Humanoids Are the Low-Hanging Fruit—What Comes Next?

Humanoid robots are having their moment. Investors feel it. Founders feel it. Factories, logistics hubs, and even homes are starting to feel it too. And honestly, it makes perfect sense: the humanoid form factor is the cleanest shortcut to commercial traction. Our entire built environment is tuned to human proportions--door handles, ladders, forklifts, countertops, toolsets. Two arms, two legs, and a sensor suite at roughly five to six feet off the ground is the built-in API for the physical world.

If you want a robot you can drop into a factory tomorrow and have it be useful, you build a humanoid. That's why Figure (which raised $675M at a $2.6B valuation in 2024), Boston Dynamics (whose Atlas began commercial production in January 2026 for Hyundai deployments), Apptronik (with Apollo piloting at Mercedes-Benz facilities), Agility Robotics (whose Digit is being tested at Amazon warehouses), and hyperscaler-backed entrants are pulling in capital at unprecedented speed. This isn't sci-fi nostalgia. It's pure practicality--the path of least resistance.

But that same logic creates a subtle trap: skeuomorphism. We're shaping robots to fit the world we built for ourselves, not asking what happens when robots start reshaping that world in return. What happens when we stop optimizing for "fits in a warehouse aisle" and start optimizing for entirely new physics, architectures, and capabilities?

That's the next frontier. And it's weirder, richer, and more commercially explosive than a fleet of metal interns.

What Comes After the Humanoid Wave

The humanoid platform will eventually commoditize. Once locomotion, manipulation, and balance are good enough--and model training becomes cheap and abundant--humanoids will become appliances. Useful, ubiquitous, margin-thin.

So the next investor question becomes: what's just beyond the obvious?

The answer: robots that ignore human constraints entirely.

1. Swarm Intelligence
Imagine hundreds of cheap, self-organizing units acting as a collective--no central body, no single point of failure. Ideal for infrastructure maintenance, hazardous exploration, or assembling structures from the inside out.

Harvard's Wyss Institute has demonstrated this with their Kilobots--a self-organizing thousand-robot swarm that can form complex shapes following simple programmed rules, mimicking ant colony coordination. The work, led by Radhika Nagpal, has been published in Robotics and Autonomous Systems. Stanford's Multi-robot Systems Lab studies distributed algorithms for swarm control, while their Shape Lab developed Zooids--palm-sized bots that demonstrate emergent path formation for swarm user interfaces. Verity already commercializes indoor drone swarms for inventory tracking, with recent pilots alongside Maersk and On achieving 99.9% accuracy at 1,000 items per second. All the ingredients--onboard compute, low-cost sensing, lightweight coordination--are falling into place.

2. Soft Robotics
Soft robots ooze, stretch, squeeze, and deform. They reach places humanoids never could: rubble, pipes, air ducts, surgical pathways.

Carnegie Mellon's Robotics Institute is building soft robots for continuum manipulation and serpentine robots for bridge inspection. Their Softbotics Lab and recent work on modular robots for gas pipeline repair showcase inspection and repair applications. ETH Zurich researchers have developed vine robots that evert through bending, while Italian researchers published groundbreaking work on growing soft robots with climbing plant-inspired behaviors in Science Robotics. Here, the body becomes part of the intelligence--perfect for search-and-rescue or minimally invasive medical work.

3. Modular and Reconfigurable Systems
Imagine robots made of modules that reassemble into whatever tool the task requires: a crane, a crawler, a long inspection arm, a climbing rig.

MIT's M-Blocks are self-transforming robot cubes that jump, spin, flip, and identify each other. With no external moving parts, they use internal flywheels reaching 20,000 RPM to propel themselves, connecting via permanent magnets. NASA's work on autonomous in-space construction and MIT Media Lab's TESSERAE self-assembling space architecture point toward machines that grow, split, merge, and reshape themselves. Recent research in Nature Communications on modular shape-changing tensegrity-blocks demonstrates robots that can self-assemble into structures for emergency deployment. Early construction and mining robots are already testing these architectures for remote and hazardous sites.

4. Bio-Hybrid Machines and Organoid Compute
The boundary between robotics and biology is dissolving. Organoid computing uses living neural tissue for adaptive, energy-efficient computation--what researchers call "organoid intelligence." Companies like Cortical Labs are building the "world's first code deployable biological computer," while FinalSpark's Neuroplatform offers remote access to brain organoids for biocomputing research. Johns Hopkins researchers have found lab-grown brain organoids show building blocks for learning and memory.

On the actuation side, biohybrid robots powered by living muscle tissue are advancing rapidly. MIT recently announced artificial tendons that give muscle-powered robots a boost, while ETH Zurich and the Institute for Bioengineering of Catalonia are developing muscles and tendons for robots. These machines won't look familiar--no faces, no chassis. Just adaptive, functional entities built on biological efficiency and plasticity. For investors, this space could echo the leap from vacuum tubes to silicon.

5. Robots That Don't Look Like Robots
This is where the frontier gets playful. Tangible Robots' Eggie is a wheeled humanoid built for home environments with dexterous hands and whole-body control. InteractionLabs' Ongo is an AI-powered desk lamp robot with emotional interaction--proving social robots don't need limbs; they can be characters, companions, or interfaces.

On the industrial side, Eyecandy Robotics' Tensaur is deliberately non-humanoid because the job doesn't require it--it's "a physical character brought to life through AI & Robotics" for education and research. Many high-value tasks won't be solved by bipedal machines but by devices built exactly for the task at hand.

Prototypes like the pendulum-driven legless rolling jumping robot, UCLA's ARTEMIS bipedal robot, or KAIST's field-programmable robotic folding sheets and microscopic metasheet robots point to a future where form follows function, not familiarity.

Why This Matters for Investors

The humanoid wave will create real value, but it also unlocks the next wave. Once embodied intelligence becomes a commodity--perception, manipulation, control--the design space explodes. Costs drop. Constraints loosen. Familiarity stops being an advantage.

That's when weird robots shift from sci-fi to inevitability.

Humanoids win the present because they slot into the world as it exists. But the next iconic robotics companies will build machines that reshape the world entirely--new species of industrial, biological, and computational agents.

On a 5- to 10-year horizon, the opportunity isn't to bet against humanoids. It's to invest just beyond them.

Because once the humanoid race gets crowded, the real frontier opens: machines that swarm, flow, grow, dissolve, and reassemble--machines that aren't coworkers but something genuinely new.

And that frontier is already coming into view.