first deployed with U.S. special operations in 2018 and has since supported missions in Syria, Afghanistan, and Iraq. In October 2023, Nova 2 helped Israeli forces rescue hostages in Gaza. It autonomously maps buildings, detects threats, and navigates indoors without GPS or remote control.

Market Signals

Funding & Growth

  • Total Funding: $1.3 billion across multiple rounds

  • Latest Round: $240 million (F-1) in March 2025

  • Current Valuation: $5.3 billion

  • Notable Investors: L3Harris Technologies, Hanwha Asset Management, Point72 Ventures, Disruptive, USIT, A16Z

  • Annual Contract Growth: $95M (2022) → $163M (2023), with continued growth through U.S. and international military sales.

Contracts & Government Traction

Shield AI has secured multiple significant defense contracts, including:

  • U.S. Coast Guard – $198.1M award for V-BAT maritime ISR through 2029, with initial deployments on major cutters.

  • U.S. Navy (NAVAIR PMA-281) – Evaluating Hivemind for carrier aviation.

  • Japan Maritime Self-Defense Force – Multi-year commercial sale of V-BAT systems for shipboard ISR—the country’s first ship-launched drone.

  • Singapore Air Force & DSTA – Joint effort to develop Hivemind as a common AI pilot architecture.

  • Korea Aerospace Industries (KAI) – Integrating Hivemind across platforms to accelerate AI-piloted capabilities.

  • Booz Allen Hamilton – Partnering on autonomous airborne solutions for the U.S. military.

  • L3Harris Strategic Teaming – Co-developing missionized Group 5 UAS using Hivemind in contested airspace.

Looking Ahead

Shield AI’s partnerships across the Indo-Pacific place it near the center of great power competition. But this week, we’re stepping away from the regional map and focusing on a more fundamental challenge: human trust in AI autonomy.

Despite AI outperforming pilots in simulations, trust isn’t built in a lab. Aviators build it through multiple flights, cowboy time in the ready room, and late-night Mid Rats. An AI wingman doesn’t do any of that—it’s expected to be trusted from day one.

To be clear, this isn’t just a Shield AI problem. Companies like Anduril, General Atomics, and Saronic will face the same challenge: integrating machines into a culture built on human trust.

That challenge is magnified by how lethal force is governed. Authority flows from the President down through field commanders—who remain bound by the laws of armed conflict. And no AI system currently fits in that chain of command.

Meanwhile, Chinese researchers are developing AI that can predict and counter human pilots by analyzing aircraft control surfaces. That makes trusted, scalable human-machine teaming not just a tactical advantage—but a strategic requirement.

To cross that gap, AI developers must build behavioral transparency into their autonomy stack—tools that help operators understand, anticipate, and trust AI decision-making. That includes transparent planning models, consistent behavior in edge cases, and post-mission debrief tools that explain what the AI “saw” and why it acted.

At the same time, Co-founder Brandon Tseng is unequivocal on use of force: “That is a human decision and it will always be a human decision.” Shield AI isn’t trying to replace people—it’s trying to scale them. If Shield AI can solve for trust as well as performance, it won’t just field better AI—it will help shape how the U.S. fights.

And it’s well on its way. The company has achieved impressive traction through multiple programs of record and international deployments. Its partnerships with legacy primes show a clear focus on fielded capability—working with those who can get systems across the finish line, not just a funding round.

Challenges

  • Adoption Risk vs. Trust Gap – Even with superior performance, cultural resistance and perceived risk could delay procurement and operator confidence.

  • Legal & Command Constraints – Integrating AI into the military chain of command requires not just better tech—but doctrinal change.

  • AI Behavior Transparency – Without explainable reasoning, full autonomy will remain opaque—limiting their use in high-stakes missions.

Bottom Line:

Shield AI is more than a prototype success story—it’s a blueprint for how modern defense companies scale, partner smartly, and deliver capability that matters. With programs of record, fielded systems, and trusted integration with legacy primes, it’s one of the most mature autonomy players in the defense space.

As autonomy becomes more capable, the next step is building the trust needed to integrate it into real-world operations. That means more than performance in controlled environments—it means making the system understandable, predictable, and accountable in the field.

Codifying digital rules of engagement, red-teaming edge cases, and designing explainable behaviors are all part of the answer. But so is the broader challenge of helping operators feel confident not just in what the system does—but in why it does it.

Autonomy is here. Trust is next.

Stay Ahead of Defense Innovation

The Future of Defense Tech Starts Here. Get exclusive insights on emerging warfare tech, cutting-edge startups, and the evolving defense industrial base—delivered straight to your inbox.

1  The views expressed in this newsletter are my own and do not represent the views of the U.S. Navy, Department of Defense, or any government agency. Mention of companies, technologies, or products is not an endorsement or recommendation. The content is for informational purposes only and should not be considered investment advice.

Reply

or to participate.