The Hollow Base

What the National Defence Strategy gets wrong about defence industry, and why the answer is not (primarily) the primes.

The 2026 National Defence Strategy (NDS26) stakes the credibility of Australia’s deterrence posture on five words: ‘Industry policy is security policy.’ (page 69) That claim is not wrong. But what follows it is eight pages of investment priorities, acquisition reform, and partnership frameworks that treat a structural problem as a procurement one. And that’s a problem.

The baseline is not in dispute. As noted in our last piece, Australia produces, by the best available estimates, roughly two-thirds of what it consumes in defence materiel. In a conflict that strains the sea lanes on which the remaining third—a critical third¹—depends, that gap is not a shortfall to be ‘managed’. It is an operational ceiling. NDS26 acknowledges the gap exists but does not acknowledge the depth of the conditions that created it, nor what reversing them needs.

Before interrogating the strategy’s assumptions, the inventory needs stating plainly. The seven Sovereign Defence Industrial Priorities are weighted toward sustaining and maintaining platforms built elsewhere. That is industrially sensible; it reflects where Australian industry has genuine comparative advantage. It also means that what is being labelled a ‘sovereign’ industrial base is, in significant measure, the back half of other nations’ supply chains. Australia is good at keeping things running. It is not, in the main, making them².

The critical minerals endowment is real and significant. The gap between extraction and refined, usable industrial inputs—the value-added transformation where strategic leverage resides—remains overwhelmingly offshore, including in China. The NDS lists aluminium, copper, and zinc as strategic materials. It does not address the processing dependency sitting between the ground and the finished input.

The deeper substrate is harder to paper over. Australia’s manufacturing share of GDP has declined from roughly 14 per cent in 1990 to well below 6 per cent now. That is the tissue from which ‘domestic industrial capability that can be scaled at the onset of conflict’ (page 34) must be conjured³.

Industrial capacity is not a stock variable replenished by investment decisions. It is an emergent property of an ecosystem: distributed skills, supplier networks, research institutions, and the cultural practice of making things, typically accumulated across generations. Ecosystems hollowed over four decades do not reconstitute themselves on a ten-year defence planning horizon.

NDS26’s treatment of supply chain resilience is among the document’s least examined elements. Descriptions such as ‘accelerating efforts…to address critical vulnerabilities’ (page 22)—diversifying global sources, reducing sea-lane dependency, strengthening civil maritime freight—comprise a direction rather than a plan. Running through NDS26 is an assumption that deserves direct interrogation: that international supply chains will remain available under precisely the stress conditions Australian strategy is designed to address.

The COVID precedent. When the pandemic disrupted supply chains for goods far less strategically sensitive than precision munitions, export controls were the first response of major producers. Australia scrambled for personal protective equipment and vaccines. The lesson was never institutionalised. NDS26 does not cite it.

The alliance queue. In a major Indo-Pacific conflict, US defence industrial output will be allocated in sequence: American forces first, treaty allies in the immediate theatre second, more distant partners thereafter. This is not speculation—it’s been evident in every major conflict since 1939, made explicit in the 2025 US National Security Strategy’s emphasis on domestic reindustrialisation and carried out in Secretary Hegseth’s delays of bought arms to Estonia last month⁴.

US prime contractors with Australian operations are not sovereign partners. In crisis, they are foreign assets subject to foreign priorities. The queue is real and Australia is not near its front.

The Ukraine lesson, partially read. NDS26 invokes Ukraine to validate the asymmetric, low-cost, drone-centric model it is building toward. The citation is selective. Ukraine’s extraordinary wartime industrial adaptation rested on three conditions Australia cannot assume: a pre-existing distributed manufacturing base; a dense network of small workshops capable of rapid adaptation under pressure; and—critically—a population that understood itself to be in existential danger and responded accordingly. Ukraine’s Brave1 program, cited approvingly in NDS26, engaged over 2,300 companies. That ecosystem was not conjured by a strategy document. It was built over years, from necessity, under threat.

NDS26 draws the technology lesson from Ukraine but ignores the social and structural lesson.

Sovereign industry is not built by procurement reform. It is built by sustained political commitment, across multiple electoral cycles, to capital-intensive programmes that produce no immediate dividend, carry enormous cost uncertainty, and whose strategic necessity citizens cannot easily evaluate because the relevant information is classified.

That’s evident through Australia’s own history of defence procurement. Programmes are initiated with ambition, then terminated under budget pressure. That’s the consequence of the structural interaction between democratic time horizons and industrial investment timescales. Defence industry asks governments to make thirty-year commitments in three-year political environments. That tension has not been resolved by any of the strategic reviews or defence white papers since 1986, including NDS2026.

If anything, the structural reality of Australia’s existing industrial landscape compounds that tension. Defence industry is dominated by subsidiaries of US and European prime contractors—Lockheed Martin, BAE Systems, Thales and others—whose investment decisions are made in foreign headquarters, subject to foreign strategic logic and shareholder returns. Calling their Australian operations ‘sovereign’ is a convenience that does not survive a crisis test.

NDS26 ‘tripartite’ partnership—government, industry, workers and, of course, unions—nods towards the right conceptual frame. But such a partnership does not exist. The incentive structures are antiquated and misaligned. The governance arrangements are immature, and subject to short-term political patronage. The long-duration commitments—and the patient capital—necessary to justify genuine private investment are missing at the required scale.

The alternative to prime-dependency is not a different procurement model. It is a different kind of country, industrially. That is the magnitude of need implied by NDS26.

To shift the centre of gravity toward genuinely sovereign capability requires, at minimum:

· long-duration commitments sufficient to justify capital investment and to nurture SME ecosystems;

· workforce pipelines built from the apprenticeship level, linked to specific capability requirements;

· research-to-production pathways that currently do not exist in most priority areas; and

· governance arrangements giving the Australian government real leverage over foreign-headquartered entities claiming sovereign status.

Finland, Sweden, South Korea, and Taiwan built sovereign industrial capacity over decades, sustained by consistent political will and publics that understood why it mattered. Each of those countries has a history that made industrial mobilisation feel necessary, not merely desirable. Australia’s history has not, so far, provided that forcing function. And while NDS26 may point in the right direction, it fails to reckon honestly with the distance.

The imminent Defence Industry Development Strategy must now do the work NDS26 does not: name the Sovereign Defence Industrial Priorities with real content, specify what ‘sovereign’ means in terms of Australian equity, design authority, and supply chain depth, and articulate how the government will sustain commitments across electoral cycles and nurture robust, innovative and scalable SME ecosystems. Aspirational language repeated from NDS26 will be a warning sign.

The hollowing-out took over forty years. The NDS allocates ten years to reverse it, on a budget being simultaneously compressed by AUKUS Pillar I. ‘Industry policy is security policy’ may be true, but so is its inverse. The absence of industrial policy has been Australia’s de facto security policy for a generation, accumulated not through decision but through deferred choice, outsourced manufacturing, and the comfortable assumption that allied supply chains would always be open.

Recognising the magnitude of what reversal requires is not pessimism. It is the precondition for a plan that might actually work. The primes will always offer to fill the gap—and they can offer good capability. But they will do so on their terms, for their shareholders, subject to their governments’ export controls. That has been the arrangement for forty years. It is why the base is hollow.

NDS26 acknowledges the problem but has not yet decided what to do about it.

1. Nor, as we are seeing now as the Prime Minister and Foreign Minister scout the region for oil supplies, does it need to be high-tech, but processed, refined, and integrated with other materials or components.

2. At least at the scale, consistency and adaptability needed.

3. Worse, still, as noted by the Chief Economist to MB Fund and MB Super in 2025, as per the latest available OECD data shows Australia with the lowest share of manufacturing of any OECD country; it is also the least self-sufficient country. And it has the lowest share of R&D expenditure as a percentage of GDP, at 1.68 per cent compared the OECD average of three per cent. That is a significant handicap in a viable defence manufacturing sector, and cannot be wished away by press releases.

4. Ironically, the government announced the purchase of additional HIMARS systems on 28 April 2026, a week after Secretary Hegseth informed Estonia, a front-line nation, that the United States was delaying its HIMARS systems shipments, because of the conflict in Iran. Unsurprisingly, European nations are reviewing their dependence on US capability.