Speaking to Defence Connect, Humble states that AI is fast becoming a core decision-support capability for the Australian Defence Force and that the recently released 2026 Defence Industry Development Strategy (DIDS) has, whether by accident or design, positioned Australian small and medium-sized enterprises (SME) to shape that transformation from the ground up, rather than simply supply into it after the fact.
The conversation, recorded roughly a month after Defence Industry Minister Pat Conroy has unveiled the strategy at the National Press Club, arrives at a pivotal moment.
Prime Minister Anthony Albanese’s recent announcement of new regulatory settings for AI infrastructure and data centres has underscored that artificial intelligence is no longer a peripheral consideration for the defence industrial base, it is becoming central to how capability is designed, tested, procured and sustained.
AI as decision support, not replacement
Humble is careful to frame AI’s role in defence in functional rather than futuristic terms.
“AI in a defence context is increasingly becoming a decision support capability,” she says. “The value isn’t about replacing people, it’s about helping operators make faster, more informed decisions by processing the vast amounts of very complex information they have, at speed and at scale.”
That framing matters for how the DIDS itself is interpreted. Rather than treating AI as a stand-alone acquisition category, Humble sees it as a connective layer across the capability life cycle, one that depends on sustained collaboration between defence and industry “throughout development, not just at the point of acquisition”.
It is a subtle but important distinction, and one with real consequences for how Defence contracts, tests and sustains software-enabled systems. Unlike traditional hardware platforms, which might remain in service for decades largely unchanged, AI-enabled software matures continuously through iteration, testing and operational feedback.
Humble says this reality undercuts the traditional model of locking in requirements upfront, well before a capability reaches the field.
“The earlier in the life cycle Defence understands what is technically possible, the more flexibility it has to adopt solutions that are going to continuously improve through their life cycle,” she says, describing the process as “spiral development” – the insertion of capability upgrades even while a platform is already in service.
She points to US General (Ret’d) David Petraeus’ remarks at a recent Morgan Stanley Australia Summit in Sydney, where he argues that large, exquisite platforms increasingly need to be complemented – not replaced – by cheaper, more disposable and asymmetric capabilities such as autonomous systems and drones.
Where the DIDS and AI adoption intersect?
If AI is reshaping how capability is built, the 2026 DIDS is, in Humble’s reading, reshaping who gets to help build it and how early they get a seat at the table.
Her chief takeaway from the strategy document was is its explicit recognition that Australian companies with specialist expertise and operational insight have an integral role in shaping sovereign capability “from the outset”, rather than being brought in only once requirements are already locked.
Sovereign capability is about a lot more than what Defence owns … It’s about the Australian companies and SMEs that innovate, design, develop, integrate and sustain capability over decades.”
- Rebecca Humble
“Sovereign capability is about a lot more than what Defence owns,” she says. “It’s about the Australian companies and SMEs that innovate, design, develop, integrate and sustain capability over decades.”
Bringing industry into the conversation before requirements are set, she says, enables companies developing AI and decision-advantage tools to “demonstrate what’s technically achievable and identify potential risks a lot earlier”, helping shape delivery pathways rather than simply responding to a finished specification.
This is also where the DIDS’ emphasis on alternative financing becomes relevant to the AI conversation. Aurizn, backed by private equity firm Pemba Capital, is among a small number of Australian defence companies to have taken on institutional capital at scale, a model more familiar in the United States among firms such as Anduril, Shield AI and Saronic. Humble describes the DIDS’ attention to this financing model as validating a strategic bet Aurizn made some years ago, in response to the 2024 iteration of the strategy.
“What private equity gives us is that optionality around capital resourcing, investment and really focused capital decisions to help us grow and scale,” she says. “It’s definitely a force multiplier.”
However, Humble is candid that Australia’s comfort with defence-sector private capital lags well behind the US and the UK, where such funding has been the norm “for generations, decades”. Part of the gap, she suggests, is a lack of familiarity among founder-led Australian SMEs with how capital markets can support growth – an education gap she believes needs deliberate attention if the DIDS’ ambitions for scaling are to be realised.
The missing middle and AI’s role in closing it
Central to Humble’s reading of the DIDS is what has been termed Australia’s “missing middle”, the underdeveloped tier of mid-sized defence companies between large multinational primes and a crowded field of early-stage SMEs.
Unlike the US and the UK, where industrial bases feature a fuller spread of large, medium and small firms, Australia’s ecosystem remains comparatively thin in the middle, limiting the customer Defence’s optionality and constraining national resilience.
AI and software-enabled capability development, in Humble’s view, is precisely the terrain on which that missing middle could be built, provided the right conditions are created. She cites Aurizn’s own experience developing Infinite Studio, a synthetic environment platform co-designed with Defence Science and Technology Group, as a template.
The platform enables defence organisations and allied partners to generate accurate synthetic environments supporting mission experimentation, sensor evaluation, AI training and autonomous systems development – sovereign Australian technology that has since expanded into use by partners in the UK, the US and Canada.
“That’s exactly what sovereign capability should look like,” she says. “Australian innovation developed in partnership with the customer, with Defence, solving Australian problems while creating export opportunities with trusted partners.”
The lesson she draws from that experience is structural: integrated teams bringing operators, engineers, analysts and developers together from the outset of a program, rather than sequential hand-offs between Defence and contracted suppliers.
Paired with continued investment in modelling, simulation and digital experimentation environments, she argues this approach reduces both technical risk and the time needed to mature emerging technologies before fielding.
Learning from but not copying allied models
Humble’s assessment is informed by 17 years working across the US and UK defence industrial bases before returning to Australia – eight years in Washington DC, nine in the UK.
She points to the US Defence Innovation Unit’s approach of presenting operational challenges rather than prescriptive specifications and to the UK’s Defence and Security Accelerator, which funds solutions against real problems and allows technologies to mature before formal procurement begins, as models with genuine parallels to Australia’s own Advanced Strategic Capabilities Accelerator. She also draws on NASA’s approach of defining a mission and desired outcome and then allowing industry latitude to solve the problem, arguing it accelerates innovation while reducing programmatic risk.
But she is equally clear that direct transplantation of allied models would be a mistake. Australia’s smaller defence market, its comparatively nascent culture of bipartisan public support for defence spending, and its underdeveloped middle tier of industry all mean that “whatever we adopt as a nation has to be obviously nuanced and relevant for our strategic circumstances”.
On AI adoption specifically, Humble notes that Australian defence companies are not moving at the pace of the commercial sector, largely because a prescribed legal and regulatory framework for AI adoption in defence has, until recently, been absent, leaving companies less certain that investment in AI capability will translate into contracted returns.
She believes that dynamic is now shifting as policy catches up but expects industry will need to “keep up with that pace and evolve quite quickly and iteratively” as it does.
Four priorities for the road ahead
Asked to distil where the DIDS’ ambitions should translate into practical action, Humble set out four priorities.
First, continuing to give industry greater visibility of Defence’s operational challenges before capability requirements are locked in, to build stronger pathways from prototype to program of record.
Second, an area she acknowledges reflects Aurizn’s own commercial interests: further expansion of modelling, simulation and digital experimentation so emerging technologies can be evaluated earlier and with greater confidence.
Third, more structured opportunities for defence, industry and research bodies to solve capability problems together, shifting away from transactional subcontracting towards genuine trusted-partnership models.
Finally, sustained investment in sovereign Australian technologies with real export potential to build the confidence venture and private capital needed to back Australian defence technology at scale.
Underlying all four, Humble argues, is a more fundamental and less-discussed challenge: the commercial relationship between the Department of Defence and its industrial base needs to change, not just in policy language but in the incentives that shape day-to-day behaviour.
Referencing a quote often attributed to Albert Einstein about the futility of repeating the same approach while expecting a different outcome, she suggests that without addressing those underlying commercial constructs, the DIDS’ stated ambitions for sovereign capability and a scaled, resilient industrial base risk remaining aspirational.
For an industry increasingly defined by software, autonomy and AI-enabled decision advantage, Humble’s message is that the DIDS has set the direction, but that translating policy intent into a genuinely reshaped relationship between Defence and its industrial base, particularly its SMEs, is the priority for 2026.
