Australia Wants to Sell. Washington Holds the Licence.

Week 34 | August 2026

Australia has decided it wants to be an arms exporter of a particular kind: a seller of the autonomous systems it builds for its own navy and air force. The National Defence Strategy (NDS) said so directly back in April, naming the Ghost Shark underwater vehicle and the MQ-28 Ghost Bat drone as sovereign capabilities with significant export potential. Anduril has been turning out Ghost Sharks in a Sydney factory since October, and the first of them reached the Royal Australian Navy in January. The sales campaigns have run since March.

So the ambition is real, and it is not new. And still, nobody has bought one. Not a single complete Ghost Shark or Ghost Bat has gone to a foreign buyer. What sits between the ambition and a signed contract is a clause of United States export law, and it hands Washington a say over sales that have nothing to do with the United States.

The language sits in Chapter 8 of the NDS, released in April. Paragraph 8.32 states that domestic production of autonomous capabilities such as Ghost Shark and the MQ-28A Ghost Bat will let Australia design, engineer and produce sovereign uncrewed systems "with significant export opportunities." Read that as an industrial policy statement wearing a capability costume.

The strategy is candid about why exports matter. Paragraph 8.8 concedes that many Australian defence firms will need revenue beyond what the ADF alone can provide, either from adjacent industries or from global sales, simply to stay viable. A sidebar points straight at Europe, where the ReArm plan and its hunger for uncrewed systems get named as a chance to grow the local industrial base. Then in July, DIDS26 sharpened all of it, lifting Ghost Shark from example to priority. Canberra wants to build these things at home and sell them abroad.

All of this is not theoretical. In April, Canberra and Tokyo signed a government-to-government arrangement on the MQ-28, covering data-sharing and Japanese observation of Ghost Bat flight testing, and building on a framework their defence ministers agreed the previous September. It stops short of any export or production commitment, which tells you something in itself. Ghost Bat is the first combat aircraft designed and built in Australia in over 50 years, and Boeing wants it sold abroad.

The previous month, Rheinmetall and Boeing paired up to pitch Ghost Bat as Germany’s collaborative combat aircraft, the German firm leading the integration. Berlin is weighing it against a genuinely crowded field of rivals, so nothing is won. An Australian-built aircraft is in the running for a major European buy.

Then came the flight that gave the game away. On 27 May an MQ-28 flew outside Australia for the first time, three sorties at Point Mugu in California. Point Mugu is a United States Navy range. The venue was the message.

Anduril has made the same moves underwater. Its president spent time in Tokyo in March setting out the Ghost Shark production model, Japan named as a prospective customer. Pat Conroy, the Australian Minister for Defence Industry, confirmed in October that Ghost Shark was a live campaign for the new Australian Defence Strategic Sales Office, and that he had briefed the Trump administration on it.

The fine print matters, though. Frameworks and pitches are not signed contracts. Not one complete platform has been sold.

Here is the part the Aussie strategy leaves out. The systems Australia wants to sell run on American software, and American software carries American rules.

Start with the AUKUS exemption, waved around as the thing that frees defence trade between the three partners. It is real. Since it took full effect at the end of last year, licence-free trade moves fairly freely between verified entities in Australia, the United Kingdom and the United States. Read the text at 22 CFR 126.7, though, and the boundary is explicit: the activity has to be to or within the territory of one of those three countries. The exemption is a bubble. It says nothing about selling onward to a fourth country.

The State Department has been blunt about what that means. Any defence article made from American ITAR-controlled data, its regulator wrote, may only be moved onward with a separate licence. When the final rule landed in December, the department said it again: the exemption exists to move goods among the three AUKUS members, not to enable reexports from them to other countries.

Now hold Ghost Shark against that rule. The hull is Australian. The autonomy stack, the part that makes it a Ghost Shark rather than a tube of batteries, is Anduril's Lattice software, and Lattice is American IP under ITAR. Built in Sydney, brain licensed from California. So any sale to Japan, South Korea, Singapore or Germany, none of them AUKUS members, needs a green light from Washington. A country trying to grow a sovereign export industry has handed a foreign capital the veto over its best products.

The people who track this call it taint. ASPI’s Tom Corben has written that even a sliver of American content can taint an allied system, and that the taint has already stopped Australian firms delivering custom versions of their products to American forces. Excellent technology does not get you around it. The paperwork decides.

Compare Israel. IAI sold its BlueWhale unmanned submarine to Germany, codeveloped there with the German firms TKMS and Atlas Elektronik, and to Greece. The German partner is not the point. No American firm sits in the design, so no ITAR and no third capital whose sign-off you need. That is what a platform without US strings looks like. Ghost Shark carries them.

None of this means Australia cannot sell hard defence kit abroad. In June, Canberra signed a A$2.5 billion arrangement to sell its over-the-horizon radar to Canada: the first international sale of the JORN system, and the largest defence export in the country’s history. Genuinely sovereign, Australian-designed, delivered through BAE Systems Australia. It settles the question for radar.

The rest is softer than it looks. The billion-dollar Boxer sale to Germany gets cited a lot, but Boxer is a Rheinmetall design assembled at a plant in Queensland. Century Engineering, an Adelaide firm, just became the first Australian company into the American nuclear submarine supply chain, machining carrier components through Newport News. Good news in itself, and the inverse of the export dream: Australia as a parts supplier to a foreign prime rather than a seller of its own platform.

The cleanest precedent is the quietest. Ocius sells its Bluebottle uncrewed surface vessel, an Australian design, and New Zealand already runs two. Bluebottle carries far less American content than Ghost Shark, which is why it crosses a border without a call to Washington. So the pattern almost writes itself. Australia exports well when it owns the whole thing, or when it sells a part. The crown jewels of autonomy sit in the hardest category there is.

Everything on the Australian side is working. The technology is in the water and in the air. Canberra has named the ambition out loud, its sales office is running live campaigns, and the minister has briefed Washington in person. Capability is not what holds back the first deal.

The block is structural. Part of it lives in export-control architecture written in another country. Part of it is a tension Canberra has not resolved: how do you sell the crown jewels while you are still fielding them for your own navy? Conroy said in October that Australia has to be stringent about who gets these systems. Fair enough. That instinct and a hard export push pull in opposite directions.

Watch one thing above all. If the United States State Department issues a licence for a specific Ghost Shark or Ghost Bat sale to a buyer outside AUKUS, the fourth-party problem is beatable in practice and the export story turns real. Until then, "significant export opportunities" is an aspiration with a foreign veto stapled to it.

We leave defence for the blue economy and follow the money into kelp. Carbon credits are being sold on the promise that seaweed locks carbon away for good. We look at what the science actually supports, and where the gap between the claim and the carbon really sits.

Since you have been, thanks for reading.

Cheers,

Mick

Ocean Tech Intelligence provides informational analysis only. Nothing in this publication constitutes financial, investment, legal, or strategic advice. Readers act on this content at their own risk. For full details see our Disclaimer.