MIT muscle-film robot swims with living cells instead of a motor

A thin robot can move through water without an electric motor. MIT researchers grew living muscle on grooved gel and arranged it into two fins. Illuminating either side makes that side move, allowing the device to turn. The work first appeared as a preprint in May 2026, before journal publication in September.

Muscle-powered swimming machines already have a history. In 2016, a Harvard-led team built a soft robotic ray using rat heart muscle, with separately illuminated fins providing steering. A 2022 Harvard–Emory collaboration developed a fish made with human cardiac cells that could sustain its own rhythmic motion. Neither muscle propulsion nor light-controlled swimming begins with this project.

The MIT study instead investigates thin films of skeletal muscle, the type normally used to move limbs deliberately. The paper reports roughly twenty times the force per unit tissue volume compared with the three-dimensional skeletal-muscle structures in its comparison. That is force density, not a twentyfold increase in swimming speed or a twentyfold reduction in total energy use.

Organization matters as much as adding cells. Think of rowers pulling together rather than against one another. Microscopic grooves guide the muscle fibers into alignment; tuning the supporting gel helps contractions move the structure instead of merely crumpling it.

The project follows the laboratory's 2025 artificial-iris work. Researchers stamped circular and radial patterns into gel, directing tissue to contract in multiple directions. That demonstration established coordinated shape changes. The swimming study uses its thin-film architecture to move an entire device from place to place.

The maze demonstration still took place in a petri dish. Researchers moved an external light source by hand to guide the swimmer. It was not observing its surroundings and choosing an exit, and the experiment did not establish that a damaged robot could repair itself.

For a reader outside a laboratory, the interesting possibility is a different construction strategy for small machines: combine contractile cells with a soft support rather than simply shrink a conventional motor. Controlled movement is one step toward delicate aquatic tasks. Reliable operation beyond the experimental setup will require further work on the surroundings and control.

Original preprint: May 8, 2026. Journal publication: September 28, 2026.