In my previous article ^[1], I noted that because of the disruptive innovation occurring in Australia’s electricity sector, traditional generators are realising that their core assets have become stranded—they’re losing money and can’t be sold. I posed the question of whether the Department of Defence would be faced with the problem of stranded assets as a range of disruptive innovations become reality. These disruptions include the increasingly lethal threats that crewed platforms face, the transition away from liquid fuels and the growth of robotic and autonomous systems.

One way to avoid having stranded assets is to not acquire them in the first place. The energy sector has reached that tipping point; while coal-fired power plants still exist, nobody wants to invest their own money in new carbon-fired generation. Defence has not reached that tipping point yet; its investment program is still built around acquiring new, extremely expensive, crewed platforms.

To be fair, nobody really knows when that tipping point will occur. It will likely occur at different points for different kinds of platforms and there will be a range of views—unlike Defence, I wouldn’t be acquiring a new manned attack helicopter ^[2] to replace the army’s Tiger—but it would be an extremely brave chief of the defence force who recommended turning off investment in crewed submarines right now. It’s understandable that in the face of uncertainty Defence defaults to what it knows, which is crewed, multirole platforms.

So it’s likely that Defence will end up with stranded assets. One way to mitigate the harm is to modify or re-role them over time. Depending on the disruptive technology that’s potentially stranding them, that may or not be possible; I’m not sure whether it will be possible to reverse engineer electric drive and batteries into a 40-tonne infantry fighting vehicle sometime in the 2030s, for example.

What about the disruptive threat posed by robotic and autonomous systems? There’s a growing consensus that the future of autonomy in militaries will take the form of MUMT (manned–unmanned teaming) or HUMT (human–machine teaming) in which large, crewed platforms continue to play a key role.

For the Australian Defence Force that recognition is significant progress in itself; its focus is increasingly on what robotic and autonomous systems can do, rather than on what they can’t. We can see this conceptual development in action as the argument for the army’s infantry fighting vehicles moves from one driven by protection (something increasingly implausible as threats proliferate) to one in which the vehicle is a key command and control node in a web. Similarly, the navy’s robotic and autonomous systems strategy ^[3] presents a quasi-fractal picture in which large, crewed platforms launch uncrewed ones, which in turn launch progressively smaller uncrewed ones.

In the shorter term, HUMT concepts are probably more viable than a force structure based largely on autonomous systems, but whether they can ensure the utility of Defence’s crewed platforms in the longer term seems doubtful, at least as they are currently designed.

To explain this, let’s unpack the benefits of autonomy. Many of its advocates argue that the key positive is removing humans from the battlefield, but for me that’s a result of a larger concept, namely disaggregation. Over many decades we’ve seen the aggregation of capability as individual platforms incorporate more sensors, more weapons, more processing power and more power generation to boost their survivability and, importantly, that of their crew. This has driven their increasing size and complexity, which in turn has driven the time required to design and build them. It also drives ever upwards-spiralling costs. We may not have quite reached the Death Star ^[4], but we’re well on the way.

Disaggregation offers a way to reverse this trend. It means distributing systems and capabilities into smaller packets. Those packets are orders of magnitude simpler, and simple things are easier to design and faster to build. Those simple systems are not identical and each is designed to do only a small number of things. That means they’re cheaper, so they can be built at scale and easily written off. But each will enjoy the benefits of significant processing power.

Removing humans from combat platforms is just another form of disaggregation, reducing the complexity, size and cost of those platforms while also delivering the benefit of removing people from danger.

Autonomy (along with the command, control and communication systems, and artificial intelligence that enable it) allows you to reaggregate those simple systems into complex networks. Whether your conceptual language describes that whole as a ‘distributed kill web’, ‘mosaic warfare ^[5]’ or something else, autonomy allows large numbers of simple things to do tasks beyond the ability of small numbers of complex things. Because they’re cheap, those simple things are attritable and replaceable, and the whole is more resilient. At the core of these concepts is a commitment to and trust in the ‘smart, small and many’.

In the near term, there’s no alternative to HUMT concepts in which the smart, small and many work with our few and large crewed platforms. Some platforms coming online are designed to work with robotic and autonomous systems. The Arafura-class offshore patrol vessel, for example, will operate uncrewed air, surface and underwater systems. But autonomy potentially allows us to reimagine and reinvent some existing platforms. With their large capacity, the Canberra-class landing helicopter docks could potentially be the ADF’s most effective anti-submarine platforms, operating large fleets of crewed and uncrewed helicopters, fixed-wing surveillance drones, swarms of small, sonar-towing surface vessels, underwater vessels and so on. They could also be capable strike platforms, whether by launching something like the loyal wingman unmanned aerial vehicle or simply hosting the army’s long-range strike missiles.

But looking further ahead, the idea of adapting existing crewed platforms becomes less tenable. There are several reasons for this. If one of the strengths of the future network is that it provides resilience by being comprised of many small, cheap and ultimately attritable components, why would you create a small number of points of failure in that network in the form of large, irreplaceable crewed platforms.

And while Defence’s mantra is ‘integrated by design’, many of the crewed platforms being acquired by Defence have fundamental limitations in terms of being able to be integrated with autonomous systems. Put another way, why would you tie your agile autonomous systems to something like a 40-tonne infantry fighting vehicle that’s difficult to deploy and sustain on operations, and is an easily detectable and targetable point of failure? If you were to design a land or maritime platform that was to be a fully integrated mothership or command node for large numbers of largely autonomous systems, it probably wouldn’t look like an infantry fighting vehicle, or a Hunter-class frigate.

HUMT concepts will no doubt be a core element of warfighting for a long time to come. Whether the humans in those concepts will crew the platforms Defence is currently acquiring is doubtful. That suggests Defence should be exploring concepts and technologies that produce a capability return on its financial investment sooner, before those investments are stranded. No company would invest in infrastructure that won’t produce cash flow for nearly 20 years, but that’s precisely what Defence is doing with its future submarine project. There may still be a place for governments to take that risk, but it should only be one element of their capability investment portfolio—and in an era of rapid technological change, it probably shouldn’t be their biggest one.