Basket

Your basket is empty

Find industrial arms, cobots, humanoids and parts from sellers worldwide.

Browse robots
← All posts

Robotics News · Robotics · Technology

Figure 02 Just Folded Laundry – And It’s a Big Deal for Humanoid Robotics

Figure AI released a video that could mark a turning point in humanoid robotics. The company’s flagship humanoid, Figure 02, was shown folding laundry

By Robotbids · · 3 min read

Today, 12th August 2025, Figure AI released a video that could mark a turning point in humanoid robotics. The company’s flagship humanoid, Figure 02, was shown folding laundry autonomously — a task that, while mundane for humans, is notoriously difficult for robots.

Founder and CEO Brett Adcock summed it up in one sentence:

"For the first time, a humanoid robot can fold laundry using a neural net. We made no changes to the Helix architecture, only new data. The same Helix architecture that solved logistic tasks was applied directly to laundry folding. There were no modifications to the model or training hyperparameters — the only addition was the dataset."


From Logistics to Laundry — With the Same AI Brain

The breakthrough centers on Helix, Figure’s Vision Language Action (VLA) model. Helix was previously seen demonstrating an hour of fully autonomous package reorientation in a logistics setting. Now, that same model is tackling an entirely different domain — folding towels — without a single tweak to its architecture.

According to Figure’s official explanation, Helix learned to:

  • Pick towels from a mixed pile

  • Adjust folding strategies depending on how the towel started out

  • Recover from mistakes, like accidentally grabbing two towels at once

  • Use fine manipulation skills such as tracing an edge with a thumb, pinching corners, or untangling fabric before folding

And it does all of this without explicit object-level representations — a critical point. For deformable objects like towels, traditional rigid object models fail because shapes are unpredictable and constantly shifting. Helix instead runs end-to-end, translating vision and language inputs directly into smooth, precise motor actions.


Why Folding Laundry Is a Robotics Nightmare

Folding laundry might sound like child’s play, but in robotics, it’s the equivalent of a tightrope walk in a windstorm. Towels are soft, unpredictable, and unstable:

  • They bend and wrinkle in countless ways

  • They have no fixed geometry or “perfect” grasp point

  • Even small finger slips can ruin a fold

  • Success demands real-time adaptation and coordinated finger control

In human terms, it’s easy. For robots? It’s a problem that blends computer vision, tactile sensing, dexterity, and adaptive planning into one challenge.


Key Results

  • First-of-its-kind: This is the first instance of a humanoid robot with multi-fingered hands folding laundry fully autonomously using an end-to-end neural network.

  • No architecture changes: The same Helix model that mastered logistics was applied directly to laundry folding — no model modifications, no hyperparameter tuning.

  • Data-driven learning: The only change was providing the right dataset.

  • Natural multimodal interaction: While folding, Helix also maintained eye contact, directed its gaze, and used learned hand gestures when engaging with people.


The Bigger Picture

The Figure 02 laundry demo isn’t just a neat party trick. It’s an important proof of generalization in robotics — showing that one architecture can transfer from one domain to a completely different one without retraining from scratch.

If Helix can move from warehouses to living rooms this seamlessly, the implications for general-purpose humanoids are massive. Instead of building and fine-tuning a different AI brain for every single task, we might be heading toward a future where one core model can learn anything — from picking parcels to making your bed.

Today, Figure 02 folding a towel might seem like a small win. But in hindsight, it may be remembered as the moment robots started mastering the messy, unpredictable tasks of human life.