Hirotaka Sato's cyborg cockroach can now work underwater for three hours

Why it matters

Miniature robots are still constrained by batteries and motors. Sato's work shows a path that uses biology for locomotion while engineering around its limits, starting with oxygen.

Hirotaka Sato, the NTU Singapore professor who has spent years trying to turn insects into field robots, has led a new cyborg-cockroach system that kept a remote-controlled Madagascar hissing cockroach moving underwater for as long as three hours.

The work, published June 29, 2026 in Nature Communications, is narrower and more useful than the viral framing around camera-equipped robot bugs suggests. Tom's Hardware described the system as a cyborg cockroach with IR cameras that can breathe underwater. The peer-reviewed paper centers on the missing layer in Sato's disaster-robotics stack: a flexible, waterproof, 3D-printed diving suit that supplies oxygen directly to a cockroach's spiracles while electronics steer the insect.

That distinction matters because remote control is old ground for Sato. Guinness World Records credits him with the first untethered controlled flight by a cyborg beetle, achieved at Nanyang Technological University on October 5, 2015. His earlier work at UC Berkeley included radio control of insect flight, part of a field that has long been pulled between defense research, search-and-rescue promise, and public discomfort with living-machine hybrids. The new paper pushes the same idea into a common disaster-zone failure mode: water.

Sato's lab has a plain thesis. The Biohybrid Robot Research Group says it is building insect-computer hybrid robots for search and rescue, unsafe work environments and exploration of unknown areas. Sato's own motivation was shaped by the March 2011 earthquake in Japan. In a 2023 Nature profile, he recalled watching search-and-rescue efforts after the quake and thinking, "I need to develop the technology to save people."

The engineering bet is oxygen, not legs

The reason cyborg insects keep attracting serious research attention is power. Small robots run into battery limits fast because motors, sensors and locomotion all need energy. A living insect brings its own muscles, gait, grip and terrain handling. The electronic controller can be comparatively small because it is steering rather than powering every movement.

Water breaks that bargain. Cockroaches breathe through spiracles, small openings that feed the insect's tracheal system. Submerge the animal, and the platform loses its respiratory supply even if the control electronics keep working.

Sato's team, with collaborators at Waseda University, built a wearable suit around that constraint. The system uses a flexible shell, an oxygen-generation tank and silicone oxygen-delivery tubes connected to thoracic spiracles. The oxygen tank avoids a tiny pressurized cylinder. Instead, it uses diluted hydrogen peroxide and a manganese dioxide catalyst on a cellulose sponge to generate oxygen, which is then routed to the insect's breathing openings.

The paper reports that the suit-wearing cyborg insects stayed active and responsive to electrical stimulation for up to three hours underwater. A control cockroach without the suit suffocated within two minutes in the underwater test condition. In locomotion tests, the suited insects moved at 87.5 mm per second on land and 78.4 mm per second underwater, a 10.4% reduction in forward speed. After three hours submerged, forward speed fell to 52.3 mm per second.

The researchers also tested the system beyond a simple tank demo, immersing suited cyborg insects at depths from 5 cm to 50 cm and running mixed-hazard trials that combined carbon-dioxide exposure with a water-filled section.

A research result, not a startup launch

There is no company, venture round or commercial deployment attached to the paper. The authors list support from Singapore Ministry of Education grant RG82/24 and Waseda's Top Global University Project. The paper names Zifu Fan, Kazuki Kai, Kewei Song, Duc Long Le, Thu Ha Tran, Mingyu Hao, Wei Yang Wan, Shinjiro Umezu and Sato as authors, with Sato supervising the work.

That absence of commercialization is important because the strongest claims here are lab claims. The system has not been shown in a live disaster deployment with the new diving suit. The Nature paper demonstrates underwater survival, guided movement, mixed hazard traversal and narrow-crevice movement under controlled conditions. It does not establish field logistics: how many insects can be prepared at once, how operators would coordinate them under rubble, how sensor payloads would perform in muddy water, or how rescue agencies would handle the biological and ethical edge cases.

Sato's advantage is that he is attacking the robotics problem from the opposite direction of most miniature robot builders. He starts with an animal that already crawls, grips, survives impacts and fits through gaps, then adds steering and sensing. The underwater suit extends that bargain into flooded spaces, which are exactly where many compact robots struggle.

The uncomfortable edge is part of the product problem

Cyborg insects will always carry an adoption problem that wheeled robots and drones do not. A rescue robot that looks like a dog, a drone or a crawler is easier for institutions to explain. A living cockroach with implanted electrodes and a breathing apparatus invites questions about animal welfare, field handling, public acceptance and failure modes.

Sato's work is still moving toward a concrete use case. Flooded rubble, drains, tunnels and partially submerged spaces are awkward for conventional robots because they combine water, tight geometry, debris, darkness and poor communications. A steerable insect that can carry a sensor into that kind of space for hours would be valuable if it can be prepared reliably and controlled at scale.

The June 29 paper does not make the cyborg cockroach a deployable rescue system. It solves one of the barriers that kept the idea from reaching flooded environments. For Sato, who has spent more than a decade moving insect cyborgs from flight control to cockroach navigation, that is the practical step: keep the biological platform alive long enough for the electronics to matter.