Jennifer Lewis is pushing 3D printing far beyond making rigid objects layer by layer. Her Harvard research group is developing printing methods that control the composition and internal architecture of soft materials, allowing researchers to fabricate structures that can bend like artificial muscles, operate as soft robots and, at the biological end of the spectrum, contain living human cells and perfusable channels resembling parts of organs.
Lewis outlined that vision at MIT’s 2025 Mildred S. Dresselhaus Lecture, where she presented work spanning soft electronics, robotic matter and vascularized human tissues. The connection to MIT is significant but needs to be stated precisely: Lewis earned her ScD in Ceramic Science from MIT in 1991 and returned as the 2025 Dresselhaus Lecturer, while her current research program is based at Harvard University, where she is the Hansjörg Wyss Professor of Biologically Inspired Engineering.
Several developments since the lecture have made that research direction more concrete. In 2026, Lewis and collaborators reported new approaches to printing programmable soft robotic structures and artificial-muscle-like materials, while another project advanced the fabrication of human kidney collecting-duct systems. Together, the projects illustrate a much bigger idea: additive manufacturing can be used to program how matter behaves, not simply what shape it takes.
Lewis’s MIT Lecture Connected Robots, Electronics and Human Tissue
The unifying idea behind Lewis’s research is surprisingly straightforward. The way a material is manufactured affects its structure, and that structure affects its properties.
Conventional 3D printing is often understood primarily as a geometric technology. A digital design is divided into layers and a printer deposits material until a physical object matching that geometry emerges. Lewis’s research asks what becomes possible when the printer also controls which materials are deposited, where they are placed and how their microscopic structures are oriented.
That shift turns a printer into something closer to a materials-programming system. Different inks can provide electrical conductivity, flexibility, structural strength or biological function, while sophisticated printheads determine how those materials are arranged inside the finished object.
Lewis’s lab organizes much of its work around three broad areas: functional matter, structural matter and living matter. Applications range from soft sensors and electronics to architected materials, robotics and vascularized human tissues.
Rotational Printing Gives Soft Robots Built-In Motion
One of the clearest demonstrations came from research published in 2026 on rotational multimaterial 3D printing.
The method allows multiple materials to emerge from a single printhead while the nozzle rotates during extrusion. By controlling nozzle geometry, rotational speed and material flow, researchers can manipulate how different materials are arranged within each printed filament.
For soft robotics, Lewis’s group used this approach to create flexible filaments containing internal pneumatic channels. The outer material was polyurethane, while a removable gel-like material formed the internal channel. After printing and solidification, the temporary material could be removed, leaving an air pathway embedded inside the structure.
The location of that channel matters. If pressure is applied through a deliberately asymmetric internal channel, one side of the structure responds differently from the other, causing the printed object to bend in a predetermined direction.
That means part of the robot’s movement can effectively be encoded during fabrication.
A Soft Robotic Hand Can Be Printed With Its Actuation Architecture
Researchers demonstrated the technique using structures including a soft hand. Instead of manufacturing individual rigid components, assembling joints and then installing separate actuators, the printing process can create the flexible body and its pneumatic architecture as part of the same fabrication strategy.
That represents an important difference between soft and conventional robotics.
Industrial robots traditionally depend on rigid links, motors, bearings and mechanical joints. Soft robots use deformable materials and can move by changing the shape of their bodies, potentially making them useful where machines must interact safely with people or delicate objects.
Healthcare is one possible application because compliant robots could eventually perform tasks in environments where rigid machinery is unsuitable. Manufacturing and wearable systems are other possibilities.
The research remains a fabrication platform rather than evidence that general-purpose soft robots are ready to replace conventional machines. Its importance lies in giving engineers finer control over how flexible robotic structures are designed and manufactured.
Artificial Muscles Take the Same Idea Further
A related 2026 project from Lewis’s lab applied rotational printing to materials designed to behave more like biological muscles.
Natural systems frequently create complex movement from organized fibers. Muscle bundles contract in coordinated ways, plant structures bend and twist, and biological filaments can generate complicated motion from relatively simple local mechanisms.
The Harvard researchers developed printed filaments whose internal material arrangement could be programmed so that heating or cooling caused them to bend, coil, twist, expand or contract.
Instead of attaching a separate motor to create each motion, the response is encoded into the material architecture itself.
This is why the term “artificial muscle” is useful but should not be interpreted as meaning the researchers have printed a biological human muscle. These are engineered, responsive structures that reproduce aspects of muscle-like movement.
The broader objective is programmable matter: designing a material so that its geometry and composition determine how it will move when exposed to a stimulus.
The Same Manufacturing Philosophy Extends to Living Cells
Moving from robotic matter to biological tissue sounds like a radical transition, but the manufacturing challenge has an important similarity.
A functioning tissue is not merely a collection of cells. Different cell types must occupy appropriate locations, extracellular materials provide structural environments, and nutrients and oxygen must reach cells throughout the tissue. That requires control over architecture at multiple scales.
Lewis’s group develops biological “inks,” or bioinks, containing materials and cells that can be deposited into carefully designed three-dimensional arrangements. The laboratory combines these printing techniques with stem-cell biology, organoids and bioreactors to create tissue models for drug testing, disease research and, over the longer term, therapeutic applications. A central challenge is vascularization.
Cells buried inside a thick engineered tissue cannot survive indefinitely without adequate access to oxygen and nutrients. Building channels through which fluid can flow is therefore one of the critical requirements for scaling engineered tissue beyond very small structures.
Vascularization Is a Major Barrier to Printing Larger Tissues
Lewis and collaborators have spent years developing methods for embedding perfusable channels within densely cellular tissues.
“Perfusable” means that fluid can be passed through the structure. In engineered tissues, those channels can help mimic part of the transport function performed by blood vessels in the body, supplying nutrients and supporting cells deeper inside a tissue construct.
This is substantially more complicated than printing a shape resembling an organ.
Human organs contain highly organized networks of cells, vessels, ducts, nerves and extracellular structures. Their biological functions emerge from interactions among those systems. Reproducing the external geometry of a kidney or heart therefore does not mean researchers have reproduced a functioning organ.
The distinction is essential when discussing bioprinting. Lewis’s work represents meaningful progress toward more realistic human tissue systems, but whole transplantable organs remain a much more difficult objective.
Kidney Engineering Shows How Specific the Challenge Has Become
A 2026 project led by Lewis illustrates the level of biological complexity researchers are now tackling.
The kidney contains more than a million microscopic filtering units called nephrons, while an extensively branched collecting-duct system helps transport urine toward the ureter and regulates water, salts and acidity. Recreating that plumbing architecture is necessary if researchers ultimately want to engineer more complete kidney tissue.
Lewis’s team developed a platform combining human induced pluripotent stem cells, organoid biology, extracellular matrices and bioengineering techniques to fabricate collecting-duct structures at different scales.
The researchers created smaller tubular networks alongside larger perfusable structures and demonstrated connections between them. The work provides a way to study collecting-duct biology and could support disease modeling and drug testing while contributing to the much longer-term challenge of organ engineering.
The Wyss Institute’s description of the kidney work explicitly frames whole-organ biofabrication as an ultimate goal rather than a capability that has already been achieved.
That distinction prevents an important scientific advance from being overstated.
Printed Tissues Could Have Value Long Before Transplantable Organs
The commercial and medical value of bioprinting does not depend entirely on reaching the point where researchers can manufacture complete replacement organs.
Engineered human tissues can potentially provide more realistic systems for studying diseases and testing drug candidates. Researchers currently rely on combinations of conventional cell cultures, animal models, organoids and clinical studies, each of which has limitations.
A three-dimensional tissue containing relevant human cell types, extracellular materials and perfusable channels can reproduce aspects of human biology that a flat cell culture cannot.
That could help researchers investigate how tissues respond to drugs, genetic changes or disease processes under more physiologically relevant conditions.
Patient-specific cells create another possibility. In principle, induced pluripotent stem cells derived from an individual could contribute to engineered tissues designed around that person’s biology.
That is the context behind the question Lewis posed during her MIT lecture: can tissues eventually be made “from you, for you?”
Soft Robots and Living Tissues Share a Manufacturing Problem
The most interesting part of Lewis’s research may be the connection between projects that initially appear unrelated.
A soft robotic hand and a vascularized tissue do not perform the same function, but both require precise control over multiple materials inside complex three-dimensional structures.
For a robot, the printer may need to position flexible polymers and pneumatic channels so the structure bends predictably. For engineered tissue, it may need to arrange biological materials, living cells and perfusable channels so cells survive and organize appropriately.
In both cases, geometry alone is insufficient. The local composition and architecture determine function.
This is why multimaterial printing matters. Rather than printing an object and adding its functional systems afterward, researchers can increasingly incorporate some of those functions during fabrication.
That approach could eventually reduce the boundary between manufacturing a structure and programming its behavior.
3D Printing Is Becoming a Materials Platform
Lewis’s work also demonstrates how far additive manufacturing has moved from its early association with rapid prototyping.
The field is increasingly concerned with controlling matter across multiple length scales. Printhead design, ink chemistry, fluid mechanics and digital toolpaths can be combined to determine not only an object’s shape but its electrical, mechanical or biological characteristics.
For soft robotics, that could enable devices with integrated actuation and sensing. For medicine, it could produce tissue models with increasingly realistic biological organization. For electronics, multimaterial printing can embed conductive and functional elements directly into devices.
These applications require very different materials, but they share a manufacturing philosophy: function can be created by controlling material placement.
Lewis’s career has helped connect those disciplines because her research combines materials science, fluid mechanics, bioengineering and additive manufacturing rather than treating 3D printing as an isolated technology.
The Biggest Goal—Printed Organs—Remains a Long-Term Challenge
The possibility of manufacturing replacement human organs understandably attracts the most attention, but it is also where careful language matters most.
Researchers can already bioprint increasingly sophisticated tissue constructs and organ-specific structures. Lewis’s group has developed vascularized tissues and kidney-related systems, while other laboratories worldwide are working on complementary problems involving cells, scaffolds, blood vessels, maturation and transplantation.
A complete human organ is much harder.
It must contain the appropriate cell types and architecture, establish functional vascular networks, integrate with the recipient’s body and continue performing complex biological functions over long periods. Manufacturing a structure with the correct external shape solves only a small part of that problem.
Lewis’s current research is better understood as building enabling technologies toward that objective.
Progress in perfusable tissues, organ building blocks and engineered duct networks can have nearer-term applications while researchers continue tackling the harder requirements for therapeutic organs.
What Comes After the MIT Lecture
Lewis’s 2025 Dresselhaus Lecture provided a snapshot of a research program spanning an unusually wide range of materials. The developments that followed in 2026 show that the common thread is becoming clearer.
Rotational multimaterial printing can create soft robotic structures whose movement is encoded in their internal architecture. Related methods can generate artificial-muscle-like filaments that bend and twist in programmed ways. At the biological frontier, bioprinting and organoid engineering are producing increasingly sophisticated perfusable human tissue structures, including components of the kidney’s collecting-duct system.
These advances do not mean fully autonomous soft robots or transplant-ready printed kidneys are around the corner. Both fields still face substantial engineering and biological barriers.
What they demonstrate is a change in what 3D printing can mean.
The printer is no longer being used only to reproduce a digital shape. By controlling where different materials, channels and cells are placed, researchers can begin to encode behavior and function directly into manufactured structures.
For Lewis’s work, that principle connects everything from a bending robotic finger to living human tissue—and it may ultimately prove more consequential than 3D printing any single object.
