IN DETAIL: How do you build a defence industry that can deliver quickly and at scale?

On a recent visit to Anduril Industries’ HQ in Costa Mesa, California, the company’s co-founder Palmer Luckey made a very important point: if you want to build something in large numbers, you need to tailor it to the capabilities of your industry base.

Gregor Ferguson

“A country can only mass manufacture arms if it designs arms that can be mass manufactured by the industrial base it already has,” said Palmer Luckey bluntly.

And he cited the US track record in World War 2 to back himself up. Before World War 2 the US was a world leader in manufacturing fairly undemanding things like cars and farm and construction equipment, but in huge numbers. So, what can you make in an automotive plant? America’s production engineers were very specific: what thickness of metal can we weld? What gauge of steel can we bend? What radiuses can we machine or bend metal to, using equipment that was designed to manufacture cars, industrial equipment and farm equipment?

It was these questions, and the answers to them, that allowed American engineers to go into John Deere tractor factories, Caterpillar bulldozer factories, Ford and General Motors car factories, and turn them into aircraft and tank factories.

“A lot of people think that we turned them into tank factories,” said Palmer Luckey. “But we actually designed tanks to be made by car factories. It was really the reverse of what people imagine.”

That’s what Anduril is trying to do, he says: “We’re trying to build submarines that can be built in car factories. We’re trying to build cruise missiles that can be built in farm implement factories.”

Because if he or one of his customers needs to make 100 times more things than they currently do (if they even make them at all), they won’t be able to do it in a brand-new facility that needs specially trained people with all-new skills.

“It’ll take you a generation to grow those people and teach them,” points out Luckey. “So, my strong belief is weapons must be aligned with the capabilities of their host nation.”

Which brings us neatly to Australia.

The Department of Defence has an explicit policy goal: to create a robust, sustainable defence industry sector that can both support the Australian Defence Force (ADF) and its allies and become a high-technology pillar of a stronger national economy.

Australia is a world leader, or at least world-class, at R&D in many things: autonomy, robotics and quantum science, for example; but it has almost never commercialised any of these technologies at scale. Although we do outstanding research in these and other vital areas, our industrial base is vanishingly small.

Yes, we can manufacture anti-ship missile decoys and naval surface-air missiles and build components that go into every F-35 Lightning II that’s built, but we currently lack the resources and supply chains to make more than that small contribution to our own and collective defence.

Incidentally, what do these programs have in common? While you’re thinking about that I’ll continue.

A new strategy

In 2024 Defence published its Defence Industry Development Strategy which addressed everything from procurement reform to prioritisation, innovation, workforce development and exports. It acknowledged the strategic challenges that Australia faces but didn’t set financial targets, nor timescales, except for the establishment of Australia’s Sovereign Defence Industrial Priorities: it named two Epochs – Epoch 1, from 2023 to 2025; and Epoch 2 from 2026 to 2030.

By 2030, Defence wants that industrial base to build a network of MRO depots, manufacture parts for nuclear-powered submarines (and eventually whole submarines), manufacture complete guided weapons in-country, establish an autonomous systems capability, and so on.

In 2022-23, Defence awarded nearly 52% of all Australian government procurement contracts. Defence accounted for $38 billions-worth of contracts of which just over $8 billion went to Australian Small and Medium Enterprises (SMEs), defined for our purposes as companies employing less than 200 staff – that is, 99.8% (I’ll say it again – 99.8%) of all Australian employers.

The Strategy included a diagram setting out what the three Tiers of Australia’s defence industrial base look like. It’s a pyramid, basically, with a small Tier 1 at the top, a broader Tier 2 half way down and the whole is supported by a thick layer of Tier 3 companies at the bottom. Most of the dozen or so Tier 1 prime contractors are foreign-owned; most Australian SMEs operate as Tier 3 companies providing parts, consumables and services; a very few Australian companies operate in Tier 2, delivering major equipment, systems, assemblies and so on to support the primes. The majority of companies in Tier 2 are foreign, or foreign-owned.

A key aim of the Strategy is to grow the number of Australian businesses in Tier 2. Another aim is to increase the amount invested in the defence industry by Australian (and overseas) sources of capital and venture funding. And this is where we run into a problem – and so does Anduril, and so do other companies like the longer-established prime contractors.

Firstly, Defence is reluctant to back up its innovation funding (and there isn’t a huge amount of that) with production contracts; secondly, even when it does order materiel an Australian defence order is very small by global standards and it may not be repeated quickly, if at all.

So, Australian funding and venture capital sources tend to look for opportunities where the technology inherent in the product is dual-use, or where there is a strong export market (and ideally both) and the chance of repeat orders. If they’re going to fund R&D and then commercialisation they have to be certain the market exists – in volume, that the companies they’re investing in have a good chance of claiming a significant proportion of that market, and that they can make some decent money on the investment.

There aren’t many such opportunities and if Canberra says it respects free markets, then it also has to respect funding sources’ reluctance to invest exactly where Canberra wants.

By the way, the three programs I mentioned earlier have one thing in common: a huge export market which dwarfs Australia’s peacetime domestic market. That’s what makes the BAE Systems Australia Nulka decoy and its tail section for the Evolved Sea Sparrow Missile, and the Australian supply chain for the Lockheed Martin F-35, viable – indeed, profitable.

There’s a third factor also: Defence’s decision-making and processes are abysmally slow. Bodies such as ASCA (the Advanced Strategic Capabilities Accelerator) and the Royal Australian Navy’s Maritime Integrated Capabilities (MIC) innovation branch are trying valiantly to hasten these, with some success, and Defence is now saying repeatedly that ‘speed to capability’ and everything that flows from that mindset is now paramount. But getting anything to happen quickly in Canberra, in anything short of an outright war, still seems to be extraordinarily difficult so observers are suspending judgement.

So companies like Anduril Industries, on the one hand, and Lockheed Martin, on the other (to take just two examples), are faced with an interesting set of problems.

Interesting times

Lockheed Martin has been building its proven Guided Multiple Launch Rocket System (GMLRS) in the USA for years. This is a sophisticated weapon with a well-established American supply chain. The company has put forward a plan to Defence to build the GMLRS in Australia. Defence is expected to name the Australian manufacturing site before the end of this calendar year and Lockheed Martin says it will eventually build some 4,000 GMLRS a year there (the company aims to build some 12,000 a year in the USA) and in due course the same company’s Precision Strike Missile (PrSM) Increments 1 through 4.

They will be identical to the ones made in the USA and at a comparable price – but these are relatively complex weapons and it has taken a while to build the skills and capability of the Australian supply chain. That will be a constant problem for sophisticated equipment – and for some ‘older generation’ equipment, too: building the skills and the supply chain to manufacture them in-country takes time and money.

Some 4,000 GMLRS a year is far more than the Australian Army can use, but Lockheed Martin says export sales to the US Army and other approved export customers will soak up the excess. Lockheed Martin is paying for some of it, but the majority of the investment will be by Defence, through its Guided Weapons and Explosive Ordnance (GWEO) Enterprise.

Anduril has a $1.7 billion contract from Defence for an undisclosed number of Ghost Shark Extra Large Autonomous Underwater Vehicles (XL-AUV) to deliver on. The RAN should get its first production Ghost Shark in January 2026. To be built in an all-new factory in Sydney, the company will build ‘dozens’ of Ghost Sharks for the RAN and approved export customers as well as the civilian version, known as Dive-XL. Awarded in September, the contract comes only a couple of months after the end of a very rapid, three-year R&D and prototyping program funded equally by Defence and Anduril. One of the reasons why the program was so quick was that it was co-funded by a commercial organisation with skin in the game and a genuine civilian as well as military market for the end-product.

On current plans just about every Ghost Shark and Dive-XL built for the RAN and for export and commercial customers will come from Sydney. A mirror factory on the east coast of the USA will build Ghost Sharks, Dive-XL and its smaller brother, the Dive-LD, for US military and commercial customers. So there is the promise of sufficient volume to keep prices down and to make it worthwhile investing in manufacturing infrastructure in Sydney.

The thing about Anduril’s Ghost Shark, like its Barracuda cruise missile, like its Copperhead AUVs (they are torpedo-like, but not exactly torpedoes) and its YFQ-44A Fury Collaborative Combat Aircraft (CCA), is that they are designed to be built rapidly and in numbers by a workforce that isn’t necessarily attuned to traditional Defence requirements but knows how to mass-produce things.

The Ghost Shark, although highly classified, uses a lot of simple fibre glass on a steel frame; it is square in section because it is easier to manufacture that way. And it is ‘wet’ – most of it is flooded, partly because it has opening doors for (unspecified) payloads, and partly because the production difficulties in manufacturing perfectly cylindrical hull sections that keep water out at great depth are immense. Only things like the guidance electronics in the nose and the battery pack at the stern employ pressure vessels to keep them dry at depths of thousands of metres. And they can have as few as one or as many as three payload bays carrying anything from torpedoes to sonars, or smaller AUVs, or anything else an end-user can imagine.

Unlike traditional cruise missiles, the Barracuda 500 cruise missile is designed to be assembled quickly using just 10 different hand tools. It reportedly costs around US$216,000 ($333,000) and is now being manufactured in Taiwan, according to the Asian Times.  The Barracuda 500 makes extensive use of COTS – Commercial Off The Shelf – equipment and its outer skin employs a couple of one-piece fibre-glass panels which include all of the fasteners necessary to attach them. As the head of Anduril’s maritime business, Dr Shane Arnott, says, if you can build a fibre-glass bath you can be part of a missile supply chain.

The YFQ-44A Fury, which flew for the first time only a few weeks ago, also uses a lot of fibre glass; and its retractable landing gear is made from aluminium and steel castings, not from exotic alloys which are very costly and difficult to fabricate or mould. Basically, any competent manufacturer can make most of the Ghost Shark, Barracuda and Fury: they are designed to be cheap and to be manufactured at scale.

And it’s not just Anduril that is challenging older business models: Boeing Defence Australia’s MQ-28A Ghost Bat CCA (for which Shane Arnott was lead before he joined Anduril) doesn’t use exotic stealth materials – it is semi-stealthy and derives its low observability from its shaping. It’s non-afterburning jet engine is a civilian power plant, reportedly off a business jet (Defence hasn’t said – the Fury also uses a civil engine, reportedly). And it uses one-piece composite mouldings for ease of assembly and servicing.

It isn’t disposable, but it is attritable – you’d rather lose one of these than a crewed aircraft. And it has completed more than 150 test flights and 20,000 hours of digital twin testing compared with the US CCA program which only made it first flights this year.

The 11.3m Leidos Sea Archer Uncrewed Surface Vessel (USV), the second example of which is already under construction near Newcastle, NSW, is also an attritable design: it can be sacrificed, if necessary, so there’s no point in spending heavily on it. It is made of aluminium sheet and can be fabricated by any boatbuilder; it has a 1,500nm range at its most economical speed, has a 900kg payload and a sprint speed, thanks to its two diesel outboard motors, of more than 40kt. And its Leidos LAVA platform autonomy system, communications and EO-IR system are all designed to be good enough, but cheap enough to be dispensable if necessary. And like every example cited here, it is designed to be manufactured quickly, at scale.

Manufacturing at scale

Manufacturing at scale is one of the emergent defence industry preoccupations; it is also becoming important in the eyes of some Defence customers: the government of Ukraine, most famously; but also those of the United States and United Kingdom and, to a much more cautious extent, Australia.

So what does manufacturing at scale mean? It means, very simply, mass production: producing lots of identical things, quickly, and then making an improved version cheaply and quickly. It means ramping up production rates for defence equipment quickly – which means in turn repurposing factories and workforces that might have made baths or truck bodies or ute trays (pickup truck trays to the uneducated). You can’t do that with an exquisite system such as an aircraft or warship, still less with the traditional missiles they carry, or which are designed to destroy them, though they are still necessary.

You need lots and lots of missiles, at a time when global supply chains might be constrained (at best) or completely non-existent. You have to be able to build them yourself, and quickly, and in numbers. And to do that you need to make them cheap and easy to manufacture.

These are the realities confronting Defence. The defence industry will do what it’s told – and funded to do – and companies will join the defence industry if there’s a prospect of making some serious money (and there may only be Defence money for a while), and partly also for patriotic reasons.

People criticise the Defence focus on nuclear-powered submarines and Hunter-class frigates, forgetting a couple of important things: firstly, we’ll need these even more in 20, 30 and 40 years’ time (so, arguably, we should have invested in them several years ago); secondly, the Australian defence budget is artificially low and could be increased quite significantly to pay for equipment that we need now and in the next couple of years as well; thirdly, however, even if we had that extra money the robust domestic supply chains that would support an Australian missile or armoured vehicle factory don’t exist properly as yet and it’ll take a while to create and refine them.

Again, we should have invested a lot more money several years ago. We don’t build solid rocket motors in huge quantities for missiles, for example; we don’t build electric motors in huge quantities for drones – in Ukraine they go through more than 10,000 a month; we don’t build batteries in huge quantities, either.

That said, the ASCA approach, now that it is part of the Vice Chief of the Defence Force’s Group, is to set problems, describe a capability which it wants the ADF to field quickly, and then calls for industry’s ideas and proposals. It doesn’t quite ask for blue-sky thinking on the part of industry, but it comes close. And it has set the bar high to deter frivolous proposals.

This is a good way of getting industry minds to focus on a problem and then come up with solutions that might be straight from left field, or build on something that might already exist.

The only problem with this approach is that it leaves the power of genesis – identifying, first, a problem, and then the capability required to address it – to the VCDF Group. Industry becomes a service provider – a smart service provider, possibly, but it is reactive and not pro-active.

It’s probably harder now than before for an Australian defence company to look at a problem and then sell a solution of its own that is either much better or that saves lots of money, when the VCDF Group hasn’t thought of it first. The only way to achieve such a thing is to build that solution for an export and/or dual use market, and that’s an expensive business: it’s an activity that’s usually confined to cashed-up primes and funding sources.

Quantity, quality and speed

This article is about the need and the ability to create a defence industry sector at scale, and quickly. In the famous triangle of attributes, we have speed, quantity and quality: we can focus on two of these but not the third. Which do we focus on?

It depends, of course, on what you’re trying to do: if you’re building GMLRS and eventually PrSM missiles to somebody else’s design, and with a proven business model, you probably focus primarily on quality and quantity. Why? Because you need to match what the prime contractor is doing in his home country and build the missile at the same price point – which also means building it at scale. This can be done relatively quickly, but it takes a lot of cooperation and good will.

If you’re building Barracuda 500 cruise missiles you would probably focus more on quantity and speed, having assured yourself first that this weapon will reliably deliver the effects you want, if built properly. And that’s the thing: building it properly shouldn’t take the combined efforts of a bunch of talented and highly trained experts who take years to train.

The emphasis here is on quantity and speed: speed to capability, in numbers. You don’t know how your new product will develop and evolve, or even if it will, so a Minimum Viable Capability (MVC) is the best starting point for that subsequent development journey and you can get to an MVC relatively quickly.

Quantity? The more the better. You’ll certainly know about it if you start to run out.

And that’s the paradigm confronting end users and industry today: if you haven’t got it, but you expect you’ll need it, you had better start stockpiling it if you can’t make it yourself – and even if you can, but expect usage rates to be high; or making it yourself right now (if quantity and speed are most important); or starting now to build something more complex, if quality and quantity are what is needed to get the project done.

Note that quantity is in every calculation: quality and speed can be dependent variables, depending on what we’re trying to achieve.

Quantity, quality and speed, however, mean different things to different nations: “We are designing weapons that can be made very easily in Taiwan,” says Anduril’s Palmer Luckey, but it would be hard to make some of those things in the US, he points out.

“And I think that there’s things that probably we wouldn’t manufacture in Australia, because you guys don’t already have a workforce, an industrial pipeline to make those things,” he adds. “The more stuff you guys are making, the easier it is for us to make more stuff.”

Good point. Now, what does Australia mass-produce in-country for any application, military or civil? Because that may determine what defence equipment we can mass produce, at least in the short term. Because this may end up as all the time we have.