The jaws are small enough to hold in your hand. But the animal they belonged to may have stretched 62 feet from arm tip to arm tip, longer than a school bus, longer than most mosasaurs that patrolled the same seas. In a study published in Science in May 2026, a team led by paleontologist Yasuhiro Iba of Hokkaido University announced the discovery of 27 fossilized octopus jaws from the Late Cretaceous period, roughly 100 million years ago. The specimens suggest that the earliest known octopuses were not the small, soft-bodied scavengers scientists long assumed. They were massive predators built to crush armored prey.
Jaws hidden inside stone
Fifteen of the 27 jaw specimens had been collected years earlier from Upper Cretaceous sediments in Hokkaido, Japan, within a geological formation called the Yezo Group. The remaining 12 were found without anyone picking up a rock hammer. Iba’s team developed a technique they call “digital fossil-mining”: they ground rock concretions into paper-thin slices, photographed each layer in full color, then fed the images to a zero-shot learning AI model trained to spot fossil structures that human eyes routinely miss.
The approach nearly doubled the known sample in a single step. A Nature news analysis noted that the jump from 15 jaws to 27 transformed what had been a handful of curiosities into a dataset large enough for meaningful statistical comparison. The AI pipeline was independently validated on sepioid beaks from similar concretions, confirming it could reliably extract small, subtle fossils from visually uniform rock.
The jaws themselves tell a vivid story. According to Hokkaido University’s summary of the research, up to roughly 10% of the jaw tip on some specimens had been ground away through repeated use, a wear pattern consistent with biting into hard-shelled prey like ammonites or marine crustaceans. The researchers assigned the 27 specimens to two distinct species, both substantially larger than any octopus alive today.
How big, exactly?
Size estimates derived from the jaws place these animals between 7 and 19 meters long, or roughly 23 to 62 feet. For perspective, the giant Pacific octopus, the largest living species, tops out around 16 feet. A colossal squid reaches perhaps 46 feet. The upper bound of 62 feet would put these Cretaceous octopuses in the size class of large marine reptiles like mosasaurs and plesiosaurs that shared their waters.
But that 62-foot figure is a ceiling, not a tape measurement. Octopus bodies are almost entirely soft tissue, which decays long before fossilization can preserve it. The scaling relationships used to project total length from jaw size are borrowed from modern cephalopods that are orders of magnitude smaller than the animals being reconstructed. The Science paper itself presents the 62-foot estimate not as a firm figure but as the upper end of a range accompanied by wide confidence intervals, reflecting the statistical uncertainty inherent in extrapolating from such small anatomical proxies to whole-body dimensions. Neil Landman, a research associate emeritus at the American Museum of Natural History, told reporters that “even modest changes in assumed body proportions can shift total length estimates by many meters in either direction.” No complete body fossil exists to anchor the math.
Readers should treat the range of 23 to 62 feet as the best current bracket, not a confirmed dimension. Future discoveries or refined modeling could tighten it considerably.
Apex predators or powerful neighbors?
Iba described the findings as evidence that the earliest octopuses were “giant top predators.” The jaw wear supports the idea that these animals regularly attacked hard, resistant food sources. But the leap from “large jaws with grinding damage” to “apex predator that dominated Cretaceous oceans” involves interpretive steps that not all paleontologists accept without reservation.
No direct gut-content evidence or bite-mark matching on prey shells has been found. Some specialists cited in press coverage have proposed alternative scenarios in which these octopuses were powerful but still shared the upper tiers of the food web with large fish, marine reptiles, and other cephalopods. The Cretaceous ocean was crowded at the top. Mosasaurs, large sharks, and predatory fish all competed for similar prey, and the ecological relationships among them remain poorly resolved.
Isotopic analysis of the jaw material could eventually help clarify where these octopuses sat in the food chain. Ratios of certain elements in fossilized hard parts can reveal trophic level, the position an organism occupies relative to its prey. That work has not yet been done on these specimens, but Iba’s team has signaled interest in pursuing it.
A preservation puzzle
All 27 jaws come from a single region: Hokkaido’s Yezo Group, a formation of fine-grained marine concretions unusually good at trapping small, hard structures like beaks and jaws. That geographic concentration raises an obvious question: were giant octopuses confined to the western Pacific, or do they only appear to be because Hokkaido’s rocks happen to preserve them?
Comparable sediments elsewhere in the world have not yet yielded similar jaws, but many have never been surveyed with the digital fossil-mining methods Iba’s team pioneered. The technique’s success in Hokkaido suggests that existing museum collections of unexamined rock concretions could harbor overlooked specimens. If other institutions adopt the same AI-assisted workflow, the known range of these animals could expand well beyond Japan.
The two-species classification is also provisional. The differences separating the species rely on subtle variations in curvature, thickness, and relative jaw-element size. With only 27 specimens, it is difficult to determine whether those differences reflect true species-level separation, sexual dimorphism, or normal variation within a single widespread population. More jaws, especially from different age classes or geographic regions, will be needed to test whether the current taxonomy holds.
What the method may matter more than the monster
The most lasting contribution of this study may not be the headline creature but the tool that found it. Paleontology has long been limited by what the human eye can spot in the field or the lab. Iba’s digital fossil-mining pipeline sidesteps that bottleneck. By grinding, photographing, and computationally scanning rock that would otherwise sit unstudied in storage, the method opens a path to discoveries locked inside collections that have already been “finished” by traditional standards.
The implications reach beyond octopuses. Any small, hard fossil embedded in a concretion, from fish otoliths to microbial structures, could potentially be detected the same way. As the technique spreads, the picture of ancient marine ecosystems is likely to sharpen across many lineages, not just cephalopods.
For now, the giant Cretaceous octopus sits at an intriguing midpoint between solid evidence and spectacular inference. The jaws are real. The wear marks are real. The size and ecological role remain educated projections, grounded in data but awaiting the kind of confirmation that only more fossils can provide. What vanished with the dinosaurs 66 million years ago may have included a tentacled predator far larger and more formidable than anything the modern ocean has produced. Proving it will take more rock, more grinding, and more patience.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
