The Undersea War

Elisabeth Braw has written the best account yet of the contest on the seabed. Read from Australia, the question it leaves is who would come to mend our cables, and what we would do while we waited.

Most Australians know the story of HMAS Sydney and the Emden . Fewer remember why the German raider was at the Cocos Islands in the first place. At daybreak on 9 November 1914, Captain Karl von Müller put a landing party of fifty men ashore on Direction Island to destroy the wireless station and the undersea cables that linked Australia to Britain.

The station’s operators had time to send a warning before the Germans reached the building. Part of it, “strange cruiser at entrance to harbour”, got through the jamming to the first convoy of the Australian Imperial Force, then steaming north from Albany, and Melbourne sent Sydney to look. By half past eleven, after being engaged by Sydney ’s guns, Emden was beached on North Keeling Island. The landing party, meanwhile, had cut two of the station’s cables but could not find the third.

We remember the gunnery, but the more useful lesson is the target. The first battle fought by a ship of the Royal Australian Navy began as a raid on a cable station, because the German Admiralty understood, as the British had when they cut Germany’s cables on the first day of the war, that whoever can sever a country’s communications can blind it.

I have been thinking Direction Island since finishing Elisabeth Braw’s excellent new book, The Undersea War: The International Battle for Supremacy at the Bottom of the World . Elisabeth is a senior fellow at the Atlantic Council , where she leads its work on threats to the global maritime order, and she reported the book from Svalbard, the Baltic, the Matsu Islands off Taiwan, Indonesia and Tanzania.

I have enjoyed most things Elisabeth has written, finding them thought-provoking and challenging. Indeed, Elisabeth has helped shape much of my work on national resilience, including my 2024 ASPI report, National resilience: lessons for Australian policy from international experience , in which I acknowledged her assistance.

With that said, in my view her book is the best account yet of the contest on the floor of the ocean. Read from Australia, it leaves one question that it does not set out to answer, and that this post is : if our cables were cut tomorrow, who would come to mend them, and what would we do while we waited?

The kingdom on the seabed

Braw calls the infrastructure on the ocean floor the ‘Undersea Kingdom’. It is roughly 1.6 million kilometres of fibre optic cable, carrying somewhere between 95 and 99 per cent of the world’s international internet traffic, plus 30,000 kilometres of pipelines and power lines. The book starts where it should, with the first transatlantic cable of 1858, over which Queen Victoria’s greeting to President Buchanan took the better part of a day to arrive, and moves briskly through the industry that grew up around it: a small number of manufacturers, a smaller number of specialised ships, and, increasingly, a handful of American technology companies that now own the cables outright rather than buying capacity on them.

The strength of the book is its reporting, from the places where cables have been broken and the people who had to work out why. Her chapter on Russia’s shadow fleet, the ageing tankers that sail under flags of convenience (or none) to carry sanctioned oil, is the best treatment of the subject I have read. She walks through the incidents in the Baltic one by one: the Hong Kong flagged NewNew Polar Bear dragging its anchor across the Balticconnector gas pipeline in October 2023; the Chinese bulk carrier Yi Peng 3 , shadowed but never detained after two data cables were severed in November 2024; and the Eagle S , the Cook Islands flagged tanker that the Finnish police and coastguard boarded in December 2024 after the Estlink 2 power cable and four data cables were damaged. In each case the shape of the evidence is the same. A vessel loiters, slows, drags something heavy along the bottom for tens of kilometres across exactly the right spot, and then insists it was an accident.

Braw is careful what that evidence can carry, while accepting that a link to the Kremlin or to Beijing usually cannot be proved beyond doubt. The ambiguity is the purpose of the method. The undersea contest, as she put it in an interview with New Eastern Europe earlier this month, “really is a war of subversion”. Russia, she argued, “wants to harm our countries without having to use military force”, because “if it doesn’t use military force, NATO isn’t in charge of responding”.

The legal chapters explain why prosecution is so hard. A coastal state has full jurisdiction only in its territorial sea, out to twelve nautical miles, and limited rights beyond it . Captains of the shadow fleet know this and stay outside the line. The Finns showed what a determined state can do when they boarded the Eagle S ; a Helsinki court then dismissed the case because the damage occurred outside Finland’s territorial waters. A Swedish prosecutor dropped the Vezhen case because the crew’s intent to release the anchor could not be proved. Braw builds her accounts of these chases through interviews with the officers and prosecutors involved, noting they rarely end in a conviction.

Chapter 4, ‘Nord Stream explodes’ carries the book’s argument attribution in its starkest form. In September 2022 almost everyone assumed Moscow had destroyed its own pipeline; the charge German prosecutors eventually laid was against a Ukrainian officer. If an explosion on that scale, in the Danish and Swedish exclusive economic zones, could take years to attribute and point somewhere few expected, a dragged anchor in the Indian Ocean may be harder still.

Chapter 13 on deep seabed mining raises even deeper questions. Cables, pipelines and nodules all lie on the same seabed and raise the same question, which is who makes the rules for the bottom of the ocean and who enforces them, and for Australia and the Pacific it is the chapter closest to . I will return to this issue below.

Elisabeth has written a work of reporting, and the policy argument cannot start until someone has established, as she has, case by case, what actually happened on the seabed and why governments could not act on it. She has made her prescriptions elsewhere, in an Atlantic Council brief on the Baltic that calls for operational relationships across navies, coastguards and cable owners, shared regional response protocols and joint repair capability, and in her argument for radical transparency with citizens what is happening under the water. I hope that her other unanswered challenge, what to do the concentration of ownership in a few companies with interests of their own, will follow in subsequent work. That question, and the ones that matter most for a country at the far end of the network, are where an Australian reader should look to pick up the argument.

The other half of the kingdom

Braw’s mining chapter follows The Metals Company and its chief executive, Gerard Barron, through the campaign to harvest polymetallic nodules (potato sized lumps rich in nickel, cobalt, copper and manganese) from the floor of the Pacific. Its sharpest passage, is her case against President Trump’s executive order of 24 April 2025 to fast track American permits for seabed mining. Under the United Nations Convention on the Law of the Sea, the seabed beyond national jurisdiction is administered by the International Seabed Authority. Braw argues the Presidential order undercuts that regime.

That matters in our region because Pacific seabed mining happens in two different legal places, and the difference decides who has a say. Patrick Kaiku of the University of Papua New Guinea set out the two layers in a Griffith Asia Institute paper in 2022. In the Area (the seabed beyond national jurisdiction), a state may only sponsor a contractor, and the International Seabed Authority licenses and regulates the work; Nauru, Tonga, Kiribati and the Cook Islands have all exploration in the Clarion-Clipperton Zone east of Kiribati, and Nauru’s sponsorship of a subsidiary of The Metals Company is what forced the Authority’s rule-making to a deadline. Inside its own exclusive economic zone, by contrast, a state exploits the seabed under its own laws, and the Authority has no role at all. The Cook Islands Seabed Minerals Authority estimates that some 12 billion tonnes of nodules lie across 750,000 square kilometres of the Cook Islands’ own continental shelf, and Rarotonga has issued exploration licences there. As its representative to the Pacific Islands Forum, Tepaeru Herrmann, put it, “the need to work collectively across the region does not trump the sovereign interests of each Forum member state to address its own development needs”.

The geopolitics has followed the geology. The Cook Islands signed a seabed minerals cooperation deal with China in 2025, New Zealand paused its aid in response, and in August that year Rarotonga agreed to work with the United States on the same resources. On 20 October Washington finalised the first deep sea mining lease sale in American history, to be held on 19 November, for 31 million acres of federal seabed off American Samoa . American Samoa banned seabed mining in its own territorial waters in 2024; the lease areas lie beyond them, inside the United States’ own zone, where neither the territory nor the International Seabed Authority can stop it. Meanwhile the Authority failed in March to finalise its mining code, and Connor Graham of the Lowy Institute warned in August that commercial mining “could begin as early as 2027”, “still with no agreed standards, liability framework, or guaranteed benefit-sharing protections”.

For Australia the implications are uncomfortable. The decisions that will shape the seabed of our near region are being made capital by capital, under domestic law, in agreements often drafted by the companies themselves, and Kaiku warns that the split between Pacific states that want to mine and those that want a moratorium could empty the Blue Pacific of meaning as a regional idea. The same seabed carries the cables this post is , and the survey vessels that map nodule fields can map cable routes just as well. Yet Australia, as Graham puts it, “is yet to take a clear policy position on deep-sea mining in the Pacific”, and Debomita Dasgupta noted in The Interpreter in June that, as of April, we were not among the forty countries that had declared their opposition, while our critical minerals strategy says nothing the seabed at all. A country that wants to be the Pacific’s partner of choice cannot be silent on the question its partners are most divided .

What has changed since Elisabeth finished writing

In the increasingly complexity and compounding environment of geopolitics, grey zone conflict and geoeconomic peril, events have continued to add prescience to Elisabeth’s ideas. On 10 September, Reuters reported that British, Norwegian and American forces had confronted Russian deep sea submersibles earlier this year near Svalbard, where vessels of Russia’s undersea warfare directorate, known by its Russian acronym GUGI, were rehearsing the deployment of a device designed to disable undersea cables in a conflict with NATO. The Norwegian Defence Minister, Tore Sandvik, said the operation sent Moscow “a clear message … that they cannot operate covertly”.

Twelve days later, the US-China Economic and Security Review Commission published a contracted report by the China Strategic Risks Institute, China’s Gray Zone Threats to Trans-Pacific Undersea Cables . Its authors, Andrew Yeh, Zack Liao, Owen Au and Athena Tong, document what the People’s Liberation Army calls the pursuit of ‘command of the deep sea’, a sixth operational dimension prized for “concealment, endurance, and high barriers to attribution”. PLA affiliated scholars cited in the report write that in a war, “cutting undersea cables is the most effective way to strike at the opponent’s vital points”. The report describes 10,000 metre class submersibles, swarming underwater gliders and work on ‘hibernating’ cutters that could be laid beside a cable in peacetime and woken in war. It also built an index of the most exposed trans-Pacific cables. Three of the ten most vulnerable have no landing in Taiwan at all, and one of them is the Southern Cross Cable Network, which the authors describe as “a major route linking the United States with Australia”.

Fifteen cables and a tanker off Perth

Braw writes from the North Atlantic and the Baltic. From the Australian perspective, our own defence minister has said a great deal that resonates with Elisabeth’s book. Addressing the Shangri-La Dialogue in Singapore on 30 May, Richard Marles told his audience that “around ninety-nine percent of Australia’s internet traffic flows through just fifteen subsea cables”, that these were “fifteen physical assets on the ocean floor”, and that “we are among the most exposed nations in the world to this threat”. He also said, with a candour I would like to hear more often, that “we have been slow, collectively slow, to recognise them as the strategic targets they have become”. The Chief of Navy, Vice Admiral Mark Hammond, had already said in November last year that the loss of either our sea lanes or our seabed cables “would be an existential threat to our island and to our people”.

Ten weeks after the Singapore speech, the argument stopped being abstract. Over a weekend in August, two fibre optic cables operated by Subco broke back to back inside the federal cable protection zone off Perth, in benign weather, while a UAE managed product tanker steamed an unusual racetrack pattern across the same water. Bevan Slattery, Subco’s founder, was measured it: “This could be a coincidence or an external aggression event.” The Australian Federal Police confirmed it had received a report of crime and was assessing it. Nothing has been proven, and I do not assume it will be. What the episode showed is how the Baltic pattern looks when it arrives here: a ship, a protection zone, a break, and no way of knowing for weeks whether it was carelessness or a message.

The Australian scholar who has thought hardest what this means is Samuel Bashfield, a research fellow at the La Trobe Centre for Global Security, whose Submarine Cables , a short study published by Cambridge University Press in its Elements in Indo-Pacific Security series in July, is the natural companion to Braw. Braw reports what has happened; Bashfield sets out why states are behaving as they are. He argues that states are competing for central positions in the cable network in three ways at once: controlling the hubs where cables land and connect (nodes); securing, and sometimes exploiting, the data that passes through them (flows); and dominating the manufacture, laying and repair of the cables themselves (production). His framing borrows Henry Farrell and Abraham Newman’s idea of weaponised interdependence, and his warning follows from it. A country that seeks a central position to reduce its own vulnerability also acquires the means to weaponise the network, whether it intends to or not. His closing sentence is, to my mind, the sharpest piece of advice: China, the United States and their partners “should operate deftly so as to not weaponise the very networks these states are attempting to secure”.

Bashfield is generous to Australia on one count. Our cable protection zones, two off Sydney and one off Perth, which criminalise anchoring, trawling and other activities that can foul a cable, are regarded internationally as a “gold standard” that others could copy. He is less generous on the rest. The zones were designed to deter careless seafarers, not hostile states. The 2024 National Defence Strategy did not engage with what he calls “seabed lines of communication”, and he notes that Indo-Pacific governments have been slower than European ones to say publicly how they would defend their cables. The 2026 National Defence Strategy does better, naming critical seabed infrastructure for the first time as vulnerable to sabotage by hostile nations , and underwater warfare now sits at the top of the capability investment priorities. But the report Bashfield co-wrote in August with David Brewster, Jocelinn Kang, Mike Constable and Anthony Bergin for the ANU National Security College, Connected & Protected , still finds cable security barely mentioned in that strategy, overlapping agency roles with no formal coordination, and no national cable strategy at all.

The critical infrastructure law does not fill that gap. The international cables are part of carriers’ telecommunications networks, and since 4 April 2025, when the old telecommunications security obligations moved out of the Telecommunications Act and into the Security of Critical Infrastructure Act 2018 , their owners have answered to the Cyber and Infrastructure Security Centre in the Department of Affairs. A responsible entity must protect its asset “from all-hazards, including security risks”, keep “competent supervision and effective control” over it, and tell the Secretary of Affairs changes to its network. For the most important assets the Telecommunications Security and Risk Management Program Rules 2025 require a written program covering “cyber and information, personnel, supply chain, and physical security and natural hazards”, carriers must report serious cyber incidents to the Australian Signals Directorate within twelve hours , and the Minister can direct an entity to do “a specified thing that is reasonably necessary to reduce or eliminate a security risk”.

Those are sensible obligations, and they are obligations on owners to manage the risks to their own assets. Nothing in them requires anyone to restore a severed cable within a set time, to hold spares or a repair ship on call, or to plan for several cables lost at once, and nothing in them reaches a vessel loitering outside the protection zone. The industry has noticed. In its December 2025 submission to the independent review of the Act, Vocus argued that “it is imperative that the Government treat subsea cables as critical infrastructure in the national interest”, and that the protection zones, while “a passive layer of defence”, are “insufficient on their own” .

Bashfield and Oscar Youlten add a complication that neither government nor industry has yet faced. In a piece for the Lowy Institute’s The Interpreter earlier this month, Fibre sensing: how undersea cables learned to listen , they explain how the glass fibre in a cable can now be used as a microphone. Distributed acoustic sensing reads the scatter of laser pulses to hear what moves near the cable; polarisation monitoring can register an anchor strike or a passing warship. In their words, “while distributed acoustic sensing can detect a trawler, it can also detect a submarine”. For cable owners this is a gift, a way to warn a careless master before he snags the line. For a strategist it changes the status of the cable. Infrastructure that listens becomes, in a crisis, a military sensor, and a military sensor is a legitimate target. Bashfield’s book makes the same point in its conclusion: by turning cables into sensing networks, “these cables may become more vulnerable in a conflict as the infrastructures take on a distinct military value”.

Braw is at her best on what happens after the cut. A cable is not mended by the navy. It is mended by a specialised ship, with a crew who can grapple a broken line from several kilometres down, haul it aboard in a seaway and splice glass fibres thinner than a hair on a pitching deck. There are not many such ships, most are owned by a handful of companies, and they are held on standby under maintenance agreements funded by the cable owners. The system works because faults are frequent and mostly predictable. The industry handles around 190 repairs a year, and the International Cable Protection Committee’s data attributes 84 per cent of them to fishing or anchoring.

For Oceania, the system comes down to one ship. Jocelinn Kang’s report for the Department of Foreign Affairs and Trade’s Cable Connectivity and Resilience Centre, The Role of Governments in Subsea Cable Repair , published in August, sets it out plainly. The vessel assigned to the South Pacific maintenance zone, which stretches from Indonesia to Hawaii, is the Lodbrog , based in Fiji, with its spares depot in Apia. At eleven knots it would take roughly eight days to reach Sydney or Port Moresby, and two weeks to reach Perth, before any allowance for weather. If the Lodbrog were busy, the nearest ship would come from Singapore, the Philippines or Taiwan and take ten to fourteen days to arrive. “No nation in Oceania”, Kang writes, “owns or operates a specialised cable repair vessel, or maintains a workforce capable of doing so.” A companion report for the same Centre by Mike Constable, Submarine Cable Repair and Maintenance Analysis in the Pacific Islands , concludes that “replacing the Lodbrog , SPMMA’s designated repair vessel and relocating the cable depot are near term necessities”. Bashfield put the Australian position bluntly in The Interpreter in June: “Australia currently has virtually no domestic cable repair capability. That gap will not be closed by underwater military drones.”

Distance is only the first delay. The cables that link Perth to Singapore cross Indonesian waters, and Indonesian cabotage rules require that repairs there be carried out by an Indonesian flagged ship and crew. Kang records that the permit process added six to seven weeks to one recent repair, because ships will not sail until the permit is in hand.

Kang is careful to separate two kinds of failure, and the distinction is the most useful idea I took from the Australian material. The first is slow repair under normal conditions: a single fault that takes longer to fix than a small island state can absorb. Most of that can be addressed with better regulation, faster permits and fewer faults. The second is the crisis case, where the question is “whether anyone will or can come at all”, because of hostilities, because a foreign government has told its operators not to serve a particular country, or because several cables have been cut at once and there are not enough ships to go round. The commercial system was built for the first kind, and nobody has yet built anything for the second.

The British have been working on the second kind for a year, and their experience is sobering. Their Joint Committee on the National Security Strategy, which reported in September 2025, heard from Alasdair Wilkie, who chairs the Atlantic Cable Maintenance and Repair Agreement, that “repair ships take three years to build, and the staff to man them 10 years to train ”. A ship can be bought within a budget cycle, but a crew cannot. The congressional commission report reaches the same place for the Pacific and asks Congress to co-finance at least five additional repair vessels for the Indo-Pacific by 2040, with priority access for the United States and its partners in a crisis. That is a sound proposal, but it also means that if the ships are built, Australia will be negotiating for a place in someone else’s queue.

There is good news in this story, and it comes from an unexpected direction. Braw, who is not given to flattery, wrote a column for the Financial Times on 1 October under the headline ‘Australia’s banks show how to prepare for cable blackouts’ . Earlier this year, National Australia Bank ran an exercise with other banks and government officials that assumed first the east coast cables and then the west coast cables were cut. Brad Carr, the bank’s executive responsible for geopolitical risk, told her what they found: switching to satellite moved only a tiny fraction of the data, which is to say the money, they needed to send. Anyone who read The Sky We Rent will not be surprised that satellites are not a substitute. What the banks then did is the interesting part. They began, with government, to design a payments system that could run mostly offline, so that Australians could still use debit cards and withdraw cash from ATMs if the country were cut off from international clearing. “It’s like field surgery, not a complete hospital,” one person involved told Braw. “There will be blood, but it will save the patient.”

The banks are not alone. The Australian Signals Directorate’s CI Fortify guidance, published in October last year, expects critical infrastructure operators to be able to “isolate vital OT and enabling systems for a period of 3 months while maintaining critical services”. Taken together, these are the beginnings of an answer to the second half of my question, what we would do while we waited. Andrew Horton asked the same question of digital systems in The Great Geo-Patriation , where he made severability , the ability to keep functioning once the foreign connections are cut, the test of digital sovereignty. Braw sets this against Vietnam, which had prepared far less and in August lost four of its eight major international cables and 30 per cent of its connectivity.

This complicates the story of neglect. On the policy architecture, neglect is the right word. On the practical question of keeping the country running through an outage, parts of the private sector and the cyber agency are ahead of the strategy documents, and a foreign observer has noticed. The gap is not that nobody has thought the problem; it is that the thinking sits with individual institutions and a single piece of agency guidance, rather than with anyone who owns the whole problem.

Some may argue that I am treating a commercial maintenance problem as a strategic one. Kang’s own report is clear that “the commercial system works under normal conditions”, that most faults are caused by fishing and anchoring, and that competing cable owners cooperate because each depends on the others for redundancy. That is true, and it is why the case for public money rests entirely on the second failure mode. A government that bought a repair ship to shave days off routine faults would be wasting its money. A government that wants assurance a ship will come when several cables have been cut and the commercial fleet is spoken for is buying something the market does not sell.

Others will say that redundancy solves the problem: fifteen or sixteen cables, landing on both coasts, with new routes across the Indian Ocean to Oman and Africa and a growing web across the Pacific. Darshana Baruah and Virpratap Vikram Singh of the International Institute for Strategic Studies show in their September mapping of Indo-Pacific cables that Australia is becoming a hub in its own right, with routes that deliberately avoid the South China Sea and new strategic nodes appearing at places like Christmas Island, which is expected to receive three new cables in 2028 and 2029 where one landed in 2018. That is real progress, and the Government deserves credit for the Pacific cable financing that sits behind some of it, which Bashfield shows has deliberately avoided Chinese suppliers in favour of American, Japanese and Australian firms. But the National Australia Bank exercise was run precisely because redundancy has a limit: several cables in the same corridor can be cut together, and the new nodes are, in Baruah and Singh’s words, assets that “an adversary may seek to disrupt”. More landing points do not by themselves guarantee resilience, as British lawmakers have also argued.

A third concern is that a sovereign ship would not even solve the Indonesian problem, since cabotage rules apply to any foreign flag, including ours. Kang makes this point herself. It is an argument for doing the diplomatic work alongside the shipbuilding, not instead of it.

What we would need when the call comes

Braw’s book leaves the policy questions at the threshold. Bashfield, Kang and their colleagues have walked through the door, and the Connected & Protected report sets out fifteen recommendations I will not repeat here. Five things stand out for me.

One, someone has to own the whole problem. The cables are privately owned, the protection zones are regulated by the Australian Communications and Media Authority, the seabed patrol belongs to Defence and the Australian Border Force, the cyber risk to the Australian Signals Directorate, and the Pacific funding to the Department of Foreign Affairs and Trade. The National Security College recommends Maritime Border Command as lead coordinator for physical threats and an interagency committee reporting to the Secretaries Committee on National Security. The European Union has already mapped, risk assessed and stress tested the cables that serve its members. Whatever the machinery, a national cable resilience strategy with a named owner is overdue.

Two, Australia should buy assured repair, not merely wait to be offered it. That means a repair ship based in Australia, co-financed with Japan, New Zealand and the United States and contracted to a trusted operator with government activation rights, together with a funded pipeline of trained crew that takes the decade Wilkie described seriously. Subco is already asking for government co-funding of a vessel that could lay and repair cables, which suggests the industry would meet a government halfway. If the American proposal for five more Indo-Pacific vessels goes ahead, we should be a partner in it from the start rather than a supplicant at the end.

Three, the permit delays should be negotiated away before they matter. Kang points to Article 22 of the Australia–Singapore Digital Economy Agreement as a model for writing cable resilience into economic agreements; a standing arrangement with Indonesia for emergency cable repair would remove weeks from the worst case at little cost.

Four, the banks’ exercise should become the national exercise. The cut-off scenario National Australia Bank ran for the financial system should be run across energy, health, logistics and government services, and its results reported publicly, so that the Australian public knows what a cable outage would mean for them and what they would be asked to do. The CI Fortify standard is a good start for operators; households and small businesses have nothing equivalent.

Five, the rules for cable sensing should be written before a crisis writes them for us. Bashfield and Youlten are right that governments will not wait for consensus. New cables landing in Australia should carry monitoring, as the National Security College recommends, and Australia should lead a regional conversation who may use the data a listening cable collects, and for what.

The station on Direction Island

The Cocos Islands have a cable again. They were connected to the Oman Australia Cable in 2022, and Bashfield has written that “forgotten nodes like this” will be critical again in the age of data. The operators on Direction Island in 1914 did three things right. They saw the threat coming, they got a warning out before the line went dead, and there were enough cables that the raiders could not find them all. The rest depended on a cruiser that happened to be over the horizon. Elisabeth Braw has written the book that should make every government see the threat coming. Australia has, in places, started to plan for the hours after the line goes dead. What we have not yet done is arrange for anyone to be over the horizon when it does. A country with fifteen cables and one repair ship eight days away is relying on luck, and luck is not a policy.