Shield AI’s V-BAT Already Has a Weapons Deal — So Why Is Co-Founder Brandon Tseng Still Arguing Against Armed Drones?

The V-BAT has flown more than 60 miles into Russian-controlled airspace through layered GPS jamming, spotted a Buk-M1 surface-to-air missile system, and handed targeting data to HIMARS crews who destroyed it. That mission — pure intelligence, zero ordnance on the drone itself — is exactly the point Shield AI co-founder Brandon Tseng wants people to understand. The tension is that Shield AI just signed a weapons deal anyway.

Key facts:

• Shield AI co-founder and president Brandon Tseng told Business Insider that experienced militaries don’t ask him to arm the V-BAT because they already know how to use joint fires.
• Shield AI has simultaneously announced a partnership with South Korean arms manufacturer LIG Nex1 to equip the V-BAT with six-pound guided missiles — contradicting the “nobody wants this” framing.
• The V-BAT is a vertical takeoff and landing (VTOL) drone that flies autonomously in GPS-denied, comms-jammed environments using Hivemind AI software.
• Ukraine used it for intelligence, surveillance, and reconnaissance (ISR) — and stopped requesting an armed version once operators understood how joint targeting works.

The V-BAT’s Role in Ukraine Is Intelligence, Not Strikes

The V-BAT is a single-rotor VTOL drone roughly 8 feet long with a 9-foot wingspan, weighing 88 pounds and capable of flying for 12 hours and covering 600 miles without GPS or pilot input. Shield AI’s Hivemind software handles navigation entirely when jamming cuts off radio control and satellite signals — which on the Ukrainian front happens constantly. We covered the V-BAT’s Ukraine deployment in late 2024, when it withstood seven simultaneous jammers during testing near Kyiv over two days.

Tseng, a former Navy SEAL, told Business Insider the armed-drone question comes up often — mostly from buyers who don’t yet grasp how modern militaries actually deliver fires. His answer is consistent: the V-BAT finds targets. Something else kills them.

“Who doesn’t ask for that? The US military doesn’t ask for that because we understand joint fires. The Ukrainians don’t ask for it anymore, either,” Tseng said.

That distinction matters operationally. An armed ISR drone carries weapons weight at the cost of endurance, range, and sensor payload. A pure ISR platform like the V-BAT can stay on station longer and carry better optics, feeding targeting data to artillery, HIMARS, or manned aircraft that can deliver far more destructive force than a small drone warhead. Ukraine proved this — the V-BAT spotted the Buk-M1 surface-to-air missile system, HIMARS destroyed it.

Shield AI Has Already Signed a Weapons Deal

Tseng’s ISR-first argument sits alongside a concrete commercial reality: Shield AI has announced a partnership with South Korean arms manufacturer LIG Nex1 to equip the V-BAT with six-pound guided missiles. This isn’t exploration — it’s a signed deal. The co-founder can argue that mature militaries don’t need armed V-BATs all he wants, but the company is building one for customers who do.

That’s not hypocrisy. It’s the defense business. Some buyers — smaller militaries, allied forces still developing ISR-to-strike integration — don’t have the joint fires infrastructure that lets a pure ISR platform shine. For them, an armed V-BAT capable of delivering a guided munition is more useful than a platform that hands off to a HIMARS battery they don’t have. Shield AI is building for both customer types.

The V-BAT has expanded well beyond Ukraine. Greece deployed V-BATs to watch its eastern borders in the Aegean, with the Hellenic Army using them for persistent surveillance from the Evros River to the Greek continental shelf. The US Coast Guard put the V-BAT through four days of ship-based trials in late July 2025, scoring 100 percent on all key performance parameters aboard the NSCs Midgett and Stone. Shield AI also partnered with India’s JSW Defence to manufacture V-BATs locally. Greece, the Coast Guard, and India aren’t asking for missiles. The LIG Nex1 deal serves a different customer set entirely.

Hivemind Is the Product, Not the Airframe

Shield AI’s core commercial argument has always been its Hivemind autonomy software, not any specific airframe. Hivemind is what lets the V-BAT navigate without GPS, maintain mission parameters when communications drop, and keep flying in conditions that ground or crash conventional drones. The V-BAT is the current delivery vehicle. The X-BAT, unveiled in October 2025, scales Hivemind to a $27 million jet-powered fighter-class drone capable of supersonic flight and a 2,000-nautical-mile range.

That software-first framing is why Tseng keeps pushing back on the armed-drone demand even as his company signs weapons deals. Adding missiles to a V-BAT makes it a modestly better strike drone for one customer type. Perfecting Hivemind makes every platform it flies — from the V-BAT to autonomous fighters — more capable across every mission. The LIG Nex1 deal is a product line extension. Hivemind is the company.

Shield AI was founded by Brandon Tseng and his brother Ryan Tseng. The company has raised over $1 billion and carries a valuation near $3 billion. We’ve tracked Shield AI’s trajectory since 2024, when the company first outlined its vision of AI software as the defining edge in modern air warfare.

DroneXL’s Take

Tseng’s ISR argument is genuinely sound — and it’s also a sales pitch, and those two things aren’t mutually exclusive. The instinct to arm every drone comes from a reasonable place: if the drone is already over the target, why not let it shoot? The answer, which Ukraine learned through hard experience, is that endurance is payload. Every pound of warhead is a pound fewer of battery, sensor, or fuel. A V-BAT that stays airborne for 12 hours and hands precise targeting data to a HIMARS battery is more lethal to the enemy than the same airframe carrying a grenade it can drop once.

That’s real doctrine. US military targeting has operated on that principle for decades. The V-BAT’s Buk-M1 mission is the cleanest recent example I’ve seen of it working on a modern battlefield against modern air defense.

But Tseng is also the president of a company that just signed a guided-missile deal with a South Korean defense firm. So when he says “experienced militaries don’t ask for armed drones,” he means experienced militaries with mature joint fires infrastructure. The LIG Nex1 partnership is for everyone else. Shield AI isn’t turning down revenue — it’s segmenting its market while keeping its core pitch intact.

What I’ll watch: the first operational deployment of an armed V-BAT variant, and which customer gets it. If it goes to a NATO ally with strong C2 infrastructure, Tseng’s framing holds — they’re filling a specific gap, not replacing ISR doctrine. If it goes to a smaller military without that infrastructure, it tells you the armed V-BAT is becoming the default product regardless of what the co-founder argues on the philosophy circuit. Expect that first armed deployment within 12 months of the LIG Nex1 deal closing.

Editorial Note: AI tools were used to assist with research and archive retrieval for this article. All reporting, analysis, and editorial perspectives are by Haye Kesteloo.