Why are humans exploring space but not the deep ocean?

More than 80% of the ocean remains unexplored not for lack of curiosity, but because the economics, physics and infrastructure of the deep sea are less forgiving than space.

The paradox is stark. Humanity has landed rovers on Mars, photographed a black hole, and sent spacecraft to the far side of the moon, yet more than 80% of the world’s oceans remain unexplored, according to NOAA and others cited across the ocean-science community. In practical engineering terms, the deep ocean is often the harsher frontier: dark, cold, corrosive, navigation-poor, and subject to pressures that make routine access extraordinarily difficult. We explore space because it is strategically compelling and institutionally organised. We lag in the ocean because the medium is less forgiving, and the supporting system is thinner.

The ocean’s difficulty is a matter of physics

The simplest answer to the question is also the most important. The ocean is hard to explore because water is a profoundly hostile environment for machines and humans alike. Oceana Canada, citing NASA oceanographer Dr. Gene Carl Feldman, puts the point with unusual bluntness: “In some ways, it’s a lot easier to send people into space than it is to send people to the bottom of the ocean.” The reasons are basic but severe: zero visibility at depth, extreme cold, and crushing pressure.

Pressure is the great antagonist. At depth, it dictates materials, hull geometry, endurance, sensor design, and cost. Vehicles built for deep-sea work need to survive conditions that NOAA notes can require wider-bodied designs able to withstand pressures thousands of metres below sea level. Spacecraft face radiation, vacuum and thermal extremes, but deep-sea systems must operate while physically compressed by the surrounding medium, often without the benefit of easy communication or navigation.

And unlike space, the ocean blocks many of the tools modern exploration relies on. GPS signals do not penetrate seawater. Radio communication is severely limited. Light disappears quickly. Even when a vehicle can reach the seafloor, positioning it precisely and returning data reliably are non-trivial tasks. The result is that exploration is not just a matter of building a capable craft; it is a matter of building an entire operational chain around a difficult environment.

The ocean is close. It is not accessible.

Why space advanced faster

Space exploration benefits from something ocean science has historically lacked: large, concentrated institutions with clear public legitimacy, durable budgets, and geopolitical urgency. NASA became a national project as much as a scientific one. Ocean exploration, by contrast, has often been distributed across agencies, research institutions, naval programmes and commercial operators. The result is fragmented investment.

That asymmetry matters. An essay in Issues in Science and Technology argues that ocean exploration is both understudied and highly promising, and suggests the United States lacks a dedicated agency capable of spearheading a major drive comparable in ambition to space programmes. Whether or not one agrees with the institutional prescription, the diagnosis is hard to dismiss. Space has had a coherent narrative and a powerful administrative machine. The ocean has had many missions, but fewer unifying structures.

Budgets follow stories. Space has long offered strategic prestige, defence relevance, and public spectacle. Ocean work, though no less consequential, is more procedural in the public imagination: mapping, monitoring, surveying, characterising. Yet those verbs underpin matters of climate, food security, ecosystems, offshore energy, marine resource management and national security, as the autonomous-ocean literature and recent NOAA-industry partnerships make clear.

This helps explain the paradox. We have not chosen stars over seas because the ocean matters less. We have chosen the domain for which the political and institutional machinery matured earlier.

Technical breakdown

If deep-ocean exploration is progressing, it is increasingly because of robotics rather than direct human presence. NOAA traces the rise of autonomous underwater vehicles, or AUVs, back to the 1960s, with deep-sea mapping specialisation accelerating by the 1980s. These platforms now come in multiple forms, each shaped by mission. Some are torpedo-like and thruster-powered for speed. Others are bulkier, designed to endure high pressure and carry specialised sensors.

Conventional AUVs are typically propeller-driven subsurface vehicles that provide stable platforms for high-resolution seafloor mapping and imaging, according to the Future Vision for Autonomous Ocean Observations paper. Their strengths are clear: they can survey remote areas, collect imagery, and monitor environmental change without placing humans at direct risk. They are increasingly used to observe ecosystems challenged by pollution, ocean acidification, warming and invasive species.

Their constraints are equally important. These systems often require a supporting surface vessel, partly for navigation and partly for recharging and data download. That requirement is one reason ocean exploration remains expensive and logistically dense even in the age of autonomy. A robot on its own is only part of the system.

Recent developments suggest a broader architecture is emerging. Fugro and NOAA Ocean Exploration signed a five-year Cooperative Research and Development Agreement to accelerate deep-ocean mapping and characterisation using uncrewed surface vessels (USVs), electric remotely operated vehicles (eROVs) and cloud-based data workflows. The significance is not only the hardware. It is the operational model: fewer people at sea, more persistent sensing, and data delivered in near real time.

At the same time, research points to a new generation of AI-enhanced vehicles designed to reduce cost and widen access. A 2024 conference abstract on the system known as Jolly Oscar describes a push toward more cost-effective AUVs so that more researchers can participate in advanced oceanographic exploration. That emphasis on affordability matters. If ocean exploration is to scale, capability alone will not suffice; the unit economics must improve.

Artificial intelligence and machine learning are likely to sharpen this transition. NOAA notes that as AUVs advance with improvements in AI and machine-learning technology, their data storage, sensor sophistication and imaging capabilities continue to develop. The promise is practical rather than theatrical: better route planning, more efficient classification of seabed imagery, improved anomaly detection, and reduced dependence on constant human supervision.

> The future of ocean exploration is less about sending heroes down and more about sending fleets out.

This is also why many experts and commentators now argue against a romantic “aquanaut” model. The Issues essay is explicit: relying largely on robots and remotely controlled submersibles is more economical, nearly as effective for investigating biodiversity, chemistry and seafloor topography, and far safer than human-led underwater exploration. In that respect, ocean science is converging on the same lesson that much of space science has already learned. Presence is not the same as productivity.

What remains at stake below the surface

The argument for exploring the ocean is not merely scientific curiosity, though that would be sufficient on its own. It is about governing a planet intelligently. The Future Vision paper links new observation technologies directly to societal needs, including the management of energy, ecosystems and raw materials, as well as the ocean’s effects on climate, weather and food security. Oceana Canada makes a parallel point from the conservation side: it is difficult to protect what we do not know, and only about 7% of the world’s oceans are designated as marine protected areas.

Mapping and characterisation are often mistaken for bureaucratic preliminaries. In fact, they are strategic assets. Better seafloor data informs responsible offshore energy decisions, marine resource management, habitat protection, and national security planning. That is why NOAA’s recent collaboration with Fugro frames deep-ocean data as useful across public and commercial domains. Knowledge in this context is not abstract. It shapes permissions, protections and priorities.

There is also a temporal issue. Environmental change is altering marine systems faster than traditional ship-based observation can comfortably track. Autonomous platforms offer persistence: they can revisit locations, monitor change, and extend coverage into places that are otherwise costly or hazardous to survey. If space exploration often asks what lies beyond Earth, ocean exploration increasingly asks what is changing beneath it, and how quickly.

Outlook

The 80% figure is frequently repeated because it captures a truth that is both empirical and civilisational. The ocean remains largely unexplored not because humanity is indifferent, but because the deep sea combines hostile physics with expensive logistics and historically fragmented institutions. Space won earlier because it had a clearer mandate and a grander apparatus behind it.

That gap may now narrow. AUVs, USVs, eROVs, cloud-connected operations and AI-enhanced sensing are making ocean exploration more continuous, more remote and potentially more affordable. NOAA’s long arc from 1960s AUV development to contemporary autonomous mapping programmes suggests steady technical maturation. Industry partnerships such as the Fugro agreement indicate that the infrastructure around these systems is also improving.

Still, one should resist easy symmetry. The ocean will not become simple because machines become smarter. Pressure will remain pressure. Navigation underwater will remain awkward. Data collection at depth will remain expensive relative to work done on land or in orbit. But the balance is shifting from heroic expeditions to scalable systems.

That is the real answer to the paradox. We explore space because we built the institutions to do so. We are only now building the tools, workflows and coalitions that could let us know our own planet with comparable seriousness. The next great age of exploration may be less about leaving Earth than finally understanding the largest part of it.