Between Hawaii and Mexico, on the floor of the Pacific Ocean, there is a plain the size of the continental United States. It sits four thousand meters below the surface, in permanent darkness, at temperatures near freezing, under pressure six hundred times that of sea level. No sunlight has ever reached it. Nothing there photosynthesizes. The rain that feeds its ecosystem is not water but a slow, continuous fall of dead matter from above — fragments of plankton, fecal pellets, the shed skins of crustaceans, the bodies of everything that dies in the upper ocean. Marine biologists call it marine snow. It takes weeks to reach the bottom.
This is the Clarion-Clipperton Zone, named for two fracture systems that bound it to the north and south. It stretches roughly 7,200 kilometers east to west and covers 4.5 million square kilometers of abyssal plain. For most of human history, nobody knew it was there. For most of the twentieth century, the few people who knew about it considered it barren — a flat, featureless mud desert with nothing on it worth noticing.
They were wrong about that.
Scattered across the floor of the Clarion-Clipperton Zone, half-buried in sediment or sitting on the surface like dark potatoes left in a field after harvest, are trillions of polymetallic nodules. They range from the size of a golf ball to the size of a grapefruit. They are black or dark brown, rough-textured, and surprisingly light for their size. Pick one up — if you could — and it would feel like a dense, misshapen billiard ball.
Inside, a nodule is built in concentric layers, like a tree ring or an onion, each layer a thin shell of manganese and iron oxide precipitated from seawater over deep time. At the center of every nodule is a nucleus — something hard enough to seed the crystallization. Often this is a fragment of rock or a piece of clay. Sometimes it is a piece of bone. Sometimes it is a shark's tooth.
In August 2026, NOAA's EX2605 expedition recovered seven fossilized megalodon teeth from the abyssal plain near the Cook Islands. The teeth were not loose on the seafloor. They were found inside polymetallic nodules, each one the hidden core around which millions of years of mineral precipitation had slowly accumulated. A tooth from a shark that went extinct roughly 3.6 million years ago, sitting on the ocean floor, gathering manganese atom by atom from the water above it, until it became something unrecognizable — a dark, rough lump of metal oxide with a fossil at its heart.
Polymetallic nodules grow at a rate of roughly two to five millimeters per million years. A nodule the size of a potato takes ten to twenty million years to form. This is not a human timescale. It is barely a geological one. The nodules currently sitting on the floor of the Clarion-Clipperton Zone began forming somewhere between eight and thirty-two million years ago. Some of them predate the genus Homo by a factor of ten. They grew while the Andes rose, while the Isthmus of Panama closed, while the Mediterranean dried out and refilled, while the ice ages came and went. They are still growing now, though at a pace that makes glaciers look hurried.
What makes them economically interesting is their composition. By weight, a typical nodule from the CCZ is 24 to 30 percent manganese. It also contains nickel (1 to 1.5 percent), copper (0.8 to 1.5 percent), and cobalt (0.2 to 0.5 percent). These numbers sound small until you consider that the CCZ contains an estimated 21 billion tons of nodules. The zone holds more cobalt and manganese than all known land deposits on Earth combined. The nickel alone could supply the global battery industry for decades. These are precisely the metals required for lithium-ion batteries, the batteries that power electric vehicles, the vehicles that are supposed to replace the internal combustion engine and slow climate change.
The nodules are just sitting there, on the surface, not buried under thousands of feet of overburden like a copper deposit in Chile. You do not need to blast tunnels or strip mountaintops. You send a machine to the bottom, and it picks them up.
Before anyone seriously considered mining the Clarion-Clipperton Zone, the CIA accidentally made it famous.
In 1968, a Soviet Golf II-class ballistic missile submarine, K-129, sank in the North Pacific, roughly 2,500 kilometers northwest of Hawaii, in water nearly five kilometers deep. The submarine carried three nuclear-armed ballistic missiles, nuclear torpedoes, and Soviet codebooks. The U.S. Navy detected the sinking through its underwater acoustic surveillance system. The Soviets searched for the wreck and did not find it. The Americans found it almost immediately.
Recovering a submarine from 4,900 meters of ocean required technology that did not exist. More problematically, it required secrecy — the Soviet Union could not learn that the United States was salvaging their submarine from international waters. The CIA needed a cover story that would explain why a massive, purpose-built ship was lowering heavy equipment to the deep ocean floor in the middle of the Pacific.
They chose deep-sea mining.
Working through Howard Hughes — the reclusive billionaire whose eccentricities made him the perfect front, since nobody expected his business ventures to make sense — the CIA constructed a 618-foot ship called the Hughes Glomar Explorer. The public story was that Hughes had become interested in the commercial potential of manganese nodules on the deep seafloor. The ship was equipped with a massive claw mechanism hidden in its hull, ostensibly designed for nodule collection.
In the summer of 1974, the Glomar Explorer sailed to the wreck site and lowered its claw. It grasped the submarine's forward section and began to lift. Partway up, a mechanical failure broke the grapple, and roughly two-thirds of the recovered section fell back to the ocean floor. What reached the surface included the bodies of six Soviet sailors (who were given a formal burial at sea, in Russian, filmed by the CIA) and portions of the hull. Whether it also included codebooks or nuclear warheads remains classified.
The operation — codenamed Project Azorian — was eventually exposed by journalist Seymour Hersh and the Los Angeles Times in 1975. But by then, the cover story had done something unexpected. It had convinced people that deep-sea mining was real, that it was imminent, and that there was money in it. The manganese nodule story, invented as a lie, generated genuine commercial interest. Mining companies began sending their own survey expeditions to the CCZ. The fiction preceded the industry.
After the intelligence mission was complete, the CIA leased the Glomar Explorer for actual mining tests. In 1979, the Ocean Minerals Company (OMCO) used it to conduct one of the first collector trials in the Clarion-Clipperton Zone — dragging a 100-ton machine across the seafloor to scoop up nodules as a proof of concept.
In March 2023, a team of scientists returned to the OMCO test site to see what had happened in the forty-four years since.
The tracks were still there. They looked, according to the researchers, "like they were made yesterday."
The persistence of those tracks is not surprising once you understand the environment. The abyssal plain of the CCZ receives new sediment at a rate of roughly one centimeter per thousand years. A five-centimeter-deep gouge made in 1979 will not be filled in for another four thousand years. The deep ocean operates on a timescale that makes industrial activity look like a bomb going off.
The 2025 study by the Natural History Museum found that the biological impact of the 1979 test was mixed, and in some ways more troubling than a simple finding of total destruction would have been. Some organisms had returned. Sediment-dwelling infauna — small worms, nematodes, creatures that burrow — had recolonized the disturbed tracks to densities comparable to undisturbed areas. Life was finding its way back into the mud.
But the organisms that depended on nodules had not recovered. The nodules themselves were gone — scooped up by the collector in 1979 — and anything that required a hard substrate to attach to had no substrate to attach to. Four decades later, the mined tracks remained biologically impoverished in precisely the category of life that was most distinctive to the region. The mud-dwellers came back. The nodule-dwellers did not, because there were no nodules to dwell on, and there would not be for millions of years.
This matters because the most remarkable organisms in the Clarion-Clipperton Zone are not the ones in the sediment. They are the ones on the nodules.
Xenophyophores are single-celled organisms. A single xenophyophore can grow to the size of a basketball.
This fact is worth pausing on. A single cell — one continuous membrane enclosing one continuous cytoplasm, with thousands of nuclei distributed through it like stars in a galaxy — large enough to hold in both hands. They are among the largest cells on Earth. They live on the abyssal plain, often attached to polymetallic nodules or sitting on the sediment surface near them, and they build structures.
A xenophyophore extends long filaments outward from its body, probing the surrounding sediment like fingers. It selects individual particles — grains of sand, fragments of shell, microscopic bits of debris — and rejects others that are too large or too small. From the chosen particles, it constructs a test: a protective casing, an architecture, cemented together with its own secretions. Under a CT scanner, the interior of a xenophyophore looks like a tree — the living cell branching through the structure it has built, reaching into every corner but leaving room to spare, an organism that is also a building and also a city.
Because each xenophyophore is also a habitat. Researchers have documented over fifteen major animal groups living inside individual xenophyophores — sponges, mollusks, crustaceans, worms. A single cell can house more than a hundred individual organisms. They have been described as "apartment houses for animals." In areas of dense xenophyophore populations — and some parts of the CCZ contain as many as twelve per square meter — the seafloor becomes a kind of reef built from single cells, each one a micro-ecosystem, each one anchored to a nodule that took ten million years to form.
When a nodule collector passes over the seafloor, it takes the nodule and everything on it.
Ninety percent of the species in the Clarion-Clipperton Zone have not been formally described. The taxonomic work is proceeding at roughly twenty-five new species per year. In March 2026, researchers published descriptions of twenty-four new amphipod species from the CCZ, including two new genera and an entirely new superfamily — Mirabestioidea — representing a previously unknown branch on the evolutionary tree. These are not minor reclassifications. They are genuinely new lineages, things that had no name and no known relatives, discovered in a region that is currently being surveyed for industrial extraction.
The phrase that recurs in the scientific literature is "mining before we know what's there." The conventional sequence — survey, catalog, understand, then decide — has been inverted. The International Seabed Authority has issued seventeen exploration contracts in the CCZ. The Metals Company, a Canadian mining firm, has an agreement with Allseas to build a commercial nodule collection system with an annual capacity of three million wet metric tons. NOAA is reviewing applications. In April 2025, President Trump signed an executive order directing federal agencies to expedite licensing for deep-sea mineral extraction. The commercial machinery is moving faster than the scientific machinery.
Against this: forty-five countries have called for a moratorium or precautionary pause. BMW, Google, Samsung, Volvo, and Volkswagen have pledged not to use metals sourced from seabed mining. The ISA has been trying to finalize a Mining Code — the regulations that would govern commercial extraction — for years. As of July 2026, the code remains unfinished. The negotiations ended in stalemate. No mining has been approved. No mining has been banned. The situation is frozen in the particular way that international institutions freeze: not through resolution but through the inability to reach one.
In July 2024, a paper appeared in Nature Geoscience that briefly changed the terms of the debate.
Andrew Sweetman, a marine ecologist at the Scottish Association for Marine Science, reported that sealed chambers placed on the seafloor of the Clarion-Clipperton Zone were registering rising oxygen levels. This should not have been possible. At four thousand meters, in total darkness, with no photosynthetic organisms present, oxygen should only be consumed — by microbes, by animals, by chemical reactions in the sediment. Every previous deep-sea study had shown exactly this pattern. Oxygen went down in sealed benthic chambers. That was the baseline expectation, confirmed across decades of research.
Sweetman's chambers showed the opposite. Over roughly two days, oxygen concentrations rose to more than three times the background level. The team attributed this to the polymetallic nodules inside the chambers and proposed a mechanism: seawater electrolysis. The metallic nodules, rich in manganese and iron, carried measurable voltage differentials on their surfaces — up to 0.95 volts. Sweetman hypothesized that these voltages were sufficient to split water molecules, producing oxygen and hydrogen without sunlight, without biology, without photosynthesis. He called it "dark oxygen."
The implications were significant. If rocks on the deep seafloor were producing oxygen through electrochemistry, it would mean that the deep ocean was not merely consuming oxygen supplied from the surface but generating its own. It could reshape models of ocean oxygenation, alter estimates of how much oxygen the deep sea contributes to (or subtracts from) the global budget, and — most immediately — add a powerful new argument against mining. You would not just be destroying ecosystems. You would be removing a previously unknown source of the planet's oxygen supply.
The paper was covered extensively: CNN, the BBC, Scientific American, Nature news. The phrase "dark oxygen" entered popular science vocabulary within days.
Then the problems started.
The first issue was the voltage. Sweetman reported a maximum reading of 0.95 volts on a nodule surface, which would indeed be close to the 1.23-volt threshold required for water electrolysis. But the average voltage measured across nodule surfaces was 0.024 volts — roughly fifty times too low to split water. The 0.95-volt reading was a clear outlier. Building a mechanism on a single outlying data point is, to put it gently, unusual.
The second issue was the controls. Several experiments conducted in 2018 had recorded rising oxygen inside chambers that contained no polymetallic nodules. Control chambers with only seawater showed the same pattern. If oxygen rose without nodules present, nodules could not be the cause. These results were not reported in the 2024 paper.
The third issue was the data itself. Of thirty-two benthic chambers deployed, only two produced potentially reliable oxygen readings. The rest showed patterns consistent with known experimental artifacts — contamination, sensor drift, leaks. Rising oxygen in sealed benthic chambers is a documented problem in deep-sea instrumentation, not a discovery.
The fourth issue was thermodynamics. In March 2025, Angel Cuesta and Marcel Jaspars, both at the University of Aberdeen, published an analysis titled "Is abyssal dark oxygen production even possible at all?" Their answer was no. Electrolysis requires an external energy source — something has to drive the splitting of water molecules. No such energy source had been identified in Sweetman's system. The nodules were not batteries. They were not connected to anything. The proposed mechanism, Cuesta and Jaspars argued, violated the first and second laws of thermodynamics. You cannot get chemical work from a system at equilibrium. The manganese and iron in the nodules are already in their lowest-energy oxide forms. There is no thermodynamic gradient to exploit.
Five independent rebuttal papers were submitted for review. Downes, Tengberg, Nakamura, and others, working separately, converged on the same conclusions: the oxygen data were artifactual, the mechanism was physically impossible, and the key results were unreproducible.
By late 2025, Sweetman had publicly retreated from core claims. He acknowledged that some chambers showing oxygen increases had not contained nodules. At the 2025 AGU meeting, his abstract abandoned the electrolysis mechanism entirely, stating that the process "remains unclear." In April 2026, Nature Geoscience attached an Editor's Note to the paper, alerting readers to concerns under investigation. As of August 2026, formal retraction has been called for but not enacted.
The Nippon Foundation had already awarded Sweetman £2 million to investigate dark oxygen further. Three research expeditions were launched in 2026, the first in January. Whether these expeditions will find something genuinely interesting about deep-sea oxygen chemistry or will simply confirm the rebuttal papers is an open question. The most likely outcome is the least dramatic one: the oxygen readings were instrument error, the mechanism was impossible, and "dark oxygen" was a mirage that briefly illuminated how badly people wanted a simple reason not to mine the ocean floor.
The argument against mining the Clarion-Clipperton Zone does not require dark oxygen. It does not require a novel geochemical process or a rewriting of thermodynamics. It requires only the facts that are already established.
The nodules take ten to twenty million years to form. The organisms that live on them — xenophyophores, sponges, corals, anemones — cannot recolonize once the nodules are removed, because there is nothing to recolonize on. The tracks from a 1979 test are still visible and biologically impoverished forty-four years later. Ninety percent of the species in the zone have not been named. Twenty-four of them turned out to belong to a previously unknown superfamily. The ecosystem is not well understood. The recovery timescale is not decades or centuries but geological epochs.
The argument for mining is also straightforward. The energy transition requires cobalt, nickel, manganese, and copper. Land-based mining for these metals involves strip-mining in the Congo, deforestation in Indonesia, child labor in artisanal cobalt operations, acid mine drainage, and carbon emissions from heavy machinery. The nodules contain all four metals in a single formation, sitting on the surface, and collecting them produces no tailings, no acid drainage, and — according to the mining companies — a smaller carbon footprint per ton of metal than any terrestrial alternative. The batteries that will replace fossil fuels need to be made of something.
This is the structure of the problem. Both sides are correct about the facts they cite. The disagreement is about what we are willing to lose. The question is not whether mining the CCZ would destroy an ecosystem — it would, in ways that are effectively permanent on any human timescale. The question is whether that destruction is justified by what the metals would be used for, and whether there are alternatives that make the destruction unnecessary.
Lithium iron phosphate batteries, which contain no cobalt, nickel, or manganese, now account for a growing share of the EV market. Redwood Materials projects it could supply raw materials for five million batteries per year through recycling alone by 2030. Sodium-ion batteries are entering commercial production. The metals in the nodules are critical today. Whether they will still be critical in 2035 is less certain. Whether they will be critical in 2050 — the timescale on which climate targets are set — is genuinely unknown.
Meanwhile, the nodules sit in the dark, growing at five millimeters per million years, encasing the teeth of sharks that went extinct before humans existed, serving as the foundation for single-celled organisms the size of basketballs that house entire communities of animals inside their bodies, in a place so remote and so hostile that most of the life there has never been given a name.
The International Seabed Authority meets again in 2027. The Mining Code remains unfinished. The moratorium coalition has forty-five members and growing. The Metals Company has no revenue and a plan to extract three million tons per year. The tracks from 1979 still look like they were made yesterday.
Nothing has been decided.