Sovereign Capability from the Periphery

In March 2026, the Strait of Hormuz closed. It didn’t stay closed for long in the formal sense, but the chaos that followed (seizures, blockades, insurance spikes, sporadic attacks) kept oil flowing like treacle. Within four weeks, petrol prices in Auckland had risen about 40 per cent. Diesel had nearly doubled. GDP forecasts were slashed, and the government’s main structural response was a $50-a-week boost to the in-work tax credit for working families.

New Zealand has been a country with no refining capacity since Marsden Point shut in 2022¹. We import every drop of refined fuel, mostly from Singapore and South Korea. Onshore stock at any given time runs at about 30 days². The electricity system is hostage to rainfall, gas is in fast decline, there’s no LNG terminal, and demand is forecast to grow 80 per cent by 2050. These are the conditions of the country we already live in.

The Hormuz episode was a small moment. It will be easily forgotten. But it was a glimpse of an architecture that everyone in this country has been quietly assuming would keep holding. It will not keep holding. The supply chains that fed Aotearoa for 40 years were built for a world that has since walked off. The umbrella of allied military superiority has fewer ribs than it has had since before I was born. Distance, the asset that once made New Zealand quietly strategic and quietly safe, no longer buys what it used to buy.

The periphery is going to have to learn to build, because the centre is no longer reliably going to build for it. The conditions under which a small, allied, peripheral country could outsource industrial seriousness have ended. We’re going to make hard things ourselves, or someone else is going to make them and decide where we stand in the queue.

---

Software was a holiday from history³.

For a generation, the most lucrative and prestigious parts of the Western economy were bits, finance, and services. Industrial work was something other people, somewhere else, did for low margins and lower status. Energy was cheap. Supply chains were reliable. Strategic materials were something historians wrote about. Somewhere between Crimea in 2014, the pandemic in 2020, and the partial collapse of the post-Cold War security architecture in the years that followed, atoms came back to the centre of policy.

Energy, materials, propulsion, computing hardware, defence: these are the substrate of everything else. A country that can’t generate, store, and move power; that can’t make steel, fertiliser, semiconductors, or fuel; that can’t project deterrence into its own neighbourhood, is a country whose decisions get made for it by whoever does control those things. Industrial capability is the basis of sovereignty, and the phrase is no longer rhetorical: advanced semiconductor manufacturing now sits in a geographic band the width of Taiwan, rare-earth processing is roughly 90 per cent in one country, and that country has demonstrated, in the past five years, an open willingness to cut off particular minerals to particular customers. Global gas prices can spike an order of magnitude when a single supply route fails; Europe ran that experiment live in 2022.

The strategic environment has shifted to match. The rules-based order is fraying, the underwriter is increasingly transactional, and the freedom of the seas is, in several places, being actively contested. The Indo-Pacific is where the new map is being drawn. China has industrialised a category of warfare that allied militaries are still learning to define: 200-drone swarms operated by a single human, centrally directed autonomous fleets at scale, the recent flight test of Jiutian, a 16-tonne unmanned drone mothership whose internal bay is claimed to carry more than 100 loitering munitions. The deterrence math that kept the post-1989 peace assumed a particular kind of asymmetry between the United States and any plausible adversary. That asymmetry is narrowing in the platforms that matter most for any kinetic contingency in our region. Wars happen when the two sides disagree about who would win⁴, and the breadth of consensus that allied superiority would prevail is the thing that has held the peace.

New Zealand’s security has run, for decades, through the security of the Pacific, the security of our partners, and the credibility of the alliance architecture our partners maintain. None of those things are under our exclusive control. All of them are getting more expensive to maintain.

The capability stack that allied militaries actually rely on, in 2026, is a network of trusted private suppliers. The defence primes are at the top of it, but the resilience of the stack is set by the depth of the small and mid-sized firms underneath them: the autonomy software, the niche sensors, the specialty alloys, the propulsion subsystems, the secure communications, the critical minerals refining. That stack has been thinned out across most of the West by a generation of consolidation, offshoring, and underinvestment. Trusted allied-nation suppliers are now genuinely scarce. This is a door that small countries with technical talent and clean institutions can walk through. It doesn’t stay open forever.

---

New Zealand has been one of the largest net beneficiaries of Pax Americana, and we’ve paid almost nothing to maintain it.

Look at the maths. Two oceans of strategic depth from any plausible adversary. Five Eyes intelligence, FPDA, Closer Defence Relations with Australia, and the security of the alliance system without the price of admission. Defence at barely 1 per cent of GDP, one of the lowest sustained levels in the developed world. Intelligence and civil defence add another fifth of a percentage point. No conscription since 1972. No defence-industrial base of any consequence since the 1940s. A foreign policy we describe to ourselves as “independent”.

The countries that genuinely run independent foreign policies pay properly. Singapore caps headline defence at 3 per cent of GDP and spends another 2 percentage points across conscription opportunity cost, civil defence, sovereign R&D, and a state-owned defence prime. Switzerland reaches a similar level at 1.5 to 2 per cent all-in, and only because of mountains, 2 centuries of unbroken militia tradition, and 365,000 civil shelters covering every resident by law⁵. Sweden held 3 to 3.5 per cent of GDP through the Cold War to underwrite genuine non-alignment, took the peace dividend down to 1.1 per cent by 2014, watched its non-alignment doctrine collapse within a decade, and joined NATO in 2024⁶. The credibility of an independent posture compounds slowly, like trust, and disappears quickly when you stop paying for it.

What we call “independence” is the cheapest seat in the Western alliance, sold to ourselves as a principle. I am aware how this lands. I would retire the formulation if I could find one that wasn’t true.

---

This is a delicate thing to say in a country that has built so much of its modern self-understanding on the moral cleanliness of the 1980s. The nuclear-free movement was a moral achievement. France was still testing nuclear weapons in the Pacific while New Zealanders were marching, and the US Navy in that era didn’t separate propulsion from weapons. The instincts at the heart of the movement, against further proliferation and against testing of that kind, were right then and remain right now. So was the bundling that found its way into the 1987 Act. That world has gone.

The bill is coming due. AUKUS submarines will be operating in our neighbourhood inside the decade. Allied vessels are increasingly nuclear-propelled across the Five Eyes. Our exclusion from arrangements that were once routine is increasingly visible and increasingly absurd, and the strategic cost of that exclusion is rising every year. The energy conversation, similarly, has reopened. A country that has banned itself from even discussing civilian nuclear power, while watching its gas decline and its refining capacity disappear, is making decisions about its grandchildren on the basis of a 1987 piece of legislation that explicitly carved nuclear power generation out of its prohibition in the first place⁷. We’re allowed to read what the law actually says.

The argument is for abandoning the inertia that has been smuggled in under the principle, while keeping the principle itself. The propulsion question and the weapons question are different questions. The energy question is a third question. We’ve been running them all together for 40 years, and the result is a country that has stopped being able to make some of the decisions that small countries, in this period, are going to need to make.

---

We’ve done this before, and it wasn’t a fluke.

Sometime in the mid-2000s, Sir Peter Beck⁸ decided to build a rocket. He didn’t have the manufacturing base, the supply chain, the regulatory framework, or any obvious reason for it to be possible. He built it anyway. 20 years later, Rocket Lab is a NASDAQ-listed company that has put more orbital launches into space than any private company outside the United States. About a third of its revenue now comes from defence customers, and that share will go past half as the Neutron rocket comes online. New Zealand quietly became one of a handful of countries that can put hardware into orbit, because the people who did the work were good enough.

LanzaTech is the same story in a different field. A small Auckland biotech turning industrial off-gases into ethanol and aviation fuel, now NASDAQ-listed and licensing its technology into mills and refineries on three continents. Both companies are part of a 20-year deep tech lineage that has never been particularly well understood inside New Zealand and is much better understood by serious investors abroad. That lineage is producing a second generation now.

In Wellington, in late 2024, a private New Zealand company called OpenStar achieved first plasma in a levitated dipole fusion reactor. There are fewer than ten private fusion companies in the world with operating plasmas, half a dozen that matter. The reactor concept is a genuine alternative to the tokamak orthodoxy that has dominated fusion physics since the 1960s, and the team building it is good enough, technically, that it has earned the public support of senior figures at MIT and at the leading US fusion companies. Patient public capital sits alongside private capital in the round, on terms that work for both sides. This is the kind of partnership between state and private capital that small allied countries used to know how to do.

In Tauranga, an unlisted New Zealand defence company is closing in on 100 million in revenue, more than doubling annually, with hundreds of staff across New Zealand, the United Kingdom, and Ukraine. One of the world’s largest manufacturers of uncrewed surface vessels by recent production volume. Its platforms operate with the United Kingdom Royal Navy and with Ukrainian forces, and its software stack is hardware-agnostic, non-ITAR, and built to plug into the NATO+ capability network without any of the friction that comes with US export controls. It has competed against Anduril, BAE, and Thales for British Army contracts, and won the parts it should have won.

In a basement in Parnell⁹ within the same walls that Rocket Lab started, another New Zealand company is building catalysts and electrochemical cells that turn ethanol directly into electricity, with industrial demonstration units in design for European construction sites and a Japanese OEM as a paid commercial partner. Across the country, teams are working on hydrogen-based direct reduced iron for our own titanomagnetite ironsands, on geothermal precious metals, on geological hydrogen, on neodymium electrolysis, on AI-driven materials discovery, on autonomy software for multi-domain platforms that allied militaries actually want to buy. This is an industrial base, in latent form, much further along than anyone in Wellington gives it credit for.

Why does any of this happen here. The answer isn’t romance. New Zealand has a small number of structural advantages that are easy to lose sight of inside the country and easier to see from outside. A talent density that punches above its weight in applied physics, mechanical engineering, and computational chemistry, sustained by university departments small enough to be hands-on rather than hierarchical. A regulatory instinct that, when it has been allowed to operate, has been willing to write new rules instead of waiting for precedent: the country licensed the world’s first private orbital launch site, and Rocket Lab’s existence is downstream of that single decision. A geography that genuinely is an asset for some kinds of work, in particular range testing for autonomy and aerospace, where low population density, diverse terrain, and a 500-kilometre-plus sea range are the relevant inputs. A Rocket Lab and LanzaTech alumni base now seeding the next generation of companies, with the kind of operational scar tissue you can’t manufacture from a graduate programme. And the ability to know a country’s whole technical community by first name, which is a strange thing to be able to claim, but it changes how fast hard things can move.

None of these advantages is decisive on its own. Together they describe a country that, when it actually decides to build something hard, has fewer of the structural disadvantages than its self-image suggests.

The structural argument that “New Zealand is too small to do this work” is empirically false. We’re already doing it. What we haven’t done is decide, as a country, that the work matters.

---

What needs building is mostly the deliberate version of what’s already happening.

A serious deep tech industrial base is a small number of sectors, capitalised and regulated as if they mattered. Six are already visible from where we stand. Energy, including the honest reckoning with civilian nuclear power, both fission and fusion, that the 1987 Act never actually prohibited. Defence and dual-use, including autonomous systems, sensors, communications, and the kind of niche, high-trust integration work that allied militaries are going to keep needing for decades. Critical minerals, where our titanomagnetite, geological hydrogen, geothermal mineral resources, and processing know-how feed directly into the Five Eyes effort to disentangle from Chinese supply chains. Advanced materials, including the AI-driven discovery that, for the first time in chemistry, makes it economic for a small country to compete at the frontier. Propulsion and aerospace, where Rocket Lab and the ecosystem it has seeded already give us an unfair advantage. Advanced computing, narrowly: hardware and architecture rather than another wave of consumer software companies.

Industrial capability here is two kinds of work at once. Factories across both islands making what the country buys. And the technologies we develop here and export to solve problems our allies and trading partners are stuck with: rare-earth dependencies they want to break, toxic tailings they want to remediate, emissions-heavy industrial processes they want to decarbonise, old-economy approaches with poor unit economics that a better technology can replace. The factories and the exports are one project. A small country builds industrial capability the same way a small company builds revenue: by being a supplier of things the rest of the system needs.

A handful of moves would matter most. Treat the deep tech industrial base as a bipartisan strategic priority, not a discretionary innovation programme vulnerable to cuts every cycle. Stand up a dedicated allied flight-test range with a defence and dual-use mandate from inception, capable of supporting work that allied companies currently can’t do at home; New Zealand has the geography, what it lacks is a facility set up with the right scope to host it. Reform investment screening so that it accelerates trusted allied capital instead of slowing it. Build, with allied partners, the kind of industrial-scale neutron capability that fusion, defence, and advanced manufacturing all need and that no single country can fund alone. Anchor public capital into industrial-scale piloting infrastructure, co-located with our existing heavy-industrial sites, because the most capital-intensive phase of deep tech company-building currently happens offshore by default, and that phase is where the talent, the supply chains, and the strategic capability accumulate. And, with respect to the 1987 Act, do the boring legal work of separating the propulsion provision from the weapons provision in the public mind: the weapons ban can stand, but the propulsion ban is a Cold War artefact that costs us alliance access, energy options, and strategic relevance, and we’re the only country in the world that confuses the two.

Countries with much smaller industrial bases than ours have already made these moves in the past decade, in response to the same set of conditions. The reason we haven’t is the absence of a political and intellectual permission to treat industrial capability as serious work.

The role of the state is worth saying out loud. Patient public capital, alongside private capital, on reasonable terms, has been part of every successful small-country deep tech story of the past 40 years. It built Israel’s defence technology base. It built South Korea’s semiconductors. It built Singapore’s pharmaceutical and biotech sector. It rebuilt the United States’ own semiconductor and aerospace base, again, after a generation of letting them slip. The small set of strategic technologies that determine New Zealand’s options for the next 40 years are not going to capitalise themselves through the New Zealand venture market alone. The state’s role is to be the patient anchor that lets the private market do its job around it.

---

Two thoughts on what any of this is ultimately for.

The first is that the technologies carry the values. Tesla decarbonised road transport more than two decades of climate rhetoric did, by selling cars people wanted that happened to be electric. The values were embedded in the product, not in the messaging, and they travelled to people who didn’t share them. The same logic runs through the deep tech base above. Autonomous systems built with meaningful human control designed in from the start carry a view of how machines and humans should share lethal authority. Mineral processing that doesn’t poison rivers carries a view of what extraction owes to the places it happens in. Ethanol-to-electricity that works in real industrial settings carry the view that decarbonisation is engineering, not deprivation. These are values this country reaches for at its best: a real attachment to individual freedom and the rule of law, a discomfort with concentrated power, a fairness instinct that runs to outcomes as well as procedure, a Treaty-shaped habit of negotiating sovereignty rather than imposing it, and a practical environmental stewardship that is older than the word “green”¹⁰. They deserve sharper articulation than they usually get, and they are worth sharing.

The second is that we should actively want these technologies adopted as widely as possible, including in places we are otherwise contesting. Interdependence is good. Architectural dependency on a contesting power for critical capability is not. A Chinese steel mill that decarbonises with a New Zealand-developed direct reduced iron process is, in any honest accounting, a better outcome for the planet and for the long-run prospects of peace than the same technology used at home while the Chinese mill keeps emitting. The same applies to mineral processing that doesn’t externalise its waste, to autonomy software that constrains itself to international humanitarian law by construction rather than by promise, to materials science that lets old industries clean up without going broke. This is where the security argument and the diffusion argument stop pulling against each other. The deterrent edge buys the time and political stability inside which the diffusion does its longer work. The diffusion is what the deterrent is ultimately for. The destination is not a permanently fortified periphery; it is a world in which the technologies that carry these values have been built and adopted widely enough to outlast any particular contest.

---

To New Zealanders. The nuclear-free movement was, and is, a moral achievement. So was the period of independence from the British framework that the 1980s ratified. Stop letting inertia extend those movements into territory they were never designed to occupy. The free-riding has stopped being clever. The energy gap has stopped being abstract. The world that paid for our independence has stopped being interested in subsidising it. The honest version of the New Zealand identity is the one that decides, as a country, to start contributing to the thing that has kept us safe and prosperous for 80 years.

To Five Eyes readers. New Zealand is not asking to be treated as a peer of the major industrial bases. Japan, the United Kingdom, the United States, and Korea carry industrial weight at scales this country does not and won’t. The case is narrower and more useful. The alliance system needs sovereign capability distributed beyond the primes, in places where small democracies can move quickly on technologies that matter. New Zealand has a credible offer in orbital launch and space services, autonomous maritime and aerial systems, sovereign sensing, trusted compute, and selected advanced materials with defence and dual-use applications. The geography is real: Southern Ocean access, Pacific positioning, southern-hemisphere ground-station latitude. The regulatory instinct is real. New Zealand’s space regime has been imitated more than it has been criticised. The alliance access is real and underused. What’s missing isn’t the underlying capacity. It’s the political and intellectual permission to treat this as work: to say out loud that a small democracy in the Pacific can be a serious contributor to allied technology resilience, and to organise the country around being one. That permission is what this argument is asking for.

We’ve built hard things in this country before, in conditions worse than the ones we face now. Whether we choose to do it again, deliberately and at scale, is the only question left.

---

AB · Auckland, April 2026