On 18 February 1873, a dredge came up over the side of HMS Challenger about 2,825 metres below the surface, southwest of the Canaries, carrying dark ferromanganese material nobody had a use for. Three weeks later, on 7 March, at Station 16 and 4,453 metres, the ship’s naturalists pulled up the first object recognised as a manganese nodule and reported it in Nature that May. John Murray and Alphonse Renard finally described the deposits properly in Deep-Sea Deposits in 1891, with assays running above 31% iron oxide and 25% manganese oxide.
Then the ocean’s metals sat, geologically curious and commercially irrelevant, for the better part of a century.
Cold War capability, and a cover story
What eventually made the abyss reachable was not mineral demand. It was war. From late 1950 the US Navy funded the hydrophone network that became SOSUS, with its first full-size array off Eleuthera in January 1952 and roughly 3,500 staff across twenty facilities by the mid-1970s. Project Mohole’s CUSS I drilled 183 metres of hole in 3,600 metres of water in 1961 and invented dynamic positioning in the process. The Glomar Challenger then drilled 624 sites between 1968 and 1983, paying out pipe to 6,243 metres.
The commercial expectation, though, was seeded by two things. The first was John Mero’s The Mineral Resources of the Sea in 1965, whose estimate of 1.5 trillion tons of Pacific nodules was, as the US Office of Technology Assessment later put it drily, “widely quoted and popularized, thus engendering a common belief… that the deep seabed nodules were a virtually limitless untapped resource”. The second was a lie. When the CIA sent the Hughes Glomar Explorer to raise a Soviet submarine from 5,000 metres in July 1974, at a cost of around US$800 million, the cover story was manganese-nodule mining. A purpose-built ship costing more than US$350 million, apparently financed by a billionaire to harvest nodules, was the most persuasive market signal the industry never had.
Scripps had already mapped the Clarion and Clipperton fracture zones in 1954, and by the early 1970s Horn and colleagues had identified the ground between them, 120° to 155°W, as “unquestionably the best” nodule province on earth. The United States Geological Survey (USGS) put its prime area at roughly 2.5 million square kilometres and published the map series in March 1978.
The first boom, and why it died
Four consortia went at it seriously: Ocean Mining Associates, Ocean Management Inc, Ocean Minerals Company with Lockheed, and the Kennecott group. And they largely cracked the engineering. In summer 1978 OMI lifted more than 600 tonnes of nodules from around 5,250 metres from SEDCO 445 — its own slide deck ended “Proved Technical Feasibility of Ocean Mining!”. That October, OMA raised 550 tonnes in eighteen hours at rates up to 50 tonnes an hour before a pump blade failed, and in February 1979 OMCO ran a successful test from the Glomar Explorer at 4,878 metres.
It died on arithmetic, not on engineering, and not on environmental regulation. The grades were what they are now — OTA’s 1987 figures were 1.22% nickel, 0.99% copper, 0.23% cobalt and 28.8% manganese, essentially the same as the ISA’s modern Clarion-Clipperton averages of 1.25% nickel, 1.08% copper, 0.22% cobalt and 27.1% manganese. What has changed is everything they were competing against.
In 1970 the average copper ore mined on land ran about 1.3% copper; today the global average is nearer 0.62%. In the 1980s the world was mining nickel at average grades near 1% and Congolese copper-cobalt at Kamoto ran 3.50% copper with 0.53% cobalt. And the market was a fraction of today’s. World copper production in 1975 was 7.25 million tonnes against roughly 23 million tonnes of mined copper in 2025; nickel 752,000 tonnes against 3.90 million; cobalt 50,300 tonnes against 310,000. Land was richer, closer and adequate.
Then prices collapsed. Nickel fell from US$2.96 a pound in 1980 to US$1.76 in 1986; cobalt from US$32.83 in 1979 to US$8.56 in 1982; copper from 101 cents to 73 cents, with about a third of US mine production idled. NOAA modelled a three-million-tonne-a-year pioneer venture in 1982 and found it needed some US$1.5 billion of 1980 dollars — over US$6 billion today — for “an unsatisfactory low internal rate of return of approximately seven percent”. Worth noting, because the modern debate assumes otherwise: NOAA’s 1980 cost study found “no real economic penalty for doing deep ocean mining in an environmentally sound manner.”
America built a regime, then handed it over
Congress legislated anyway. The Deep Seabed Hard Mineral Resources Act became Public Law 96-283 on 28 June 1980, giving NOAA authority to issue ten-year exploration licences and commercial recovery permits. NOAA issued four exploration licences in 1984 — USA-1 through USA-4 — and never issued a single commercial recovery permit. Two were relinquished in the late 1990s; Lockheed Martin still holds USA-1 and USA-4, now running to 2 June 2027. Lockheed’s own stated reasons for dormancy, on the record with NOAA, were “conditions in the metals markets” and “the lack of international recognition of the DSHMRA licenses”.
That second reason was the international waters problem. Arvid Pardo’s 1967 speech at the UN and General Assembly Resolution 2749 of 17 December 1970 had made the seabed beyond national jurisdiction the “common heritage of mankind.” The Third UN Conference on the Law of the Sea ran from December 1973, adopted the Convention on 30 April 1982 by 130 votes to 4, and UNCLOS entered into force on 16 November 1994 — creating, that same day, the International Seabed Authority based in Kingston, Jamaica. The ISA administers roughly 54% of the world’s oceans, with an Assembly of all members, a 36-member Council, and a duty under Article 145 to write the environmental rules for exploitation.
The United States did not sign. Reagan’s statements of 29 January and 9 July 1982 set out the objections: mandatory transfer of private technology, production ceilings, a subsidised Enterprise competing with commercial operators, revenue sharing that could reach national liberation movements, a Review Conference amendment procedure that could bind the US without Senate consent, and voting arrangements giving American firms no assured access and the United States no permanent seat on the Council. He noted that the states declining to vote for it “produce more than 60 percent of the world’s gross national product,” and in March 1983 declared that “deep seabed mining remains a lawful exercise of the freedom of the high seas”.
The 1994 Implementation Agreement was written largely to fix exactly those defects. It disapplied mandatory technology transfer and the production limits, stripped the Enterprise back, grandfathered pioneer investors, and made consensus the general rule in the Council with four chambers holding effective vetoes. The US signed it on 29 July 1994 and sent it to the Senate that October as a treaty document. The Senate Foreign Relations Committee voted it out 19–0 in February 2004 and 17–4 in October 2007, held three hearings in 2012 — and it has never again reached the floor.
The scepticism was prescient
Notice what the 1994 fix did not touch: the decision rules. It made them tighter. And that is precisely where the thing has ground to a halt.
Since the first six exploration contracts were signed in 2001, the ISA has issued a total of 31 contracts in force as of 31 January 2026 — 19 for polymetallic nodules, 8 for sulphides, 4 for crusts — held by 21 contractors under 20 sponsoring states, covering 1.445 million square kilometres, or 0.7% of the Area. Exploitation contracts issued in that time: zero. Private companies cannot apply directly. Under UNCLOS Article 153(2)(b), a non-state entity (a company) may explore or exploit in the Area only if it has the nationality of a State Party and is sponsored by that State. The State issues a Certificate of Sponsorship that accompanies the application.
The Mining Code has been in drafting since 2014. A draft went to the Council in July 2019 and has gone back to council and failed at almost every meeting since. A road map adopted in 2023 targeted adoption in 2025 and missed it. Second reading finished in July 2025. In March 2026 the Council worked through 29 of 32 outstanding issues, and in July 2026 it again failed to adopt, deferring to March 2027 while extending Nauru Ocean Resources’ exploration contract by five years.
This is not incompetence and it is not ignorance. The ISA has a serviceable technical framework and, after a quarter-century of contractor baseline work, a serious evidence base: 255 contractor research cruises since 2001, and a DeepData repository exceeding 10 terabytes holding some 89,000 occurrence records. One correction worth making, because the number circulates: the 200,000 figure often cited is DeepData’s count of species observations, not peer-reviewed papers. There were roughly 1,935 deep-sea-mining publications in the whole literature between 1968 and 2021, and there have been many published in the last 5 years as well. The point stands regardless: the data exists and it is public.
What blocks adoption is the consensus architecture Reagan’s negotiators warned about. Article 161(8) defines consensus as the “absence of any formal objection,” and the 1994 chambers hand small groups a veto. Around that structure has grown a permanent campaign: 46 states (most of which have zero vested interests in deep sea mining) now back a moratorium or pause, the Deep Sea Conservation Coalition supporting a global moratorium claims over 140 member organizations, 64 companies and financial institutions have signed its business statement, and the ISA has accredited 129 observers of which 67 are anti deep sea mining focused NGOs. The ISA itself has had to run a public fact-check page, and in February 2024 its then Secretary-General wrote to the Convention on Migratory Species calling a COP document “seriously flawed,” with “numerous factual (and legal) errors,” noting that “almost half of the references cited are to materials published by anti-deep-seabed mining campaign groups”. The dark-oxygen affair followed the same pattern: a striking Nature Geoscience claim, a detailed industry rebuttal alleging an unreported nodule-free control in which oxygen also rose, and independent scientists publicly unconvinced.
Nauru triggered the two-year rule on 25 June 2021. The “two-year rule” is a 1994 UNCLOS implementing-agreement provision to force a deadline on ISA exploitation regulations. The clock ran out on 9 July 2023 with the rules unfinished. Three years on, there is still no mining code.
Re-enter the United States
Meanwhile the strategic case inverted. Cobalt is 73% Congolese; China refines roughly 48% of the world’s copper and holds a top-refiner share of about 72% across critical minerals, with the IEA warning of a 25% copper deficit by 2035. The 2025 US critical minerals list runs to 60 entries and added copper, nickel manganese and cobalt were already there. And the resource is not marginal: USGS puts the Clarion-Clipperton Zone at 21.1 billion dry tons of nodules, with ISA modelling implying on the order of 6,000–8,700 million tonnes of manganese, 270–393 million tonnes of nickel and 46–68 million tonnes of cobalt — multiples of terrestrial reserves for all three.
The technology also stopped being hypothetical. Allseas’ Hidden Gem lifted more than 3,000 tonnes of nodules from beyond 4 kilometres in 2022 through a 4.3-kilometre riser, and in May 2026 signed a production contract for 3.0 million wet tonnes a year with commissioning targeted for Q4 2027. China ran its Kaituo 2 collector at about 4,100 metres in 2024 and put a Clarion-Clipperton collector test at 5,000–5,500 metres before the ISA’s technical commission; Japan recovered rare-earth mud from near 6,000 metres off Minamitorishima in early 2026.
So the United States did the obvious thing: it woke up the statute it never repealed. Executive Order 14285, “Unleashing America’s Offshore Critical Minerals and Resources,” signed 24 April 2025, directed NOAA to expedite DSHMRA licences and permits within 60 days. NOAA proposed its rule on 7 July 2025 and published the final rule on 21 January 2026, creating a consolidated licence-and-permit pathway.
The Metals Company who had triggered the 2 year rule with Nauru jumped first, and for a rational reason: it had spent hundreds of millions inside a system that could not tell it when, or whether, it would ever be allowed to mine. TMC USA filed two exploration applications and a commercial recovery permit application on 29 April 2025, then a consolidated application over roughly 65,000 square kilometres in January 2026; NOAA certified a second licence area of about 122,000 square kilometres in May 2026, and the consolidated application went to public comment on 19 August 2026, with a hearing set for 13 October. TMC guides to a permit before the end of Q1 2027.
It is no longer alone. NOAA reported more than ten applications by late January 2026, with exploration applications publicly noticed from TMC USA, American Metal Resources, SeaX and American Deep Sea Minerals; Deep Sea Rare Minerals filed in August 2025; American Ocean Minerals reached substantial compliance on a 147,054-square-kilometre area in July 2026 and announced a US$1 billion merger with Odyssey Marine in April. Separately, BOEM has scheduled a lease sale for two blocks off American Samoa on 19 November 2026. Within 200 mile national limits countries like the Cook Islands, Japan and America, are able to exploit their own resources under national rules.
The ISA’s choice
The Authority’s response has been legal rather than regulatory. Secretary-General Leticia Reis de Carvalho declared on 30 April 2025 that “no State has the right to unilaterally exploit the mineral resources of the Area outside the legal framework established by UNCLOS,” and that “this prohibition is binding on all States, including those that have not ratified UNCLOS”. In July 2025 the Council invoked Article 139 and demanded information from contractors at risk of non-compliance. NORI and TOML took the ISA to the Seabed Disputes Chamber in May 2026 and obtained provisional measures on 18 July; the Secretary-General has now put three questions on Article 137 and non-recognition to ITLOS.
The strategy is to make the metal untradeable — as one legal commentary puts it, “a robust non-recognition obligation, applied to mineral supply chains, could render mining under” the US statute “commercially unviable”. Which, from Washington’s point of view, is not a threat at all. If treaty parties are legally discouraged from refining American nodules, the refining happens in the United States. That was the industrial policy objective in the first place and exactly what America wants.
Forty-four years after Reagan refused to sign, the specific defect he identified, a governance structure in which any determined minority can prevent a decision indefinitely, has produced exactly the outcome he predicted. The ISA holds 1.445 million square kilometres of the world’s largest untapped supply of nickel, cobalt and manganese, twelve years of drafting, a quarter-century of environmental data, and no way to say yes. Its contractors have begun to notice that somebody else can.
The International Seabed Authority still has the better claim to legitimacy, and a code adopted in March 2027 would restore most of its authority overnight. But that window is narrowing, and it is closing not because the science is unsettled or the engineering unproven, but because a body designed to require unanimity was handed a decision that will never command it. Read the tea leaves, or become a registry of permissions nobody waits for. I predict that if the ISA fails to adopt legislation in 2027, companies that have spent hundreds of millions developing deep sea mining licenses will follow TMC out the door because their shareholders will give them no choice.
This is an excellent piece. What strikes me most is that the deeper lesson may be less about whether Reagan was “right” in every respect, and more about institutional design.
From years of working at the heart of military and state institutions, I have seen the same principle repeatedly: legitimacy and procedure matter, but an institution that cannot convert them into timely decisions eventually begins to lose practical authority.
Once actors conclude that the formal system cannot produce a predictable decision, they do not necessarily abandon the objective — they begin building alternative pathways around the system. That is where institutional delay stops being administrative and becomes strategic.
The ISA may still possess the stronger claim to international legitimacy, but legitimacy without decision capacity can gradually become a liability rather than a source of authority.
A system does not lose relevance only when its rules are rejected. Sometimes it loses relevance when everyone keeps respecting the rules — but stops waiting for the decision.
Fascinating, and important. Thank you!