The water pressure at the bottom of the Mariana Trench reaches roughly 8 tons per square inch, and sunlight never reaches that depth. The first crewed vessel to arrive there did so in 1960, and the first robotic vehicles capable of returning video footage followed decades later. In 2018, researcher Sanae Shiba avoided the ocean entirely and searched an archive of video instead. She reviewed 5,010 previously overlooked dives recorded with submersibles, most of them controlled remotely or dropped as autonomous landers. One scene from 1998 stopped her: a plastic bag floating on the seafloor at 10,898 meters.

The trench was first identified on March 23, 1875, when the Royal Navy survey ship HMS Challenger lowered a weighted rope into the water. The rope kept running out. The reading came back at 4,475 fathoms, or roughly 8,184 meters, and a second survey in the same location confirmed it. That expedition, part of a four-year global voyage collecting oceanographic data systematically for the first time, recorded the number and moved on. Later researchers recognized it as the deepest place on the planet. Every measurement since has revised the figure as instruments improved, from weighted lines to sonar pulses to pressure sensors placed directly on the seafloor. In October 2010, the U.S. Navy survey ship USNS Sumner measured the bottom from the surface at 10,994 meters. The most precise figure comes from June 2020, when Victor Vescovo ran six transects across the deepest basins in the submarine Limiting Factor, with three autonomous landers alongside, cross-checking independent pressure readings taken at depth rather than relying on sonar from a ship above. The pressure sensors put the deepest observed seafloor at 10,935 meters, give or take 6 meters. That equals 35,876 feet, roughly 2,000 feet less than the figure of 38,000 feet that occasionally circulates online. Pressure rises by about one atmosphere for every 10 meters of descent. At Challenger Deep, that means roughly 16,000 pounds per square inch, about a thousand times what a person feels at sea level.

The bathyscaphe Trieste, the first vehicle to reach the bottom in 1960, had a steel sphere with walls 5 inches thick and room for two men in a space 38 inches wide. The crew endured the cramped, cold cabin for the five hours the round trip took. The challenge of building something that survives that environment is separate from the challenge of recovering usable footage from it. The solution turned out to be low-tech. Most of the filmed footage from deep-ocean trenches was not captured by a piloted submersible. It came from landers, frames carrying weights, floats, lights, and a camera, often with a baited trap. A ship drops them over the stern and they sink to the bottom. There are no cables and no joystick. The lander sits on the seafloor filming anything that approaches the light. When the timer runs out or an acoustic release is triggered, the anchor weights drop and the lander floats back to the surface on its own, often surfacing miles from where the ship expects it, depending on deep currents. Recovery crews sometimes spend hours tracking a radio beacon before they can retrieve the unit.

Alan Jamieson, a marine ecologist who has led much of the hadal research of the past two decades, designed the landers his teams used in the Pacific around a philosophy of simplicity and multiple redundant safety systems, specifically because a failure at depth cannot be repaired or rescued once the vehicle reaches the bottom. The advantage of simplicity is that nothing can snap and nobody is at risk, so multiple landers can be dropped at different depths at once. That is how depth records were broken. In August 2022, the research vessel DSSV Pressure Drop carried three high-resolution camera landers across three Japanese trenches over two months as part of a ten-year study of the planet’s deepest fish, a project built around systematic repeated surveys rather than a single dramatic attempt. The researchers involved noted that finding a rare animal at that depth generally requires dozens of drops before one confirmed sighting is recorded.

The Japanese remotely operated vehicle Kaiko worked differently, tethered to a ship, and made hundreds of dives into hadal depths beginning in 1995. It became the second vehicle in history, after Trieste, to reach the bottom of Challenger Deep, and continued collecting biological samples and footage across dozens of expeditions over the following eight years. The Woods Hole Oceanographic Institution built Nereus in 2008, a hybrid vehicle that could work freely or be steered from the surface through a fiber-optic cable the thickness of a human hair. The design combined the broad survey capability of a free robot with the precise real-time control of a tethered one. Nereus reached Challenger Deep in 2009 and sent back high-quality video over that glass thread, becoming the third vehicle to touch the bottom of that specific trench, after Trieste in 1960 and Kaiko in 1995. Neither machine remains in service. The way each ended illustrates what that depth does to equipment.

Kaiko was lost at sea in 2003 when its communication cable parted during a storm while it was working in a completely different part of the Pacific, far from the Mariana Trench. The failure was mechanical, caused by ordinary storm conditions rather than deep pressure. Nereus ended differently. On May 10, 2014, Nereus was seven hours into a planned nine-hour dive in the deepest part of the Kermadec Trench northeast of New Zealand. It had settled on the sediment at 9,990 meters for more than six hours. The team aboard the research vessel Thomas G. Thompson lost contact. Standard emergency recovery procedures did not work. The crew searched near the dive site and spotted debris floating on the surface, later confirmed to be parts of Nereus. Scientists concluded that a portion of the vehicle had collapsed under the pressure of roughly 16,000 pounds per square inch. That came on day 30 of a 40-day expedition, and Nereus was then the only operational scientific vehicle that could work at that depth. Timothy Shank, a Woods Hole biologist who helped create it, said Nereus allowed them to explore places they had never seen and ask questions they had never thought to ask. By the end of that voyage, the vehicle had already found previously unknown trench habitats and about a dozen new species, discoveries that remained valuable to researchers after the vehicle responsible for them was gone.

Risk existed from the first landing. On January 23, 1960, as Jacques Piccard and Lieutenant Don Walsh approached the bottom in Trieste, a sound like a gunshot came from inside the cabin. One of the exterior viewing windows made of Plexiglas had cracked. Pressure pushed the remaining layers tighter into the frame rather than shattering them inward, so the window held. The two men spent about twenty minutes on the seafloor and reported seeing living creatures there, which surprised nearly everyone, because prevailing scientific opinion at the time held that such pressure would make complex animal life impossible. Sixty-five years of footage later, the most frequently filmed object at that depth was not an animal. It was something discarded by humans: a plastic bag. At 10,898 meters in 2017, the Japan Agency for Marine-Earth Science and Technology opened its deep-sea debris database to the public. The archive held more than 30 years of images and videos of waste collected by Japanese submersibles and remotely operated vehicles, accumulated gradually across decades of separate research missions that had nothing to do with monitoring pollution when the footage was recorded.

A team led by researcher Sanae Shiba reviewed the archive and published findings in the journal Marine Policy in 2018. The team examined 5,010 dives. They counted more than 3,000 pieces of human-made waste: metal, rubber, fabric, fishing gear, and plastic. About a third of the classifiable debris was plastic, and of that plastic, 89 percent was single-use. The deepest item in the entire database was a plastic bag found at 10,898 meters in the Mariana Trench. The footage had been recorded in 1998 and sat in the archive for two decades before anyone specifically looked for it. Below 6,000 meters, more than half of the debris the team recorded was plastic, and most of it was single-use. Deep-sea creatures appeared in 17 percent of the debris images, including sea anemones attached to bags and plastic entangled in cold-seep communities, evidence that marine life was not only coexisting with the debris but in some cases growing directly on it. Many of the dives were more than 1,000 kilometers from the nearest land, ruling out a simple explanation of nearby coastal waste drifting a short distance offshore.

A year after that archival study, someone was at the bottom looking through a window. On April 28, 2019, Victor Vescovo took the submarine Limiting Factor to the bottom of the Mariana Trench as the fourth leg of his Five Deeps Expedition, an attempt to reach the deepest point of each of the five oceans over a continuous mission lasting about ten months. He spent about four hours there and reported seeing several creatures, some possibly new to science, alongside a plastic bag and candy wrappers. The observation carries weight because so few people have had the opportunity to make it. By July 2022, 27 individuals had reached Challenger Deep across about 22 crewed dives. The submersible DSV Limiting Factor alone logged 15 of those trips between April 2019 and July 2022, nearly doubling the number of people who had visited the site individually within a few years. The list includes Kelly Walsh, son of Don Walsh from the 1960 dive, and Kathy Sullivan, former NASA astronaut who became the first person to visit both space and the deepest point of the ocean. For comparison, human spaceflight began in 1961, and by November 2021 the 600th person had left the planet. The number of people who have seen Earth from orbit is more than twenty times the number who have sat at the bottom of Challenger Deep. The direct human observation sample is small, and waste was found in it twice.

The most common organism in Challenger Deep is the amphipod Hirondellea gigas, a scavenging crustacean that consistently appears in baited traps on the bottom and feeds on whatever organic material drifts down from above, including whale carcasses that support scavenger communities for years. Vertebrates have a much harder time. The deepest known one is the snailfish Pseudoliparis swirei, described in the Mariana Trench in 2017. It looks like a tadpole made of wet tissue paper. Chinese researchers who sequenced its genome found that its skull does not fully close, and a mutation seems to halt bone calcification early, leaving a thin skeleton that bends rather than shatters, an adaptation that helps the animal withstand pressure that might crush a fish with denser bones. The camera record for any fish was set in August 2022 and announced in April 2023. Jamieson’s team, working with the University of Tokyo’s marine science and technology department, filmed a small Pseudoliparis snailfish at 8,336 meters in the Izu-Ogasawara Trench, breaking the previous record of 8,178 meters in the Mariana Trench. Days later they caught two Pseudoliparis belyaevi at 8,022 meters in the Japan Trench, the first fish caught below 8,000 meters, meaning researchers had physical specimens rather than only video. They were 20 to 25 centimeters long and fed on small crustaceans. Jamieson noted that the Izu-Ogasawara Trench is the warmest of the deep trenches, less than one degree warmer than the Mariana Trench, and that the difference matters. If anyone finds fish at greater depth, it will not be by much.

No fish has been found in the deepest quarter of the ocean, roughly between 8,400 and 11,000 meters. For a long time nobody could explain why. In 2014, a team including Paul Yancey of Whitman College, Mackenzie Gerringer, Jeffrey Drazen, Ashley Rowden, and Jamieson published an answer in the Proceedings of the National Academy of Sciences. The key is trimethylamine N-oxide, or TMAO, a compound that belongs to a group known as piezolytes. Pressure tries to distort the shape of proteins at a molecular level, and TMAO stabilizes them, acting essentially as a molecular scaffold that protects the three-dimensional shape of the protein from collapsing under the weight of water. Fish carry more of it the deeper they live, rising from about 40 to 261 millimoles per kilogram between the surface and 4,850 meters. The problem is that TMAO also alters the salinity of the animal’s body from the inside. Shallow-water fish operate at about 350 units of internal osmotic concentration compared to 1,100 units for seawater. Add enough TMAO and those two levels equalize, at which point the fish’s internal fluid balance works against rather than for it. The team caught five snailfish from 7,000 meters in the Kermadec Trench, measured the highest TMAO levels ever recorded in a fish, and identified where the curves intersect. The answer was about 8,200 meters. Below that depth, the fish’s body would have to reverse its own water balance. Yancey pointed to a second limit. Too much TMAO makes proteins so stable they stop working entirely. Myosin in muscles must move for contraction, and enough TMAO would freeze it completely, meaning that even if a fish solved the osmotic problem, it would eventually reach a point where its muscles could no longer contract. There is a hard ceiling for vertebrates, and the snailfish filmed at 8,336 meters sits right at that boundary. Anything living deeper must be built on different rules.

In July 2025, a Chinese submarine showed what that might look like. Between July 8 and August 17, 2024, the research vessel Tansuo Hao carried the submarine Fendouzhe, meaning “Striver,” through sections of the Kuril-Kamchatka Trench and the western Aleutian Trench in the northwest Pacific. The submarine had already reached Challenger Deep in 2020, carrying three Chinese scientists to 10,909 meters. On dive number 271, at 9,533 meters, the crew found dense clusters of tube worms on black mud where the trench floor meets the base of the slope. Elsewhere they found expanses of clam-like bivalves buried in the sediment so densely that one site was called a “clam bed.” The findings appeared in Nature on July 30, 2025, from a large international team led by Shao Tong Bang and Muqin Du of the Institute of Deep-Sea Science and Engineering at the Chinese Academy of Sciences. These communities stretched across roughly 2,500 kilometers, with depths ranging from 5,800 meters down to 9,533 meters. They constitute the deepest and most widespread chemosynthetic community known on Earth. Nothing there depends on sunlight. Fluids rich in hydrogen sulfide and methane move through faults that penetrate deep sediment layers, and isotope analysis indicated the methane is produced by microbes breaking down buried organic matter. Bacteria convert those chemicals into usable food, and many of the tube worms and clams host those bacteria inside their tissues, an arrangement biologists call chemosynthetic symbiosis. That possibility had been proposed for decades, but actual documentation was exceptionally rare. The authors noted that given the geological similarity among other trenches, such communities may be far more common than anyone assumed.

In February 2017, a team led by Jamieson published a study in Nature Ecology and Evolution that changed how researchers discussed deep-ocean contamination. They used landers to collect amphipods, small shrimp-like scavengers, from depths between 7,000 and 10,250 meters in the Mariana and Kermadec trenches. The two trenches sit about 7,000 kilometers apart on opposite sides of the Pacific, making it difficult to dismiss the results as a local anomaly. They tested the fatty tissue for persistent organic pollutants. Two stood out: polychlorinated biphenyls, or PCBs, used for decades as electrical insulators, and polybrominated diphenyl ethers, used as flame retardants. Regardless of depth, species, or trench, the animals were loaded with these pollutants. The highest PCB levels found in Mariana Trench amphipods were 50 times higher than those in crabs from China’s Liaohe River system, one of the most polluted river systems in the country. PCBs were banned starting in the 1970s, and total global production from the 1930s to the ban was roughly 1.3 million tons. These compounds do not break down naturally.

In February 2019, Jamieson and colleagues published a study in Royal Society Open Science examining the hindguts of 90 amphipods collected from six trenches around the Pacific rim: Japan, Izu-Bonin, Mariana, Kermadec, New Hebrides, and Peru-Chile, at depths ranging from 7,000 to 10,890 meters. Sixty-five of the 90 animals, more than 72 percent, had ingested at least one microplastic particle. In the Mariana Trench the rate was 100 percent. Every animal examined from the deepest place on Earth contained at least one human-made fiber. Chemical analysis revealed a list resembling a clothing label: nylon, polyethylene, polyamide, polyvinyl alcohol, polyvinyl chloride, plus rayon and lyocell. Jamieson offered a comparison for what these fibers do to an animal that small, describing the human equivalent as swallowing a two-meter polypropylene rope and expecting no side effects. While examining amphipods collected in November 2014 from the South Sandwich Trench, Joanna Weston, a doctoral student at Newcastle University, found microfibers in the hindgut of a specimen belonging to a species new to science. Analysis showed it matched roughly 84 percent of PET, the plastic used in drink bottles. The team named the species Eurythenes plasticus and described it in Zootaxa on March 5, 2020, specifically to show that a species new to science from an unexplored habitat had already been contaminated before humans even knew it existed. Heike Vesper of WWF Germany, whose organization supported the research, said there are species living in the deepest, most remote places on Earth that have ingested plastic before humanity was aware of them.

Nuclear weapons tests in the late 1950s and early 1960s nearly doubled the amount of carbon-14 in the atmosphere, an increase researchers call “bomb carbon.” In 2019, a team led by geochemist Ning Wang of the Chinese Academy of Sciences published a study in Geophysical Research Letters using that signal as a tracer. They collected amphipods from the Mariana, Mussau, and New Britain trenches to depths of 11 kilometers. Muscle tissue showed carbon-14 levels far higher than the organic matter floating in the deep water around them, and instead matched organic material collected from the sunlit surface of the Pacific. The conclusion was that these animals feed on debris falling from the surface rather than living on local material, meaning the surface and the trench floor are linked far more closely than anyone assumed. The study also found that trench amphipods grow larger and live longer than their shallow-water relatives, giving pollutants more time to accumulate inside a single animal over its lifetime.

Human activity releases more than 2,000 metric tons of mercury into the atmosphere each year. In November 2020, a team led by Joel Blum of the University of Michigan published research in the Proceedings of the National Academy of Sciences showing that mercury isotope ratios in snailfish and amphipods from the Mariana and Kermadec trenches matched mercury found in fish feeding at about 500 meters depth in the central Pacific, waters dominated by mercury of human origin. Blum summarized it simply: the mercury that was once in the stratosphere is now in the deepest trench on Earth. A second team, led by Ruo Sun of Tianjin University, reached a compatible conclusion using landers in the Mariana and Yap trenches.

On May 7, 2025, researchers from the Ocean Discovery League, led by Katy Croff Bell, published a study in Science Advances compiling 43,681 submersible dive records dating back to 1958 across the waters of 120 countries. The team calculated that all those dives combined visually observed about 3,823 square kilometers of deep seafloor, less than 0.001 percent of the total deep seafloor worldwide. That is an area slightly larger than Rhode Island, or about one-tenth the size of Belgium. Sixty-five percent of all observations came from the waters of just three countries: the United States, Japan, and New Zealand. The authors compared that to trying to understand all terrestrial wildlife from a sample of land the size of Houston. Bell’s point was not that nothing is known, but the opposite: consider how much has been extracted from that tiny fraction, then ask how much remains. Mapping the seafloor with sonar is faster than observing it visually, and even that lags behind full coverage. On April 20, 2026, the Nippon Foundation and the GEBCO Seabed 2030 project announced that 28.7 percent of the world’s ocean floor has now been mapped to modern standards, about 104 million square kilometers, with roughly 5 million square kilometers added in one year. When the project began in 2017, the figure was 6 percent.

The current picture is straightforward. Cameras have reached the deepest parts of the ocean since the 1990s, mostly on simple landers that drop to the bottom, film, and float back to the surface. The deepest point, Challenger Deep, is 10,935 meters, or 35,876 feet, according to the most accurate measurements available, not 38,000 feet. Only 27 people had reached it as of July 2022. Machines do not always return intact. Kaiko was lost in 2003, and Nereus collapsed under pressure at 9,990 meters in 2014. The most striking part of the footage has never been a monster. It is a plastic bag filmed in 1998 at 10,898 meters, among more than 3,000 documented pieces of human waste, with PCBs, microplastics, bomb carbon, and mercury already found inside the animals that live there. There is also confirmed extraordinary life: a snailfish filmed exactly at the chemical boundary researchers predicted for vertebrates in 2014, and an entire chemosynthetic community of tube worms and clams living on methane and hydrogen sulfide up to 9,533 meters deep. Humans have visually observed less than 0.001 percent of the deep seafloor. The plastic bag arrived there before the cameras did.