London startup morph launched a soft robotic cell platform that combines perception, control, and mechanical deformation in a single module. The cells change their morphology in real time — adapting shape and stiffness to contact with an object or a person. The company positions itself as a design and manufacturing partner for other robotics firms, not as a robot maker.
Key takeaways
- morph builds soft cells with embedded physical AI — intelligence is built directly into the material, not managed by external systems
- The cells are made of deformable, fluid-actuated materials simulated using high-fidelity physics-based simulation
- B2B strategy: morph sells the technology as a component to robot manufacturers rather than building a consumer end-product
- Initial use cases focus on healthcare — athletic performance support, injury prevention, and rehabilitation
- Investors include 8VC, Copper, Valia Ventures, Qubit Health Capital, and Blue Lion
Soft versus rigid: the problem morph is trying to solve
Most robotic platforms treat hardware and AI as separate layers. Sensors collect data, computational models process it, and actuators execute commands. morph inverts this logic: the control intelligence is embedded directly into the material of the cell.
Founder and CEO Dr. Jean Nehme — a former reconstructive surgeon and founder of Digital Surgery (acquired by Medtronic) — explains this through a biological analogy. Octopuses change shape and stiffness without a central processor managing each movement. morph builds on the same premise.
The cells are fluid-actuated and made from deformable materials. Historically, modeling the dynamics of such materials was computationally prohibitive. The rise in computing power and access to high-fidelity physics simulators — those supporting Google DeepMind and NVIDIA, for instance — has enabled morph to build design models, simulate them, and move to production.
Platform architecture
The morph platform is not a single product but a technology stack.
Cells collect environmental signals — contact, pressure, temperature, position — process them locally using embedded models, and respond by changing shape or stiffness. Data from deployed cells feeds back into the models: training continues after the cells enter operation in the field.
morph offers three services to partner companies: designing cells for specific applications using its proprietary design engine, manufacturing them on its own production stack, and deploying and updating control models throughout the operational lifecycle.
Soft robotics versus hard robots
The market is dominated by rigid manipulators and humanoids. Why soft robotics?
Nehme points to two reasons. First, safer human contact. A soft robot absorbs impact energy by its nature. This matters both in factories and in physiotherapy or motion-assist devices.
Second, production scalability. Deformable materials are cheaper to manufacture than precision rigid structures and easier to adapt to new form factors. The company expects per-unit cell costs to be lower than conventional actuators.
The limitation is symmetrical: soft materials are mechanically unpredictable. morph addresses this with its own simulation environment and continuous learning — the model for each deployment learns the specific operating conditions after field launch.
Market and competition
Other companies work on soft robotics — Festo with bionic cobots, Boston Dynamics on active impact damping, and several academic spinouts (Harvard Wyss Institute, MIT CSAIL) on pneumatic soft actuators. None, however, have adopted a B2B component supplier strategy with embedded AI at the scale morph is signaling.
The investment round was not disclosed in terms of amount.
Why it matters
Embedded physical AI — intelligence built into hardware rather than managed externally by software — is a direction also being pursued by NVIDIA Isaac GR00T and Google DeepMind (Robotics). morph attacks the problem from the opposite side: instead of teaching existing rigid robots better control, it changes the material the robot is made from.
If this approach proves scalable, it could reshape the architecture of robots working in close contact with people — particularly in medicine, rehabilitation, and wearable assist devices. The company has not yet released production timelines or volumes, so assessing technological maturity remains difficult. The B2B strategy, however, eliminates the need to build a consumer end-product — morph can grow alongside the companies that buy its technology.
What next?
- morph announced a focus on healthcare applications as its first commercial deployments, with no specific timeline given
- The company plans to expand the platform into automotive and industrial safety segments after healthcare
- They are opening conversations with technology partners — inquiries accepted via morph.inc
Sources
- The Robot Report — Soft, robotic cells from morph embed physical AI into hardware
- morph — official product website
