• • • DEFENCE TECHNOLOGY • • •


PHYSICAL AI IN DEFENCE DEPENDS ON TRUSTED DATA


misunderstand a question and give an incorrect recommendation. In defence, the consequences are very different


BY RAHN TWITCHELL, SENIOR ACCOUNT MANAGER, TUXERA T


he defence industry is no stranger to technological change. Over the past decade, we have seen a shift from


hardware-centric platforms to software-defined systems, where new capabilities are increasingly delivered through software rather than hardware upgrades. Artificial intelligence (AI) is moving beyond


analysing information to making decisions and taking action in the physical world. From autonomous drones and robotic vehicles to intelligent surveillance systems and collaborative human-machine teams, AI is becoming embedded in defence platforms that must operate in complex, unpredictable environments. This next wave, described as Physical AI,


enables machines to perceive their surroundings and act on those decisions in real time. As AI moves from centralised computing towards distributed edge intelligence, engineering priorities are shifting too. Success will increasingly depend not only on smarter AI models, but on whether autonomous systems can trust the data they rely upon in every operating condition.


Defence AI doesn’t get


a second chance In many enterprise environments, an AI error may be inconvenient. For example, a chatbot might


because an autonomous platform may need to continue operating after losing connectivity, a drone may have milliseconds to avoid a collision, or a surveillance system must preserve sensor data for post-mission analysis. In these environments, failures can have operational consequences rather than merely reducing efficiency. These systems are continuously sensing,


deciding and acting and every decision depends on data being captured and retrieved correctly. If that data is corrupted or unavailable, even the most sophisticated AI model cannot make reliable decisions. In the era of Physical AI, mission-critical increasingly means ensuring systems behave predictably in the real world, even under degraded communications and unpredictable operating conditions. Data integrity and resilience are becoming engineering fundamentals rather than desirable features.


The hidden challenge behind


autonomous systems Autonomous military platforms continuously generate data from cameras and electronic warfare sensors meaning the challenge isn’t simply storing that information, but ensuring the right data is always available exactly when an autonomous system needs it. Much of the industry’s attention has


understandably focused on training larger AI models and increasing edge compute performance. However, these systems only perform as well as the data they rely on because if


32 ELECTRICAL ENGINEERING • JULY/AUGUST 2026


sensor data is corrupted or unavailable when it’s needed, even the most advanced AI model cannot compensate. AI has advanced rapidly across compute,


software and cloud infrastructure. What remains largely unaddressed is the infrastructure responsible for ensuring data remains complete, consistent and reliable once AI leaves the data centre and enters the physical world. This is the Physical AI data layer, responsible for


reliably capturing and managing data throughout an autonomous system’s lifecycle. Unlike traditional embedded systems designed primarily to log and store data, AI-enabled platforms demand high-throughput and deterministic storage that remains resilient under continuous workloads. Without this foundation, even the most advanced AI is limited by the quality and availability of the data it relies on.


Reliability as a strategic


capability Unlike consumer technologies, defence platforms often remain in service for decades while software and AI capabilities evolve continuously. Over-the-air updates allow new functionality to be deployed without replacing hardware, but they are only as effective as the underlying data layer that protects system integrity before, during and after deployment. Poor data resilience doesn’t simply increase


technical risk, but it can delay platform upgrades as well as increase maintenance burdens which ultimately reduce operational availability throughout a programme’s lifetime.


electricalengineeringmagazine.co.uk


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48