Shield AI completes its first X-BAT prototype, an autonomous, stealthy combat drone capable of vertical takeoff and landing. Performance, missions, and program outlook.
Summary
On November 3, 2026, Shield AI unveiled the first full-scale prototype of its X-BAT, an autonomous combat drone engineered for vertical takeoff and landing (VTOL) without needing a runway. Built in twelve months at the company’s Frisco, Texas facilities, the aircraft is now slated to transition to its testing center in Newton, Kansas.
Powered by a General Electric F110 turbofan engine equipped with thrust vectoring, the X-BAT promises a range exceeding 3,700 km, low-observable stealth characteristics, and internal carriage of air-to-air and air-to-ground ordnance. Its Hivemind artificial intelligence is designed to operate with minimal human oversight, even in heavily jammed environments.
The program now benefits from $400 million in public and private financial commitments, including $200 million provided by the U.S. government. Shield AI aims for a maiden flight before the end of 2026, a fully mission-equipped airframe in 2028, and serial production starting in 2029.
The First X-BAT Prototype Marks a Major Milestone for Shield AI
On November 4, 2026, Shield AI confirmed the completion of the first full-scale X-BAT prototype. Designated X-BAT 001 and registered as N228XB, the airframe was assembled at the X-Forge 1 facility in Frisco, Texas.
Revealed visually during the International Fighter Conference in Munich, the airframe is structurally complete, fitted with its propulsion system and hardware necessary for initial ground tests.
Unlike previously displayed static mockups, this represents a functional experimental aircraft.
The prototype is scheduled to be transported to X-Forge 2 in Newton, Kansas, for a series of ground and flight trials. Shield AI plans to build two full-scale prototypes to mitigate technical risk before locking in the final serial production configuration.
The first flight is targeted before the end of 2026. At this stage, this target remains an industrial goal rather than a demonstrated capability.
Shield AI Expanded from Small Drones to Autonomous Air Combat
Founded in 2015 in San Diego by former Navy SEAL Brandon Tseng, his brother Ryan Tseng, and engineer Andrew Reiter, Shield AI initially specialized in autonomous systems tailored for tactical military operations.
The defense firm started with Nova, a small indoor drone designed to search buildings without requiring constant manual operator inputs.
It subsequently developed Hivemind, a software architecture enabling unmanned aircraft to execute complex missions despite satellite navigation loss or radio frequency denial.
In 2021, acquiring Martin UAV brought the V-BAT into its portfolio, a vertical takeoff reconnaissance drone. This tactical system has notably seen operational deployment in Ukraine starting in 2024.
By April 2025, Shield AI reported over 130 operational sorties executed alongside Ukrainian forces in environments severely disrupted by electronic warfare.
Hivemind has also undergone flight testing on crewed military aircraft, including the X-62A VISTA testbed derived from the F-16.
This flight heritage forms the technological foundation of the X-BAT. Shield AI is no longer simply retrofitting software onto existing airframes, but building an entire combat system purpose-designed around artificial intelligence from inception.
The X-BAT Combines Stealth, Jet Propulsion, and Vertical Takeoff
The X-BAT utilizes a tailless flying-wing configuration with an angular fuselage designed to reduce its radar cross-section.
The reported wingspan measures 11.9 meters (39 ft). The manufacturer claims a maximum unrefueled ferry range exceeding 3,700 km (2,000 NM) and a operational ceiling above 15,240 meters (50,000 ft).
Targeted maneuvering capabilities exceed 4 g. These flight envelope projections remain subject to flight test validation.
The aircraft is powered by a General Electric F110-GE-129E turbofan, derived from the engine family powering platforms such as the F-16 Fighting Falcon.
This engine produces approximately 129 kN (29,000 lbf) of maximum thrust in afterburner.
The primary engineering innovation lies in integrating an Axisymmetric Vectoring Exhaust Nozzle (AVEN), which directs exhaust flow to control the aircraft during vertical flight regimes.
Initial static engine integration and thrust vectoring nozzle tests were successfully completed in July 2026.
Vertical Takeoff Relies on a “Tail-Sitter” Architecture
Unlike traditional helicopters or tilt-rotors, the X-BAT does not rely on separate lift fans or lift engines.
The aircraft sits vertically on a mobile launch and recovery platform.
At takeoff, raw jet thrust lifts the airframe directly upward. The aircraft gains altitude before pitching forward to transition into horizontal flight, where aerodynamic lift is generated by the wings.
Landing follows the exact inverse sequence.
This tail-sitter architecture eliminates the need for runways, but introduces significant flight control challenges. The engine must operate reliably across radical orientation changes, while flight control algorithms maintain stability at hover and near-zero airspeeds.
Thermal management and jet blast deflection also present engineering hurdles, particularly when operating from ship decks exposed to high-temperature exhaust.
Shield AI plans to conduct horizontal engine runs, vertical static burns, and tethered crane flights prior to unconstrained free hover attempts.
The X-BAT Drone Is Designed for Air-to-Air and Air-to-Ground Combat
Shield AI conceives the X-BAT as a multirole platform intended to operate within heavily defended airspace.
Its core mission sets encompass air-to-air combat, precision strike, ISR (intelligence, surveillance, reconnaissance), and electronic warfare.
The aircraft is designed to carry air-to-air missiles and precision-guided air-to-surface munitions inside internal weapons bays. External hardpoints can also accept additional stores, albeit at the expense of radar stealth.
The manufacturer projects a maximum payload capacity reaching up to 2,720 kg (6,000 lb), split between internal and external carriage options.
This payload metric remains a design target; no weapons integration or live-fire trials have been publicly reported yet.
The X-BAT is envisioned to operate alongside advanced fighter jets such as the F-35 Lightning II to conduct forward reconnaissance, suppress air defenses, or strike high-threat targets without placing human aircrews in direct jeopardy.
Hivemind AI Enables Autonomous Operations Without Continuous Data Links
Autonomous mission execution centers on Hivemind, Shield AI’s proprietary software framework.
The system integrates navigation processing, tactical data fusion, trajectory planning, and real-time mission execution.
Its objective is to allow a single human operator to supervise multiple uncrewed platforms without micro-managing individual flight maneuvers.
The drones receive broad tactical objectives and autonomously adapt their flight paths according to dynamic environment changes.
Shield AI asserts that Hivemind maintains full mission capability even when satellite navigation signals and communications links are completely severed.
This level of autonomy is particularly relevant in the Indo-Pacific theater, where military planners anticipate operating under contested, high-density electronic warfare environments.
However, specific rules of engagement and weapon release authority mechanisms for operational deployments remain subject to ongoing military policy definition.
U.S. Navy Backs $400 Million Program Effort
The key operational appeal of the X-BAT lies in its complete runway independence.
In the Indo-Pacific, fixed military runways represent high-value targets vulnerable to long-range ballistic and cruise missile strikes.
X-BAT units could be dispersed across island chains, austere forward operating locations, or naval vessels lacking catapults or flight decks.
In October 2026, Shield AI announced that cumulative program investments reached $400 million, split evenly between government defense funding and private venture capital.
Public funding includes $50 million awarded under the U.S. Navy’s Runway Independent Maritime Expeditionary Strike program, followed by an additional $150 million dedicated to continued development.
This funding footprint does not represent a formal procurement contract for production aircraft.
The U.S. Navy is evaluating whether the X-BAT can deliver distributed long-range strike capabilities from platforms far more numerous and survivable than traditional aircraft carriers.
Production Timeline Remains Highly Ambitious
Shield AI plans to fly a fully mission-equipped X-BAT prototype in 2028 and initiate serial production in 2029, pending formal military procurement requirements.
The company is establishing its X-Forge 3 manufacturing plant in Washington state, scheduled to open in 2027. The site will feature approximately 39,000 sq m (420,000 sq ft) of advanced industrial manufacturing space.
Over the long term, Shield AI projects potential production capacities reaching up to 150 aircraft per year.
However, achieving this production velocity hinges on flight test outcomes.
The X-BAT must conclusively demonstrate transition stability between vertical hover and wing-borne flight, recovery system reliability, maritime corrosion resistance, and verified performance figures.
Unit cost metrics have not been disclosed, at a time when defense acquisition programs are prioritizing collaborative combat aircraft that offer lower price points relative to manned fighters.
Shield AI offers a radical departure from conventional CCA designs. The ultimate success of the X-BAT will depend on its ability to turn its primary theoretical advantage—runway independence—into a robust, cost-effective, and operationally reliable defense capability.
The coming years will test the validity of this technical vision. Achieving a vertical takeoff is a significant milestone, but deploying armed, autonomous jet drones at scale from ships or remote island outposts presents a vastly more complex challenge.