The paper's swimmer follows a light in a dish, at four body lengths a minute
A September paper describes a two-fin robot a handheld light steered through a dish. The abstract's straight-line pace is about four body lengths a minute.
The short version
The paper and MIT's account describe a two-fin gel robot lined with a thin, light-responsive muscle film. In a dish, a handheld light steered it through a maze, at up to about four body lengths a minute in the straight-line figure the abstract gives. The swim is a laboratory demonstration.
The swim the abstract is willing to claim
The paper, first published September 28, 2026, describes a locomotive robot whose fins are independent thin films of aligned skeletal muscle on a micropatterned hydrogel. The abstract says those fins can be switched and sped by the control inputs the authors use, and that the device reached steered movement at up to four body lengths per minute in a straight line and 1,200 degrees per minute in rotation. A body-length speed is a laboratory ratio. It is not a river crossing.what a control loop closes.
MIT's dish is the scene
MIT News, dated September 29, says the skeleton is a gel about the length and width of a stick of gum, each fin covered by a muscle layer thinner than a hair, and that the cells twitch when lit. Shining a light on one fin flaps that fin. Alternating the light steers. The engineers put the robot in a petri dish and moved the lamp by hand through a simple maze. The fastest pace in that account matches the abstract's four body lengths a minute.the robotics desk.
The multipliers belong to the paper
The abstract says the films generated forces on the order of ten times prior 2D skeletal-muscle actuators, raised force per volume about twenty times versus 3D muscle in the comparison the authors draw, and extended untethered actuation from under ten minutes to more than thirty days, which they describe as roughly a 4,500-fold longevity gain. Those are the authors' comparisons inside one study. This page does not repeat the fabrication steps that produced the films.Futureweb.
October's wire is a reprint of September's result
Robohub's item, carried on the October 19 robotics wire, retells the MIT demonstration: a paper-thin robot, muscle cells, a watery maze. The experiment's date is the paper's date. The wire's date is the day a robotics index surfaced it again. A retelling can be accurate and still not be a new trial. Asked directly, What speed does the abstract report? Up to four body lengths per minute in straight movement, and up to 1,200 degrees per minute in rotation, with on-off and frequency-dependent control of independent fins. Those figures are the paper's.a different machine, on display.
Light from outside is the control loop
The news account says the robot has no onboard electronics for steering. The loop closes in a person's hand, above the dish. That is a real control loop, and it is an external one. A robot that follows a lamp in a dish has not been shown to hold a heading in a current, and the authors say the next goal is a faster body, with environmental monitoring as a someday use rather than a cruise they have run. A process node is a different measurement.what a process node measures.
What this page will not teach
It will not describe how the cells were prepared, altered, or trained. It will not treat a petri-dish maze as a waterway. It will not adopt the abstract's list of possible futures, from surgery to exploration, as work the device has done. The record is a paper, a university account, and a later reprint. Asked directly, Where did MIT say the maze was? A large petri dish of water. A researcher moved a light by hand. The robot followed by flapping the fin the light struck. That is the news account of the demonstration, not an open-water trial.
What a later trial would change
A swim outside a dish, a speed measured by a second laboratory, or a body that carries its own light would be a different robot from the one in the September paper. Until then, the dependency is the thin-film fin, the external light, and the speeds the abstract prints. Asked directly, Who are the authors, and where did the paper appear? Maheera Bawa and colleagues at MIT, including corresponding author Ritu Raman, in Advanced Functional Materials. The journal dates the paper September 28, 2026. MIT News ran its account the next day.
A preprint is not a second result
A May 2026 preprint carries a related manuscript identifier. The version this brief files is the journal article of September 28 and the MIT account of the demonstration. A preprint can share sentences with the paper and still not be the version of record. The DOI on the journal article is the one in the sources. Asked directly, What does the abstract claim relative to thicker muscle robots? It says the thin films produced higher force per volume than the team's comparison to 3D muscle constructs, and that untethered longevity in the study ran past 30 days against a prior window under 10 minutes. Those multipliers are the paper's comparison, not this desk's rerun.
What stays on the record
September 28, a two-fin thin-film swimmer, a handheld light, a dish, about four body lengths a minute, a rotation figure the abstract states, and an October reprint on the robotics wire. The muscle is a laboratory film. The water is a dish. The short answer this page will stand on is this: The paper and MIT's account describe a two-fin gel robot lined with a thin, light-responsive muscle film. In a dish, a handheld light steered it through a maze, at up to about four body lengths a minute in the straight-line figure the abstract gives. The swim is a laboratory demonstration.
Sources
The reports this brief is filing. Futureweb did not republish them.
Questions
Who are the authors, and where did the paper appear?
Maheera Bawa and colleagues at MIT, including corresponding author Ritu Raman, in Advanced Functional Materials. The journal dates the paper September 28, 2026. MIT News ran its account the next day.
What speed does the abstract report?
Up to four body lengths per minute in straight movement, and up to 1,200 degrees per minute in rotation, with on-off and frequency-dependent control of independent fins. Those figures are the paper's.
Where did MIT say the maze was?
A large petri dish of water. A researcher moved a light by hand. The robot followed by flapping the fin the light struck. That is the news account of the demonstration, not an open-water trial.
What does the abstract claim relative to thicker muscle robots?
It says the thin films produced higher force per volume than the team's comparison to 3D muscle constructs, and that untethered longevity in the study ran past 30 days against a prior window under 10 minutes. Those multipliers are the paper's comparison, not this desk's rerun.
Why is Robohub a source?
The robotics wire carried Robohub's October account of the same MIT demonstration. It is a second telling. It does not add a new experiment.