Humanoid labour.
At work today.
We’re solving Europe’s physical labor bottleneck with a humanoid workforce factories can hire by the hour.
The next step for AI
is on the factory floor.

€25–€45 / hour
Humanoid capacity for your factory. Let’s find the right fit for your tasks.
Talk to salesPut a humanoid to work.
See what the Nucleus AI model can do today— turning perception into action on the factory floor.
We learn and act
in the real world.
Reliable AI has no shortcuts.Explore Nucleus AI
Vision & depth
Spatial signals for embodied perception.
Camera frames and depth measurements capture geometry, object positions and changes in the scene. Paired observations across time provide a signal for self-supervised representation learning—helping a model connect what it sees with the physical world.
- Visual observations
- Depth measurements
- Self-supervised learning
Motion & contact
From sensor state to action trajectories.
Arm and gripper positions, motion trajectories and contact pressure describe how an action unfolds. Linking these signals to visual observations creates state–action sequences for policy post-training, connecting perception with controlled physical interaction.
- Proprioception
- Contact signals
- Policy post-training
Human guidance
Corrections as a learning signal.
Task prompts, expert instructions and teleoperation corrections capture human intent and where a policy needs support. Recording the robot’s observations alongside corrected actions creates targeted examples for imitation learning and post-training, with context around the moments that matter.
- Task prompts & interventions
- Corrected actions
- Imitation learning
Task outcomes
Outcomes that close the learning loop.
Task results, placement checks and error events connect an action sequence to its outcome. These records support evaluation and failure analysis, and can inform reward signals for reinforcement learning—turning operational feedback into a path for policy improvement.
- Outcome labels
- Failure analysis
- Reward signals
Turning experience
into physical intelligence.
Our humanoids generate multimodal data through real work: visual observations, motion and contact signals, expert corrections and task outcomes. Together, these signals connect robot experience to the data pipeline for model evaluation and post-training.How our AI works
Gets the job done.
From start to finish.
Intralogistics is a chain of coordinated movements. Inside an industrial parts store, Nucleus connects loading, storage transfers and material staging into a continuous workflow. Watch the complete recording, from the first container to the next job.
Set the material flow in motion.
Intralogistics starts with a coordinated sequence: grasp, lift, transfer and place. Nucleus loads parts containers onto a cart, connecting individual handling movements into a repeatable transport workflow.
Bridge transport and storage.
From the cart to the shelf, each transfer demands controlled movement and precise placement. Nucleus handles the physical handoff between transport and storage, one container after another.
Stage the next operation.
Nucleus retrieves parts containers from shelving and consolidates them on a cart. A sequence of retrievals becomes a staged material flow, ready for the next operation.
The first OS for
humanoid labour.
One workspace to manage your humanoid workforce. From the first check-in to the last task, keep your robots, operators and daily operations connected.


Every shift, accounted for.
Guide check-in and check-out with structured checklists, photo evidence and operator sign-off.
Check-in & check-out
Know your robot’s state.
See connection status, battery levels and available robot telemetry alongside the shift record.
Robot data
A view of the work.
Open a connected robot’s live camera feed and follow what is happening on the factory floor.
Live video
Digital E-Stop.
Request a stop for a connected robot or the managed fleet, with deliberate confirmation and event logging.
Digital stop control
A stronger Europe.
A better working life.
We’re a team of builders working to give European industry more capacity—and people more room to grow.
Accelerate Europe’s growth.
Our ambition is to accelerate GDP growth by helping factories do more. Humanoid workers add capacity where physical labor is in short supply.
From doing the task to leading the robots.
We want workers to become robot managers: teaching new tasks, guiding a team of robots and taking responsibility for the results. Less repetitive physical work. More judgment, skill and ownership.