Living Muscle Cells Power a Tiny Swimming Robot: How MIT Is Changing the Future of Robotics
A Tiny Robot Powered by Living Muscle
Robots are usually associated with electric motors, metal gears, batteries and sophisticated electronic components. However, a new generation of machines is beginning to challenge this traditional approach by combining engineered structures with living biological tissue.
Researchers at the Massachusetts Institute of Technology (MIT) have developed an exceptionally thin swimming robot powered by living skeletal muscle cells. Instead of relying on a conventional motor, the device uses biological tissue that contracts when exposed to light, generating the force needed to propel it through water.
The robot successfully swam and changed direction through a simple underwater maze in laboratory experiments. Although it remains a research prototype rather than a practical autonomous machine, the experiment represents an important step towards more efficient biohybrid robots.
The central idea is surprisingly simple: rather than building an artificial motor that imitates muscle, engineers can use real muscle tissue as the robot's actuator.
This approach could eventually help researchers design machines that are softer, more adaptable and potentially better suited to delicate environments than some conventional robotic systems.
The research was published in Advanced Functional Materials in September 2026. It demonstrates how carefully engineered muscle tissue and flexible materials can work together to produce controlled movement.
What Is a Biohybrid Robot?
A biohybrid robot combines living biological components with artificial materials and engineered structures.
Traditional robots generate movement through components such as electric motors, hydraulic systems, pneumatic actuators or shape-changing materials. Biohybrid machines take a different approach by incorporating living cells or tissues into the mechanism responsible for movement.
These biological components can perform functions that are difficult to reproduce using purely artificial materials.
Skeletal muscle cells, for example, naturally contract when stimulated. In living organisms, this process produces movement through the coordinated activity of muscle fibres. In a laboratory, scientists can cultivate muscle tissue and integrate it with a supporting structure so that its contractions move an artificial device.
The result is a machine whose body is partly engineered and partly biological.
Biohybrid robotics is not entirely new. Previous research has explored small swimming devices powered by cardiac cells, skeletal muscle tissue and other biological components. Scientists have also developed soft robots inspired by animals such as fish, rays and worms.
The MIT project builds on this wider field but addresses an important engineering challenge: how to make thin layers of skeletal muscle generate enough force to move a robot efficiently.
Instead of relying on bulky three-dimensional muscle structures, the researchers created a design based on thin films of aligned muscle cells supported by a specially engineered hydrogel.
This approach could reduce the amount of biological material needed while improving the efficiency with which muscle contractions produce useful movement.
How Does the Living Muscle Robot Work?
The device consists of a thin, flexible gel structure that forms two small fins. Each fin is covered with a layer of living skeletal muscle cells.
The supporting material is gelatin methacrylate, commonly known as GelMA, a hydrogel used in tissue engineering. Its properties can be adjusted to provide a suitable environment for cells to grow and generate force.
The researchers engineered the surface of the gel with microscopic grooves. These patterns encourage muscle cells to align in a coordinated direction, helping them form stronger muscle fibres.
The muscle cells are genetically engineered to respond to light. When the researchers illuminate a fin, the cells contract, causing the fin to move. This movement pushes against the surrounding water and propels the robot forward.
Because the two fins can be activated independently, researchers can influence the robot's direction as well as its movement.
Illuminating one fin produces an asymmetric movement that can help turn the robot. Activating both fins can produce coordinated movement, allowing the machine to swim.
The experiment demonstrated that a very thin layer of living tissue can generate enough mechanical force to move a small device through water.
This is significant because water creates considerable resistance for small swimming machines. Even a lightweight robot must generate sufficient force to overcome that resistance and change its position.
The researchers therefore focused not only on the biological cells themselves but also on the material supporting them.
Why the Researchers Changed the Robot's Skeleton
One of the most important elements of the project was the development of a better supporting structure for the muscle tissue.
In earlier experiments, the researchers used fibrin, an extremely soft biological gel. Although this material can support living cells, the team found that it could shrink or deform under the forces generated by contracting muscle.
When the supporting material changes shape too easily, some of the force produced by the muscle is lost rather than being transferred efficiently into movement.
The researchers therefore investigated different gel compositions, stiffness levels and surface patterns.
They found that gelatin methacrylate provided a useful alternative. By adjusting its formulation, they could create a firmer structure that better supported muscle growth and contraction.
The shape of the microscopic grooves also mattered. Cells growing in square-bottomed grooves aligned more effectively than cells growing in some of the alternative patterns tested.
This alignment encouraged the cells to form organised muscle fibres that could contract together.
The team also adjusted the thickness of the supporting gel. A film approximately half a millimetre thick provided enough support while remaining sufficiently thin for the muscle tissue to move it.
The research demonstrates an important principle in biohybrid engineering: successful biological robotics depends not only on cultivating living cells but also on designing the surrounding materials so that their forces can be used effectively.
Training Living Muscle to Produce More Force
The researchers also explored whether repeated stimulation could improve the performance of the muscle tissue.
Muscle is a responsive biological system. Its behaviour can change depending on how it is stimulated and the conditions in which it develops.
To take advantage of this property, the team exposed the engineered muscle to repeated pulses of light as a form of training.
The stimulation encouraged the muscle cells to contract regularly before the tissue was used to power the robot.
The researchers reported that this training routine increased fin displacement approximately fourfold compared with an unstimulated control group.
This finding suggests that biological actuators can be improved through a combination of material engineering, cellular organisation and carefully designed stimulation.
It also illustrates a broader difference between conventional and biohybrid robotics. Engineers building a conventional motor primarily optimise its mechanical and electrical components. In a biohybrid system, they must also consider how living cells grow, organise themselves and respond to repeated activity.
The performance of the final device therefore depends on both engineering design and biological behaviour.
How Fast Can the Robot Swim?
The robot's performance is impressive for a device powered by a thin layer of living muscle, but its current speed remains modest.
According to the published research, the machine can swim at approximately four body lengths per minute during straight movement.
Its rotational movement can reach approximately 1,200 degrees per minute under the reported experimental conditions.
These measurements describe different aspects of movement: the first measures forward travel relative to the robot's size, while the second measures how quickly it rotates.
The researchers demonstrated that the robot could swim and steer through a simple maze in a laboratory dish. A researcher manually directed a light source towards the fins to control its movement.
This is an important distinction. The robot demonstrated controlled locomotion, but it was not an independently navigating underwater vehicle. Its movement depended on an external light source operated by a researcher.
The device also does not yet have the autonomy, durability in open water or navigation systems needed for real-world underwater missions.
Nevertheless, the experiment establishes that an exceptionally thin, muscle-powered robot can move through water rather than merely contract in place.
The researchers' next objective is to improve its body design and increase its swimming speed.
A Major Improvement in Muscle Efficiency
The research paper reports several quantitative improvements over earlier muscle-powered designs.
The team demonstrated that its optimised two-dimensional muscle actuators could generate forces on the millinewton scale and movements on the millimetre scale.
The authors also reported approximately 20 times greater contractile force per unit volume of muscle than in the three-dimensional muscle actuators used for comparison.
Another notable result concerned longevity. The optimised actuators remained functional for more than 30 days under laboratory culture conditions, compared with less than 10 minutes for the previous two-dimensional actuators used as a reference.
The paper describes this as an improvement in untethered actuation longevity of approximately 4,500 times.
These results suggest that carefully engineered thin muscle layers could address some of the limitations that have made muscle-powered robots difficult to scale and operate over longer periods.
However, laboratory longevity should not be confused with operational lifetime in the natural environment. Maintaining living tissue requires appropriate conditions, and the reported results do not establish that the robot could survive for a month in an ocean, river or lake.
The results instead provide a foundation for future research into biological actuators that can remain functional for longer periods while generating useful mechanical force.
Could Living Muscle Robots Monitor the Environment?
One potential application of biohybrid robotics is environmental monitoring.
Many conventional underwater robots are designed for relatively demanding tasks and may be too large, rigid or expensive for certain delicate environments.
A smaller, softer robot could potentially operate in confined spaces or near sensitive biological structures where contact with hard mechanical components might cause damage.
Future biohybrid swimming machines might be investigated for tasks such as:
- Monitoring water conditions in confined aquatic environments.
- Exploring delicate biological habitats.
- Conducting experiments in laboratory water systems.
- Investigating movement through small or complex spaces.
- Providing researchers with new ways to study interactions between biological tissue and engineered structures.
These are possible future applications, not capabilities demonstrated by the current prototype.
The present robot is slow, requires external light control and has been tested in a controlled laboratory setting. Real-world environmental monitoring would require additional developments in navigation, sensing, power delivery, biological maintenance and protection against environmental conditions.
Nevertheless, the research suggests that biological actuators could become useful components in specialised soft robots.
Could Biohybrid Robots Transform Medicine?
Medical robotics is another field where soft, small-scale machines could eventually become useful.
Some medical procedures require devices to move through narrow spaces or interact with delicate tissue. A rigid mechanical component may not always be the ideal solution for such environments.
Researchers are therefore exploring soft robotic systems that can bend, deform or respond to external stimulation.
Biohybrid robotics could eventually contribute to this field by providing alternative ways to generate movement at small scales.
For example, future research might investigate whether biological actuators could help power miniature devices designed for carefully controlled laboratory experiments or specialised medical applications.
However, the MIT swimming robot is not a medical device. It has not demonstrated safe operation inside the human body, and the study does not establish that it can perform surgery, deliver medication or navigate through blood vessels.
Those applications would require substantial additional research into biocompatibility, immune responses, control systems, sterilisation, biological maintenance and safety.
The most realistic near-term contribution is therefore likely to be fundamental research into how living tissue can generate controlled movement within engineered structures.
The Challenges of Building Robots from Living Cells
Despite the promising results, several obstacles remain before biohybrid robots can become widely practical.
1. Living tissue needs suitable conditions
Unlike a conventional motor, living muscle is not simply a mechanical component that can be stored indefinitely and switched on whenever required.
Cells need appropriate conditions to remain viable. Temperature, nutrients, oxygen availability and the surrounding chemical environment can all influence their performance.
A robot intended for long-term operation would therefore need a reliable way to maintain its biological components.
2. The robot is still slow
The current device swims at approximately four body lengths per minute during straight movement.
This is sufficient to demonstrate locomotion but limits the range of tasks it could perform.
Improving speed will require better body geometry, more effective transfer of muscle force into movement and potentially new approaches to controlling the fins.
3. Navigation remains externally controlled
The experimental robot followed a light source manually directed by a researcher.
A practical autonomous machine would need some combination of sensors, control systems and a reliable way to receive or generate movement commands.
Developing such capabilities without making the device too large or complex will be a significant engineering challenge.
4. Scaling up is difficult
A design that works at a very small scale does not necessarily translate directly into a larger robot.
Larger devices may require more force, more extensive biological tissue and different supporting structures. Smaller devices, meanwhile, can experience proportionally greater resistance from the surrounding fluid.
Engineers will need to determine which aspects of the design can be scaled and which must be redesigned.
5. Biological and engineering systems must work together
The performance of the robot depends on the interaction between living cells, the hydrogel skeleton and the surrounding water.
Changes in one component can affect the entire system. A material that supports cell growth particularly well may not necessarily provide the ideal mechanical properties for movement.
Future progress will depend on improving these components together rather than treating the biological and mechanical parts as independent systems.
What Does This Mean for the Future of Robotics?
The MIT project is part of a broader shift towards robots that are not defined solely by conventional motors, gears and electronics.
Soft robotics already explores flexible materials and structures inspired by biological organisms. Biohybrid robotics takes this approach further by incorporating living components directly into the machine.
The potential advantage is not that biological systems are always better than artificial ones. Conventional motors remain far more suitable for many applications requiring high speed, large forces, precise positioning or operation in harsh environments.
Instead, biological actuators could provide new capabilities in situations where flexibility, small size and interaction with delicate surroundings are particularly valuable.
The most promising future may involve a combination of technologies rather than the replacement of conventional robotics.
Engineers could use living muscle where its mechanical properties are advantageous while relying on electronic sensors, synthetic materials and computer control for other functions.
Such hybrid systems would introduce new design challenges, but they could also expand the range of tasks robots are able to perform.
For now, the MIT swimming robot is best understood as a research demonstration showing that a very thin layer of living skeletal muscle can generate useful movement when combined with an optimised supporting structure.
That is a significant engineering achievement, even if practical applications remain some distance away.
When Biology Becomes Part of the Machine
The development of a swimming robot powered by living muscle cells demonstrates how biology and engineering can be combined to produce movement in new ways.
By growing aligned muscle fibres on a patterned hydrogel skeleton and controlling their contractions with light, MIT researchers created a thin device capable of swimming and changing direction through water.
The study reports substantial improvements in muscle efficiency and functional longevity compared with earlier two-dimensional actuators. It also shows that the design of the supporting material can be just as important as the biological tissue itself.
The robot is not yet an autonomous underwater explorer, a medical device or a replacement for conventional machines. Its speed remains modest, its navigation depends on external control, and its living components require appropriate laboratory conditions.
Nevertheless, the research offers a valuable demonstration of what biohybrid robotics might eventually achieve.
As scientists learn to organise, stimulate and maintain living tissue more effectively, biological actuators could become useful in specialised soft robots for research, environmental monitoring and other applications that demand delicate movement.
The broader lesson is that the future of robotics may not be exclusively mechanical or biological. It may emerge from carefully engineered combinations of both.
Sources and Further Reading
The research findings described here come from a published scientific study. Potential applications in medicine and environmental monitoring are prospective possibilities, not demonstrated capabilities of the current prototype.