For the U.S., the autonomous battlefield must have a tether

The Last Pilot Has Not Yet Been Born

In September 2024, I was standing next to the Air Force Research Laboratory’s booth at the Air Force Association’s annual Air, Space & Cyber Conference. Throughout the week, crowds large and small eddied around the nearside corner of the booth, where the nose of a full-scale model of Collaborative Combat Aircraft (CCA) jutted out toward the walkway.

While in attendance, I made a habit of taking laps around the model, stealing glances, and getting a sense of the discussions abuzz within the constant, amorphous crowds observing what was promised to represent an evolution toward the future of air dominance. Standing there, I heard a man’s plainly stated observation from over my shoulder that struck me. He was part of a small group comprised of servicemembers and industry alike: “And with that, the death knell for the combat air pilot. The last one may have already been born.”

There was laughter and more chatter among the small group. It was clear to me that day that autonomy had arrived, in earnest. By the end of the conference, then Secretary of the Air Force Frank Kendall’s signature could be seen scrawled across the airframe, as if to cosign that message for good measure.

The Autonomous Battlefield Needs a Tether

We enter an era that will increasingly be defined by the question of whether the last combat pilot, field operator, sailor, or submariner has already been born. For the foreseeable future, that question will remain rhetorical, as the preference for fighting with retained control over battlefield technologies imposes a requirement to keep a human “on,” if not directly “in,” the kill chain.

Much attention has been paid to the fundamental building blocks of autonomous technologies, the vehicle itself, and the algorithms. But just as the human brain makes sense of the world based on the data delivered by its senses, autonomous platforms will require sensing payloads to serve as the eyes, ears, and voice of the system. Among those will be systems that guarantee assured communications to the command and control (C2) center. We can think of these systems as a tether that keeps the human warfighter in, or on, the loop.

The preference for keeping a human on the loop is more than a normative one. Department of Defense Directive 3000.09 governs the design, development, and fielding of autonomous and semi-autonomous weapon systems.[i] That preference is often assumed to convey a prohibition on autonomous systems from engaging targets on their own, though the directive requires only that “[a]utonomous and semi-autonomous weapon systems will be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force” (emphasis added).

As written, the directive does not prohibit autonomous systems from selecting their own target per se, only that they must permit commanders and operators appropriate levels of human judgment over the use of force. A series of subclauses qualifies what counts as enabling an appropriate level of human judgment. Most notable is the stipulation that such systems “complete engagements within a timeframe and geographic area…consistent with commander and operator intentions,” and that “[i]f unable to do so, the systems will terminate the engagement or obtain additional operator input before continuing the engagement.”

Read closely, the directive’s language can be understood as establishing a policy floor for autonomous systems. A system that finds itself operating outside mission parameters effectively has two paths back to compliance. It can terminate the engagement, self-terminating when it drifts beyond a pre-designated geographic area or exceeds the time set for the engagement. Or it can obtain additional operator input to continue the engagement. Only the first option is something the system can do on its own without additional human input. What the directive actually imposes, then, is an assured-termination standard rather than an assured-communications one.

The second path toward compliance is preferable, as it preserves the operational utility and flexibility within an engagement. But obtaining operator input assumes the ability to communicate with an operator. Thus, while assured communications is not an explicit standard, it is highly desirable, if not an operational necessity. A de facto requirement for assured communications has emerged. How that requirement is satisfied depends on the domain. Domains and mission sets impose their own constraints on establishing and maintaining that tether, but the operational value offered by the ability to maintain it is apparent.

The Market Is Building the Tether

A recent pattern of acquisitions, partnerships, and product developments shows that firms are building technologies that satisfy an assured communications requirement in the air domain for autonomy. This August, Joby, an air mobility company that has historically focused on developing a vertical take-off and landing air taxi, entered into an agreement to acquire Resonant Sciences, a provider of radio frequency mission systems, apertures, and resilient communications, for approximately $500 million.[ii] Notably, Joby intends to consolidate its defense activities under Resonant, which will continue operating under its current name and leadership. The parties characterized the acquisition as an opportunity to unify Joby’s commercially derived vehicle and autonomy with Resonant’s self-described “warfighter focused” expertise in RF, sensing, and low-observable aircraft design.[iii]  This is not the first time Joby has reached for the enabling layer to pair with its proprietary vehicle and autonomy. A year prior, Joby announced a collaboration with L3Harris Technologies with the stated intention of “exploring opportunities to develop a new aircraft class for defense applications.”

Joby reveals an observable dynamic among some autonomous vehicle providers entering the defense market. The autonomy, while not properly characterized as the “easy” part, is the component that non-traditional entrants are most commonly bringing to the table of defense acquisition. What each firm must go out and buy to field a viable product varies with what it already holds, but generally has followed a trend: make the software, buy (or acquire) the hardware. Joby, having built a commercially optimized platform and autonomy stack already, seeks to acquire the mission-enabling payloads and survivability expertise in preparation for its entrance into defense markets. The mission systems layer — apertures, electronic warfare suites, low-observable design, and assured communications — has no commercial analog produced to a military standard, and in the case of low observability and electronic warfare, no meaningful commercial market at all.

Firms specializing in assured communications are bringing to market components targeting the burgeoning autonomous platform market. BAE, for its part, introduced its Silver Link line of low-cost, conformal apertures in March of this year, marketing their release with a visual of a fleet of CCA-inspired UAV platforms networked together with line-of-sight data links.[iv]  In June 2026, the German firm Rohde & Schwarz announced NEMACS (Networked Multipoint Array Communications System), a low-SWaP directional communications system for low-observable uncrewed platforms.[v] The NEMACS system combines an AESA antenna with a software-defined radio designed to support datalinks with a low probability of detection and interception to make assured communications more robust in contested environments. Rohde & Schwarz is reportedly teaming with Boeing to explore integration of NEMACS with its CCA-class UAV, the MQ-28 Ghost Bat.[vi]

A Concept Built on the Link

Current operational concepts for CCA focus on them acting as “missile trucks,” with cueing coming from offboard of the platform, supplied through line-of-sight, and beyond-line-of-sight datalinks.[vii][viii][ix] That is a concept of operations built on the link. With CCA Increment 2 requirements taking shape, and ancillary opportunities abounding, the drive to provide high-end capabilities at lower costs is motivating defense firms to demonstrate what might be characterized as exquisite technologies on UAV platforms. Operational concepts that envision CCAs serving in differentiated roles would undoubtedly increase the budget allocated for mission systems components. Greater mission systems requirements could lead firms to pursue multi-function arrays, single apertures which, while at a higher price tag, would provide a suite of capabilities, thus offsetting their relatively higher costs by providing a host of capabilities. SRC has pursued this path with its Generic Multi-Function Array (GMFA), referred to as Payload B, which has been demonstrated aboard Boeing’s MQ-28A Ghost Bat CCA during a February 2026 live fire demonstration.[x] In June 2026, SRC announced that GMFA development would continue, pursuing capabilities such as “airborne moving target indication, ground moving target indication, synthetic aperture radar, RF emulation, sense and avoid, airborne weather monitoring, and AMS GRA alignment.”[xi] It is plausible that an airframer might utilize such an MFA to satisfy or bolster assured communications capabilities in addition to a menu of potential functionalities to include electronic warfare and organic sensing.

While CCA Increment 1 reached its June 2026 production awards with a clear mission in mind, so far, Increment 2 has no equivalent anchor. Its mission, cost target, and design vector remain ambiguous. In December 2024, service acquisition executive Andrew Hunter confirmed the Air Force had not yet set Increment 2 requirements, explicitly leaving decisions about payloads, and whether the aircraft would be more or less sophisticated than Increment 1 to the incoming Trump administration. The program proceeded under new leadership on deliberately agnostic terms, with nine “concept refinement” contracts for Increment 2.[xii] That award appears to have been a deferment of requirements definition. More recently, Air Force leadership indicated that requirements for Increment 2 are likely to focus on designs optimized for air-to-ground missions or that offer multirole capabilities in what is characterized as a “gentle evolution” of Increment 1 requirements.[xiii]

The Air Force has spent two years deliberately not writing Increment 2 requirements, and that deferral is defensible for genuine design variables, but not for assured communications. Based on the Increment 1 concept of operations, which established a reliance on datalinks for offboard targeting and C2, and the value of keeping autonomous platforms in the fight versus tolerating a floor of self-termination, assured communications should not be a design variable. While design vectors and cost targets may shape the characteristics of the apertures that provide assured communications, they will not impact whether those links are required. That requirement effectively lives in how the Air Force intends to fight, and it should be written into Increment 2 without foreclosing a single design decision still under consideration.

The Tether Should Be the Requirement

When Frank Kendall scrawled his name across the nose of the CCA in September 2024, he was, in effect, signing onto a vision of an Air Force increasingly defined by autonomous systems. Two years later, his description of autonomy remains consistent. For Kendall, confidence in the manned-unmanned teaming concept rests on assured command and control of these platforms across domains.

“The keys to all of these concepts that I lay out [in his book on autonomy] are going to be reliable, secure communications [with autonomous platforms] … You can have default behaviors, you can give them mission orders, but at some level you don’t ever want to lose control of them.”[xiv]

A reading of Kendall’s 2026 book, Lethal Autonomy: The Future of Warfare Whether We Like It or Not, underscores the point. Kendall tackles the challenges of autonomy across every domain, and in each, robust C2 concepts enabled by assured communications are presented as a critical building block of autonomous operations.

The recommendation follows. The Air Force should write an explicit assured communications requirement into CCA Increment 2, and it should do so before the concept refinement field narrows. Not because Directive 3000.09 compels it, the directive is satisfied by a system that self-terminates, but because the requirement preserves operational functionality and flexibility, and is already embedded in how the Air Force says it intends to fight.

Industry has reached that conclusion on its own. Joby did not spend half a billion dollars on an autonomy stack. It spent it on the apertures and communications that make one useful. Firms are building toward a requirement the services have not finished writing, and the last combat pilot may not have been born yet, but the systems that will determine that pilot’s role are being designed now.

---

Jack Schwencke is a strategy analyst at BAE Systems Space & Mission Systems and is a student in Missouri State University’s Defense and Strategic Studies program. BAE Systems competes in the market discussed in this article. The opinions expressed are his own.

notes:

[i] U.S. Department of Defense, DoD Directive 3000.09: Autonomy in Weapon Systems (Washington, DC: Office of the Under Secretary of Defense for Policy, January 25, 2023), https://www.esd.whs.mil/Portals/54/Documents/DD/issuances/dodd/300009p.PDF.