An AI copilot is now riding along on US Air Force missions. Beacon AI’s Murdock pilot assistant has flown aboard Air Force refueler-and-transporter aircraft, advising flight crews and steadily accumulating real-world flight experience.
Co-founder Matt Cox, a former Navy fighter pilot, describes the goal as building an “R2-D2 for pilots.”
What Murdock actually does in the cockpit
Murdock is not flying the plane. It is designed to help the people who are.
The platform handles real-time checklists, performance calculations, and route optimization. Those are tasks pilots manage constantly, often while juggling radio calls, weather, and everything else that makes flying a heavy aircraft demanding.
Beacon AI ran flight tests in 2025 that validated Murdock’s features across critical phases of flight, including taxi and landing.
During operational flight tests, Murdock received an over-the-air software update that was completed in hours. That pioneering mid-flight update capability came out of an earlier AFWERX contract, and it drastically reduces downtime.
A growing stack of Pentagon contracts
Beacon AI, founded in 2021 and based in San Carlos, California, has secured at least 13 contracts with the US Department of Defense.
In January 2026, the company announced a $4 million contract with Air Mobility Command. The deal covers adapting Beacon AI’s platforms for operational airlift and aerial-refueling missions over a span of two years.
In April 2026, Beacon AI obtained a $49.5 million contract with USSOCOM for its AI suite, covering over 350 special operations and Air Force aircraft.
For a company that raised $15 million in a Series A funding round in October 2024, that is a meaningful jump. A single defense contract worth more than three times its Series A suggests the military is moving past curiosity and into deployment.
The pitch: AI without ripping out the cockpit
Beacon AI’s design philosophy, as articulated by Cox, emphasizes a streamlined, adaptable approach. The aim is to serve both military and civilian aviation.
That flexibility aligns with a broader Department of Defense objective: integrating advanced AI without requiring significant hardware overhauls.
Cox’s own background shapes the product’s framing. As a former Navy fighter pilot, he understands that cockpit workload, not raw aircraft capability, is often the limiting factor in high-pressure moments.
What this means for military aviation and the AI race
The most important word in Beacon AI’s recent progress is “real-world.” Plenty of companies can demo AI in simulators. Fewer can point to hours logged aboard operational Air Force aircraft with actual crews.
The Air Mobility Command work runs over two years, so the results of adapting Murdock for airlift and aerial-refueling missions will take time to show. Scaling across more than 350 special operations and Air Force aircraft under the USSOCOM contract will test whether the system performs consistently across different airframes and mission types.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.