For a century, air power has started in the same place: a long strip of concrete. Shield AI, a US defence start-up, wants to remove it from the equation. In early November 2026 the company said the first full-size prototype of its X-BAT, an uncrewed, AI-piloted combat jet that takes off and lands vertically, is finished and on its way to flight testing.
It is easy to file this under “another flashy drone.” That would miss what matters. The X-BAT is less a bet on artificial intelligence than a bet on geography, on where an air force can fight from when its bases are the first thing an enemy targets.
What happened
On 4 November 2026, Shield AI said its first full-scale X-BAT prototype was complete and was moving from its development site in Frisco, Texas, to its flight-test facility in Newton, Kansas. The company says the aircraft was built in about a year. Aviation Week and The War Zone reported the unveiling a day earlier.
The test plan is careful and step by step. First, the aircraft will lie horizontally on its launch-and-recovery vehicle and run its engine. Then it will be raised upright and run vertically. Next it will hang from a crane, a safety tether used only for testing, before any untethered flight. Shield AI has said it aims to fly before the end of 2026.
How the X-BAT works
A fighter that sits on its tail
The X-BAT does not sit level on landing gear like a normal jet. It is a “tail-sitter”: it rests upright on its tail on a trailer, launches straight up, tips over into normal forward flight, and on return pitches its nose up to come back down vertically.
One engine, steered thrust
Lift comes from a single GE Aerospace F110, from the same engine family that powers some F-16 and F-15 fighters, according to Air & Space Forces Magazine. A thrust-vectoring nozzle swivels the exhaust in pitch and yaw to steer while hovering, and small jets at the wings handle roll. According to The War Zone, the nozzle is the AVEN design first tested on an experimental F-16 in the 1990s.
The pilot is software
There is no cockpit. The aircraft is flown by Shield AI’s Hivemind autonomy software. The company says it is designed to keep working when GPS and communications are jammed, and that one human commander can supervise several X-BATs at once. These are design claims; none has yet been shown in flight on this aircraft.
The numbers so far
| Item | Figure | Status |
|---|---|---|
| Announced funding | $400 million | US Navy $200M + Shield AI $200M (company release, Oct 2026) |
| Combat radius | about 1,000 nautical miles | Company projection |
| Maximum range | more than 2,000 nautical miles | Company figure |
| Size | about twice the USAF’s first-batch CCA drones | Company statement |
| Deck footprint | 3 X-BATs in the space of one legacy fighter | Company claim |
| Missionized aircraft | 2028 | Company target |
| Production | 2029 | Company target |
| Unit price | Not public | — |
The money is real. The Navy funded the project through a programme called Runway Independent Maritime Expeditionary Strike (RIMES), run by the Navy’s Rapid Capabilities Office with the Defense Innovation Unit: $50 million in August 2026 and a $150 million contract modification in October, each matched by Shield AI’s own capital. The company names aircraft carriers, amphibious assault ships and expeditionary sea bases as possible homes.
Why runways are the real story
Here is the number that explains why the Pentagon is paying for this. In its widely cited January 2023 wargame of a Chinese invasion of Taiwan, the Center for Strategic and International Studies (CSIS) found the US and its allies lost hundreds of aircraft, and that 90 percent of aircraft losses happened on the ground, not in air combat. CSIS also noted that US forces would be limited to “a handful of air bases in the Western Pacific.”
In other words, the most dangerous place for a modern fighter may be parked on its own runway. A long, fixed strip of concrete is easy to find and hard to hide. An aircraft that can launch from a trailer, a small island or the deck of an amphibious ship spreads the target out. As the Navy’s J.J. “Yank” Cummings put it in Shield AI’s October announcement, a runway-independent design “can increase sortie generation.”
That is the strategic logic: you do not need to win every dogfight if the enemy cannot find your aircraft before they take off.
Our take: three things most coverage misses
1. The funding source hints at exports. The $150 million October tranche came through the Pentagon’s Defense Exportability Features programme, according to Shield AI. That programme exists to design export-friendly versions of US systems early. Our reading: Washington sees X-BAT not only as a US Navy tool but as something allies could one day buy.
2. The deck-space claim matters more than the AI. If the company’s “three for one” footprint claim holds up, an amphibious ship could carry far more strike aircraft than it does today. That would change what smaller ships can do, which may matter more than whether the software can out-fly a human.
3. Skipping the runway is not free. Lifting a fighter-sized aircraft straight up on jet thrust is demanding. The open question is how much of the X-BAT’s projected 1,000-nautical-mile radius survives a vertical launch with a full weapons load; Shield AI has not published a figure for that case. Until it flies, the projections are just that.
Who gains, who loses
- Gains (our assessment): The US Navy (more places to launch from), allies with small carriers or island bases, and newer defence companies challenging the traditional primes.
- Under pressure: Planners who rely on a few large, fixed air bases, and expensive crewed fighters whose value depends on surviving the first missile salvo.
- Watching closely: China, whose large missile force threatens the few US air bases in the Western Pacific, according to CSIS.
What to watch next
- First engine runs and tethered hover in Kansas. Aviation Week reports GE modified the F110’s oil system so the engine can run for long periods while upright, one of the programme’s key technical risks.
- First untethered flight, targeted for before the end of 2026.
- A unit price. Without it, claims about “affordable mass” cannot be judged.
- Ship trials. Aviation Week lists managing exhaust heat on ship decks as another technical risk.
- Whether the Air Force follows the Navy with its own money.
FAQ
Is the X-BAT a drone or a fighter jet? Both. It is an uncrewed combat aircraft built to do fighter-style missions (air-to-air, strike and electronic warfare) with an AI pilot instead of a human.
Has the X-BAT flown yet? Not as of 10 November 2026. The first full-size prototype has just been completed. Shield AI plans engine runs, crane-tethered tests and then a first flight, which it aims to do before the end of 2026.
How does it take off without a runway? It sits upright on its tail on a trailer. A single F110 engine with a steerable nozzle lifts it straight up, then it tips over into normal flight.
Who is paying for it? The US Navy has committed $200 million through its RIMES programme, and Shield AI says it is matching that with $200 million of its own money.
When could it enter service? Shield AI’s targets are a missionized aircraft in 2028 and production in 2029. Those are company goals, not a government schedule.
Why does runway independence matter? In a 2023 CSIS wargame of a war over Taiwan, 90 percent of aircraft losses occurred on the ground. Aircraft that can launch from many small sites are harder to destroy before they take off.
References
- Shield AI, “We built the first X-BAT prototype in one year. Now we put it to the test,” 4 Nov 2026: https://shield.ai/we-built-the-first-x-bat-prototype-in-one-year-now-we-put-it-to-the-test/
- Shield AI, “Shield AI and U.S. Government investment in X-BAT reaches $400 million,” 8 Oct 2026: https://shield.ai/shield-ai-and-u-s-government-investment-in-x-bat-reaches-400-million/
- Shield AI, X-BAT product page: https://shield.ai/x-bat/
- Aviation Week, “Shield AI’s X-Bat Moves Toward First Flight,” 3 Nov 2026: https://aviationweek.com/defense/aircraft-propulsion/shield-ais-x-bat-moves-toward-first-flight
- The War Zone, “Prototype X-BAT VTOL ‘Fighter’ Drone Unveiled,” 3 Nov 2026: https://www.twz.com/news-features/x-bat-vtol-fighter-drone-prototype-unveiled
- Air & Space Forces Magazine, “Shield AI Moves X-BAT Drone Into Testing After Rapid Prototype Construction,” 4 Nov 2026: https://www.airandspaceforces.com/shield-ai-moves-x-bat-drone-into-testing-after-rapid-prototype-construction/
- CSIS, “The First Battle of the Next War: Wargaming a Chinese Invasion of Taiwan,” Jan 2023: https://www.csis.org/analysis/first-battle-next-war-wargaming-chinese-invasion-taiwan
- Naval News, “CSIS Wargame: China’s Invasion of Taiwan in 2026,” 30 Jan 2023: https://www.navalnews.com/naval-news/2023/01/csis-wargame-chinas-invasion-of-taiwan-in-2026/
Featured graphic: Strategic Lens Defence.

Leave a Reply