#247 Brandon Tseng – Shield AI’s X-BAT: The First AI Fighter Jet to Outsmart Top Gun

Key Takeaways

Brandon Tseng, Shield AI co-founder, transitioned from Navy SEAL to defense tech leader.
Shield AI develops AI-piloted autonomous systems for military and national security applications.
V-BAT drones are operationally deployed, assisting in counter-drug missions and in Ukraine.
The XBAT, an AI-piloted VTOL combat jet, aims to revolutionize air warfare and deterrence.
Autonomous systems are projected to reduce casualties and enhance strategic military capabilities.

Shield AI's V-BAT aircraft are used by the U.S. Coast Guard and Navy's 4th Fleet for counter-drug missions in the Caribbean.
These operations have successfully interdicted over $500 million worth of cocaine.
The V-BAT is described as a miniature Reaper/Predator drone, providing intelligence, surveillance, and reconnaissance (ISR).

Brandon Tseng's interest in becoming a Navy SEAL began at age 10, influenced by movies like 'Under Siege'.
He attended the Naval Academy, a rigorous military institution, confirming his desire to become a SEAL in 10th grade.
The competitive selection for Naval Academy graduates involved a junior year 'SEAL screener' and a three-week SEAL cruise.
Initial non-selection for SEALs led him to serve as a surface warfare officer before eventual acceptance into BUD/S.

Tseng described Basic Underwater Demolition/SEAL (BUD/S) training as demanding but enjoyable, appreciating instructors' humor.
He personally struggled with the tread water exercise with fins, failing it four times before successfully passing.
Hell Week involved ruthless standards enforced by boat crews and grueling evolutions like 'camp surf,' causing severe chafing.

Tseng's first Afghanistan deployment in 2015 involved augmenting a DEVGRU troop, coordinating airstrikes, and conducting direct action raids.
He observed the challenges of working with Afghan Commandos, noting a 10:1 force ratio requirement and varied motivations.
His combat experiences, coupled with entrepreneurial aspirations, motivated him to found Shield AI to solve critical warfare problems with AI.

Shield AI, founded in 2015, aims to protect service members and civilians by developing AI systems for unmanned platforms.
Their 'AI pilot' technology enables autonomous operations without GPS or communications and facilitates swarming capabilities.
The initial vision was for AI and autonomy to power every military asset by 2035.

Shield AI's first product was an AI-piloted quadcopter designed for autonomously clearing buildings of threats, deployed in Syria, Iraq, and Ukraine.
A 2017 'John Henry test' demonstrated the quadcopter clearing a house faster than a SEAL squad, despite initial flaws.
These approximately two-and-a-half-pound drones were used by SEAL teams and the Army, offering a safer alternative for dangerous close-quarters combat.

Warfare in 2035 is envisioned as heavily reliant on autonomous systems, which will augment human roles rather than fully replace them.
The widespread adoption of drones is seen as making traditional military assets and human roles, like pilots, obsolete, leading to reduced friendly casualties.
Long-term, a 'robot-on-robot deterrence' is predicted, significantly reducing human involvement in conflicts.
Autonomous systems are considered the next major strategic technology, enabling 'near-infinite, intelligent, maneuverable mass' for deterrence.

The Shield AI V-BAT is an 180-pound vertical takeoff and landing (VTOL) aircraft designed for ISR and targeting, costing less than larger drones.
VTOL capability eliminates reliance on vulnerable runways, allowing flexible deployment from diverse locations, such as sprinter vans in Ukraine.
Its AI-piloted system, HiveMind, enables effective operation even with GPS and communications jamming, crucial in contested environments.
Shield AI develops both the hardware and software for these systems, ensuring seamless integration of the AI pilot.

Shield AI, through its acquired company Heron Systems, won the DARPA Alpha Dogfight competition, defeating all other AI pilots in simulated F-16 dogfights.
A DARPA-funded simulation further showcased Shield AI's AI pilot outperforming a Top Gun instructor pilot with a 5-0 score.
These results highlight the advanced coordination and execution capabilities of their AI, demonstrating the need for investment in AI for future uncrewed fighter jets.

The XBAT is Shield AI's first AI-piloted, vertical takeoff and landing (VTOL), multi-role combat strike jet platform, eliminating the need for runways.
It is projected to cost $27.5 million with a $6,000 hourly operating cost, significantly less than an F-35's estimated $200-250 million cost and $35,000 hourly operation.
XBAT prototypes are expected to fly by 2026-2027, with production slated for 2029, aiming for 300-500 units annually.
The XBAT has a 2,100 nautical mile range with payload, can carry four AMRAMs internally and two medium-range missiles, and features electronic warfare capabilities.

Shield AI's V-BAT drone has flown over 130 sorties in Ukraine, providing crucial long-range ISR and targeting capabilities.
Initial V-BAT deployments faced failures under GPS and communication jamming conditions until the AI pilot, HiveMind, was re-integrated.
Following successful extensive electronic warfare tests in 2024, Ukrainian forces are now effectively using V-BATs, conducting 4-6 strikes daily and destroying hundreds of millions in Russian equipment.
The primary threat to these drones in Ukraine is kinetic, such as S-300 surface-to-air missile systems, which is viewed as an acceptable trade-off given the cost differential.

Tseng suggests autonomous systems like the XBAT will initially augment human-led operations, similar to how self-driving cars assist drivers.
Over decades, with continued AI innovation, these systems are expected to evolve and replace some existing military platforms, co-mingling with human pilots.
Mission requirements will dictate scenarios where manned aircraft, such as F-35s or F-22s, could operate alongside unmanned systems like XBATs.
Politicians and senior Department of Defense leadership are supportive of collaborative combat aircraft, recognizing their potential to redefine air warfare by removing runway dependency.