A robot arm descended toward the seafloor somewhere off the coast, in water so dark it could swallow a skyscraper whole. It stopped. For life, its camera scanned. When it saw none on the target rock, it reached down, picked up a manganese and nickel-encrusted nodule the size of a potato, and tugged. A tiny cloud of silt blossomed and vanished. The machine continued. According to Impossible Metals, a startup that has centered its entire pitch around the notion that the ocean floor can be harvested without hollowing it out, this is what “responsible” deep-sea mining looks like.
The CEO and co-founder, Oliver Gunasekara, is certain of that assertion. His company’s autonomous robot, which was tested in a harbor in November of last year, is 95% accurate at identifying lifeforms that are one millimeter or larger and is designed to steer clear of them. The arms are quick, accurate, and minimally disruptive, just like warehouse picking machines. When you consider what’s really being suggested—industrial extraction from one of the least understood ecosystems on Earth, scaled up enough to matter economically—it’s the kind of technology that sounds almost comforting. Thousands of square kilometers of open-ocean seabed are far from a single test in a harbor.
| Category | Details |
|---|---|
| Topic | Deep-Sea Mining & Green Technology |
| Key Company | The Metals Company (TMC) |
| Key Startup | Impossible Metals |
| CEO / Co-Founder | Oliver Gunasekara (Impossible Metals) |
| Robot Accuracy | 95% lifeform detection accuracy (≥1mm organisms) |
| Target Materials | Nickel, Manganese, Zinc, Molybdenum, Rare Earth Elements |
| Nodule Type | Polymetallic nodules (potato-sized, seabed-scattered) |
| Regulatory Body | International Seabed Authority (ISA) |
| US Involvement | Unilateral seabed exploration announced March 2025 with TMC |
| Scientific Concern | Ecological damage to undiscovered species and seabed substrate |
| Key Academic Finding | Deep-sea mining described as “multi-billion-dollar solution to problems that do not exist” |
| Publication | PMC / NIH (Alger et al., 2025) |
However, like most pitches from Silicon Valley, this one is compelling. The argument goes something like this: EV batteries and renewable energy infrastructure require cobalt, nickel, and manganese. These metals are mined on land, which is hazardous, politically sensitive, and frequently linked to grave violations of human rights. In contrast, the ocean floor is dotted with nodules like loose change; there are no workers in hazardous shafts, no villages to relocate, and no rainforests above. It’s more tidy. It’s more secure. It is sustainable. Or so the tale goes.
After closely examining those claims, researchers at the NIH and a number of universities concluded that they were inadequate. According to a 2025 PMC article, deep-sea mining is “a multi-billion-dollar solution to problems that do not exist.” They contend that the mineral scarcity argument is mostly fabricated. The metals in question are not even close to running out in terrestrial deposits. Although supply chain issues exist, they can be addressed with current mining, recycling, and circular economy initiatives. If deep-sea mining merely increases global supply instead of displacing damaging land operations, the social benefit argument falls apart. Additionally, there isn’t yet a profitable example of the economic case for international seabed extraction.
Whether or not the startup community has fully addressed these criticisms is still up for debate. There is a perception that venture capital moves more quickly than science once it discovers a compelling narrative, such as the green transition, ocean abundance, or AI-assisted precision. There is a truly fascinating robot in Impossible Metals. The Metals Company’s potential resource value is in the billions. The deeper question of whether the seabed should be mined at all and whether the ecosystems there—many of which are still unidentified—can withstand industrial contact is not addressed by either of those facts.
The state of governance complicates matters. In order to avoid the International Seabed Authority and the multilateral framework that oversees the ocean floor, the US declared in March 2025 that it would start unilateral commercial seabed exploration in collaboration with The Metals Company. It was described by critics as a slow-motion land grab. It was described as pragmatic by supporters. The move indicates that some governments have quietly decided that the regulatory debate is taking too long and that waiting isn’t in their economic interest, though the reality is likely somewhere in the middle.
It’s difficult to ignore the fact that this pattern has previously occurred in industrial agriculture, offshore oil, and hydraulic fracturing. A new technology promises to have a small environmental impact. In controlled settings, preliminary tests proceed fairly smoothly. Policymakers and investors hear the pitch. Then extraction scales, conditions shift, and it becomes more difficult to quietly fulfill the initial commitments. Perhaps the most honest thing to say about deep-sea mining at this point is that it is genuinely impossible to predict whether it will follow that trajectory.
The robots do exist. The nodules are genuine. They contain genuinely valuable metals. The part where all of this occurs without causing irreversible harm to an environment that science hasn’t even completely mapped out is still up in the air. It’s not a minor uncertainty to ignore. Regardless of their motivations, the startups promoting this technology will eventually have to deal with it—not in a harbor test, but at depth, at scale, in the actual ocean.
