An octopus-inspired vision for physical AI
Image: morph
London-based startup morph has emerged from stealth with a soft robotics platform inspired by the adaptability of octopuses, positioning itself at the intersection of artificial intelligence (AI), robotics and biomimicry. The company unveiled its technology on June 2nd, 2026, after developing it privately while attracting backing from investors including 8VC, the venture capital firm co-founded by Palantir co-founder Joe Lonsdale.
morph says its platform is designed to address one of the biggest challenges facing artificial intelligence today: enabling AI systems to interact with the physical world. Rather than building humanoid robots, the company is developing flexible robotic components that can be integrated into products designed to sense, adapt and respond to changing conditions. Its platform is built around what it calls "soft robotic cells", programmable, deformable building blocks that can be simulated, configured and integrated into products across multiple industries.
The startup enters a robotics market that is attracting increasing attention from investors and technology companies as advances in artificial intelligence begin to converge with developments in sensors, simulation tools and robotics hardware.
Looking beyond rigid robots
The robotics industry has traditionally been dominated by rigid mechanical systems built from metal components, motors and fixed structures. These machines excel in controlled environments such as factories, where tasks are predictable and repeatable.
The real world presents a more complicated challenge.
Humans routinely interact with fragile, irregular and constantly changing objects and environments. Recreating that adaptability in machines has proven far more difficult than automating repetitive industrial tasks.
This challenge has fuelled growing interest in soft robotics, a field that draws inspiration from biological systems rather than traditional engineering approaches.
morph's approach also reflects a broader engineering discipline known as biomimicry, sometimes referred to as biomimetics. The practice involves studying and emulating solutions that have evolved in nature over millions of years. Engineers have borrowed ideas from nature for decades, from the kingfisher-inspired nose of Japan's Shinkansen bullet train to gecko-inspired adhesive materials. In robotics, researchers increasingly look to animals such as octopuses, jellyfish and insects for insights into movement, adaptability and efficiency that are difficult to replicate using conventional mechanical systems.
The octopus has become one of the most influential biological models in soft robotics research. Unlike animals that rely on rigid skeletons, octopuses use flexible limbs capable of adapting to highly variable environments. Their combination of dexterity, strength and adaptability has made them a recurring source of inspiration for researchers seeking to build more versatile robotic systems.
morph's technology sits firmly within this emerging discipline.
A founder with a track record in deeptech
morph was founded by Jean Nehme, a former reconstructive surgeon whose previous startup, Digital Surgery, was acquired by medical technology giant Medtronic in 2020.
Digital Surgery combined artificial intelligence, simulation technology and data analytics to improve surgical training and planning. The acquisition gave Nehme experience in building and commercialising complex technology platforms within highly regulated industries.
That background helps explain morph's emphasis on human-centred applications rather than industrial automation alone.
According to the company, its initial focus areas include human performance, mobility and injury prevention. Future applications could extend into healthcare, automotive systems and industrial safety.
This distinguishes morph from many robotics startups currently focused on building humanoid robots capable of replacing human labour in industrial settings. Instead, the company appears to be developing adaptive technologies designed to work alongside people or enhance existing products and systems.
Physical AI moves into the real world
morph's emergence comes as technology companies increasingly focus on what has become known as physical AI.
The term refers to artificial intelligence systems that interact directly with their surroundings through sensing, movement and autonomous decision-making. Unlike software-only AI applications, physical AI combines machine learning, robotics, sensors and hardware.
The concept has gained momentum as advances in generative AI have highlighted the limitations of systems that exist only in digital environments.
While large language models can generate text, analyse information and answer questions, they cannot directly manipulate physical objects or navigate real-world environments. Bridging that gap has become one of the industry's major areas of research and investment.
So how does this actually work?
The company says it combines reinforcement learning, a form of machine learning in which systems improve through repeated trial and error, with physics-based simulation. This allows robotic behaviours to be trained and tested in virtual environments before being deployed in physical products.
The use of simulation reflects a growing trend across robotics development. Virtual testing environments can accelerate development cycles, reduce costs and improve safety before technologies reach the real world.
However, translating successful simulations into reliable physical systems remains one of robotics' most persistent challenges.
As with many companies emerging from stealth, morph has presented an ambitious technological vision but revealed relatively few details about commercial deployment. Detailed performance metrics, customer deployments and large-scale commercial partnerships have not yet been disclosed publicly.
Also, soft robotics in general faces additional challenges. Flexible systems can offer significant advantages in adaptability and safety, but they can also introduce complexity in manufacturing, durability and system control. Despite substantial academic research over the past decade, relatively few soft robotics companies have successfully achieved large-scale commercial deployment.
morph's progress will therefore be closely watched, both as a company and as a test case for the commercial viability of soft robotics more broadly.
For now, morph's emergence from stealth offers a glimpse into how the next generation of physical AI companies may look beyond traditional engineering and towards solutions that have already been tested by nature itself.
Further reading on MoveTheNeedle.news:
University of Twente researchers develop electronics-free soft robot for stomach diagnostics
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.