The device is no longer the product

The total artificial heart illustrates the direction of medical technology. It remains grounded in electromechanical performance, but its future value lies in responding intelligently to the patient. Recognising whether someone is resting, walking or exercising, and adjusting accordingly, transforms a pump into an adaptive platform built around physiological needs. 

The controller, sensors, algorithms and user interface collectively determine how the implanted system performs. The differentiator is no longer pump mechanics alone, but the intelligence used to manage physiological variability safely and effectively. 

Counter-drone systems face a related challenge. A system that responds to an adversary the same way every time becomes predictable, and predictable systems can be defeated. Operational conditions change, and an adversary will actively search for predictable patterns or weaknesses. Future capability therefore depends on combining sensing, data interpretation and decision logic so the system can recognise changing behaviours and adapt its response. 

Medical imaging has undergone a similar transformation. In systems such as PET-CT, diagnostic value increasingly depends on algorithms that compensate for motion, align multiple imaging modalities, reconstruct anatomy, remove artefacts and present clinically meaningful information. The scanner acquires data; software determines what can ultimately be trusted and acted upon. 

My experience in medical imaging reinforced this transition. Whether performing motion correction in PET-CT systems, co-registering information across multiple imaging modalities, supporting DBS (Deep Brain Stimulation) surgical planning or automating bone subtraction in CTA workflows, the challenge was ensuring that software and algorithms could transform complex data into information that clinicians could trust and act upon. 

The missions differ, but the engineering problem is closely related. Both must interpret incomplete inputs, make time-critical decisions and operate safely as conditions change. Hardware establishes the physical capabilities of the system, while software and AI increasingly determine how those capabilities are expressed in the real world. 

It is no longer just a device; it is an intelligent platform. That platform includes the physical device, controller, data, algorithms, connectivity, cybersecurity and the interfaces through which people make decisions. 

This evolution increases, rather than reduces, the importance of physical engineering because sensing, computation and actuation must now operate as a coordinated system. 

A traditional regulated device might be validated and then remain substantially unchanged for years. A software-defined platform must be configurable, connected and capable of improvement, while remaining safe, secure and controlled. 

This creates a shared challenge for medical and defence organisations. When software is updated, an algorithm is modified or AI is introduced into decision-making, engineers must determine whether the complete system is still behaving within its approved and assured boundaries. 

The question increasingly becomes one of assuring behaviour, not simply validating software or an algorithm in isolation. 

An imaging algorithm, a surgical guidance platform, a robotic system or an autonomous defence capability may all achieve excellent algorithmic performance in isolation. The engineering challenge is assuring how those capabilities behave when integrated into the complete operational system. 

Cybersecurity makes the connection stronger. A connected hospital system can provide an entry point to a life-support device, turning a cyber vulnerability into a patient-safety risk. Defence systems face the same interaction between connectivity, security and physical consequence. 

The potential harm differs, but neither industry can treat software, safety and cybersecurity as separate concerns. As systems become more connected and autonomous, cybersecurity becomes part of functional safety. 

Mechanical, electrical and software development can no longer operate as a sequence of separate activities. System behaviour emerges from their interaction, together with sensors, communications, algorithms and the people making or receiving decisions. 

At Hydrix, this is where our medical technology experience directly strengthens our defence capability. Hydrix operates at the intersection of regulated medical technology, advanced industrial systems and mission-critical defence technologies, bringing together hardware, electronics, embedded systems, software and AI into deployable products. 

While the application domains differ, the underlying challenge is increasingly the same: engineering trusted behaviour in software-intensive systems operating under uncertainty. 

Medical development has embedded disciplines around traceability, risk management, human factors, verification and validation, and designing for failure modes. Defence extends that foundation into mission assurance, cybersecurity and performance in degraded or contested environments. The standards and operating contexts differ, but the underlying engineering habits are highly transferable. 

The challenge facing many organisations today is not developing better algorithms. It is turning those algorithms into trusted system behaviour. That requires far more than AI expertise alone. It requires architecture, systems engineering, human factors, cybersecurity, verification, validation and an understanding of how complex systems behave when deployed in the real world. 

Physical AI will not be won by the organisations with the best algorithms alone. It will be won by those that can integrate sensing, computation and physical action into trusted systems. 

Over the past two decades, industries have evolved from hardware-defined products to software-defined machines. The next decade will be defined by AI-enabled behaviour, where systems sense their environment, interpret incomplete information, make decisions and act in the physical world. 

Whether the application is a connected medical device, an imaging or surgical robotics platform, or a mission-critical defence system, the challenge is increasingly the same: integrating hardware, software, algorithms, edge AI, cybersecurity and human decision-making into a system whose behaviour can be trusted. 

This capability cannot be added near the end of development. It must shape the architecture, risk controls and assurance strategy from the beginning. 

The physical device still matters, but it is no longer the complete product:  

The next generation of critical technology will not be defined by hardware alone, or by AI in isolation. It will be defined by the ability to combine advanced hardware, AI and software into trusted behaviour that can be deployed safely and reliably in the real world.