Shield AI’s Hivemind mission autonomy software flew multiple U.S. Air Force Collaborative Combat Aircraft (CCA) in a multi-ship formation on October 26, 2026, and carried out tactical beyond-visual-range air combat maneuvers, according to the company. Shield AI says it was the first time Hivemind flew several CCAs as a coordinated team.

In a press release dated November 9, 2026, Shield AI said the formation flight was one of two firsts for the CCA program within three weeks. The other came on October 8, when an F-22 pilot gave high-level directives to an FQ-44A in flight. The company presents the pair as a step from one fighter pilot directing one autonomous teammate to several autonomous aircraft maneuvering together.

Two flight tests in October

On October 8, the F-22 pilot acted as an airborne “quarterback,” Shield AI said. Hivemind interpreted the pilot’s directives and carried them out in real time, so the FQ-44A performed dynamic maneuvers and simulated combat tasks under the pilot’s control. AirPro News covered that test when an F-22 pilot directed a YFQ-44A Fury in October.

On October 26, Hivemind enabled multiple CCAs to fly in formation and execute collaborative beyond-visual-range counter-air tactics. The company said the flight showed collaborative mission execution, not aircraft simply operating in the same airspace. The release does not say how many aircraft took part, which aircraft types flew on October 26 or where the flights took place.

Todd Wesley, vice president of Hivemind engineering at Shield AI, said the company’s mission autonomy runs on each aircraft and translates human intent into collaborative actions. He said the flights also advance the integration of the sensors, data links and interfaces that the autonomy depends on, and that the aircraft and mission systems are reaching a level of integration that lets Hivemind show what it can do.

Software updates between sorties

Wesley said every flight produces data for the next software release. According to Wesley, the combined Hivemind and aircraft software release process is nearing the point where the Air Force can fly a sortie, update the software overnight and fly again the next day.

Shield AI credits the Air Force’s Autonomy Government Reference Architecture (A-GRA) for making these milestones possible. A-GRA separates mission autonomy from aircraft-specific functions through standardized interfaces, so mission software can improve independently of the airframe. The company says this supports competition and lowers the integration burden for new capabilities.

The company describes a repeatable cycle of develop, verify, fly, learn and update across its aircraft deployments, with improvements on any aircraft maturing the software on all of them, limited only by data handling restrictions. Shield AI said Hivemind’s common software foundation has been integrated on more than 40 uncrewed systems across air, land, sea and space, and that the CCA flights ran on that same software.

How Shield AI reached this point in the CCA program

The Air Force selected Shield AI as a mission autonomy provider for the CCA program after a competitive evaluation supporting Technology Maturation and Risk Reduction (TMRR) efforts, as AirPro News reported on its CCA selection in February 2026. In a February 26, 2026, release, Shield AI said the choice of Hivemind was the first time the Air Force had independently chosen both mission autonomy software and an aircraft.

The same release said Hivemind had completed its first CCA flight test aboard Anduril’s YFQ-44A over the Mojave Desert, including mid-mission updates and initial operational behaviors. The November 9 release refers to the aircraft as the FQ-44A.

On June 17, 2026, Shield AI announced a production contract for CCA mission autonomy. Under that award, the company said, it would develop collaborative combat autonomy behaviors that let multiple autonomous aircraft operate together under human supervision, reducing operator workload.

Shield AI, founded in 2015, also builds the V-BAT and X-BAT aircraft, according to the company.

What comes next

Shield AI said it will keep testing, assessing and refining Hivemind with the Air Force, and that data from these flights will feed that work. Lessons from each flight can be delivered as software updates, according to the company. The release gives no dates for further flight tests or for testing and fielding decisions.

The Airbus A320 ATRA fitted with a Eurofighter nose has completed its commissioning flights at the Airbus site in Manching, Germany, and is now operational as the flying test bed for the Eurofighter’s new E-Scan radar. Airbus said on November 6, 2026, that the commissioning flights were the final step in the validation phase.

The aircraft, owned by the German Aerospace Center (DLR), carries the Eurofighter Common Radar System Mk1 that will equip new Eurofighters for Germany and Spain. Testing the radar on an airliner instead of a fighter lets engineers watch the hardware and software in real time from workstations in the cabin, and Airbus said in January 2025 that the A320 has a shorter clearance process and longer endurance than a Eurofighter.

Which Eurofighters get the radar

“The upcoming Eurofighters of the German Quadriga and Tranche 5 as well as the Spanish Halcón programmes will be equipped with a new radar called the Eurofighter Common Radar System Mk1,” said Erik Lorch, Head of the E-Scan Mk1 Programme at Airbus. Airbus said the current focus is on equipping these jets, with retrofits of Eurofighters already in service also planned.

The radar is an active electronically scanned array (AESA), also known as E-Scan. “This radar greatly enhances the capabilities of the Eurofighter in both air-to-air and air-to-ground operations,” Lorch said.

Radar maker HENSOLDT, which developed the ECRS Mk1 with Indra, said in October 2025 that about 200 of the radars will be delivered to the German and Spanish Air Forces in total. HENSOLDT announced at the time that NETMA and Eurofighter Jagdflugzeug GmbH had signed a contract for 20 new Eurofighter Tranche 5 aircraft for the German Air Force, which will carry the radar as the Quadriga fleet already does. AirPro News has reported on the German order for 20 new Eurofighters.

How the A320 was converted

To carry the Eurofighter nose, the A320 fuselage was reinforced and hundreds of metres of data cables were laid, according to Airbus. A liquid and air cooling system was installed to regulate the temperature of the radar, and the cabin holds several workstations where specialists see everything the radar provides to the pilot as well as a full view of what is happening under the nose.

“It is a distinguished platform that has already spent years flying as a research aircraft,” said Sebastian Ritz, who leads the A320 Flying Test Bed Programme at Airbus Defence and Space. “We can monitor how the hardware performs and also test new software versions or features,” he said. “The insights we generate will inform and accelerate the maturation and further development of the system.”

The modified aircraft first flew on January 21, 2025, over DLR’s airfield in Braunschweig, Germany. In its story on that flight, Airbus said the nose had been developed and assembled specifically for the test aircraft by its engineers in Manching. Engineers and mechanics from Airbus Defence and Space and Airbus Commercial Aircraft designed the new front section and reinforced the airframe, and all modifications followed the procedures of the type-certification holder, Airbus Commercial Aircraft in Toulouse. That first flight with the Eurofighter nose was made without an active radar, to test basic airworthiness.

Why test a fighter radar on an airliner

Airbus E-Scan radar project manager Thomas Hirsch said in January 2025 that the aircraft was being operated with DLR and the German Armed Forces “to test a new radar for the Eurofighter and bring it to maturity.” He added: “The A320 ATRA has a significantly shorter clearance process and can stay in the air longer than a Eurofighter.” Airbus said airborne testing on an A320 starts earlier, runs faster and lasts longer than it would on a Eurofighter, which significantly speeds up radar development.

After the first flight, the radar went through extensive ground tests before the commissioning flights. “During the recent commissioning flights, we confirmed that the aircraft and the radar also perform effectively in the air,” Ritz said. “This was the final step in the validation phase, and the E-Scan Mk1 flying test bed is now operational.”

Airbus said the A320 ATRA, short for Advanced Technology Research Aircraft, will now become a regular presence on the flight line at Manching as development of the Eurofighter’s E-Scan radar continues. The company has not published a schedule for the remaining test flights.

Saab AB has successfully demonstrated autonomous Beyond Visual Range (BVR) air combat capabilities using proprietary artificial intelligence agents integrated into a Gripen E fighter aircraft. The Swedish aerospace manufacturer officially unveiled the in-house technology on November 3, 2026, following approximately two years of airborne testing.

In a press release issued today, Saab confirmed the flight trials mark a critical step toward developing collaborative combat aircraft networks, where crewed and uncrewed systems operate as a coordinated force. The milestone underscores the company’s strategy to maintain sovereign control over its technology stack rather than relying entirely on external software providers.

Sovereign AI development and the Sferical facility

Saab centralized its artificial intelligence development under the AI Fighter Pilot Academy (AIFPA), an infrastructure designed to create, train, and evaluate AI agents. The department driving this initiative, Saab Advanced Programs, features a workforce where 40 percent of the staff hold doctorate-level qualifications, according to Aviation International News.

Peter Nilsson, Head of Advanced Programs at Saab Aeronautics, emphasized the operational focus of the project.

“We have not only built our own AI agents, but we have also flown them. This demonstrates our capability to develop AI agents that can be integrated into operational fighter aircraft. The future of air combat will not be defined by individual platforms acting alone, but by networks of crewed and uncrewed systems working together as a coordinated force. Through our ownership of the entire technology stack, we are building the capabilities needed for that future.”

To support this development, Saab recently implemented the Sferical AI facility in Linköping, Sweden. Part-owned by Saab, Ericsson, and AstraZeneca, the sovereign computing asset utilizes Nvidia’s DGX SuperPOD architecture. Aviation International News reported that the facility is powered by 1,152 Nvidia Grace Blackwell GB300 graphics processing units. This computing power allows Saab to run a one-hour flight test simulation one billion times in the same period, compared to just 1,200 times in a traditional central processing unit environment.

Speaking to Breaking Defense, Nilsson outlined the primary challenge facing the AIFPA following the successful flight trials.

“The harder question is how we turn this success into a dependable capability that can learn, improve, and be relevant over time… that is the true purpose of the AI Fighter Pilot Academy.”

Gripen E architecture and testing timeline

Saab, which employs approximately 31,000 people globally, utilized the Gripen E as the primary testbed for its AI agents. The aircraft features a software-defined architecture that segregates mission software from flight-critical functions. This separation enables the rapid integration and testing of new technologies without requiring dedicated experimental platforms or compromising safety of flight.

Prior to unveiling its in-house agents, Saab collaborated with third-party developers to validate the aircraft’s autonomous potential. On May 28, 2025, a Gripen E trials aircraft flew for the first time with the Centaur AI agent developed by Helsing, demonstrating initial BVR air combat capabilities.

Following these third-party trials, Saab transitioned to testing its proprietary agents. The company briefed journalists on the readiness of its in-house AI agents for operational deployment on October 30, 2026, before the official public unveiling on November 3, 2026.

Collaborative combat and future integration

The integration of artificial intelligence into combat aircraft aligns with a broader industry shift toward Collaborative Combat Aircraft (CCA) concepts. Defence contractors globally are developing networks where crewed fighters command uncrewed loyal wingman platforms.

Saab is actively testing the sensor networks required for these collaborative operations. On October 28, 2026, Saab and General Atomics Aeronautical Systems, Inc. (GA-ASI) successfully flight-tested an unmanned airborne early warning (AEW) capability, integrating Saab’s LoyalEye radar with an MQ-9B unmanned UAV.

Looking ahead, Saab anticipates that the Swedish Defence Materiel Administration (FMV) and the Swedish Air Force will absorb these AI agents into real-world operations within months or single digits of years, according to Aviation International News. In the interim, Saab plans to expand the capabilities of its agents by flying them in a broader range of scenarios, establishing the foundation for future uncrewed platforms.

AirPro News analysis

Saab’s insistence on sovereign, in-house artificial intelligence development contrasts with competitors who heavily outsource to commercial technology firms. By owning the entire technology stack, Saab positions itself to offer highly secure, exportable AI packages free from third-party licensing bottlenecks or foreign regulatory hurdles. The massive investment in the Sferical AI facility indicates we are seeing a shift from experimental AI flight tests to industrialized, mass-scale tactical training for autonomous systems. If Saab can deliver on its timeline for Swedish Air Forces integration, the Gripen E could become one of the first Western fighters to field operational, combat-ready AI agents in frontline service.