<p>Figure’s Helix humanoid robot is shown folding towels autonomously from a mixed pile in a video published on August 12, 2025. Figure presents the clip as the first demonstration of a humanoid robot with multifunction hands folding laundry entirely autonomously. The result is a concrete manipulation demonstration, but it is not a report of household or industrial deployment.</p><h2>Helix adapts to the pile</h2><p>Helix is Figure’s Vision Language Action model. In the video, it selects towels from a mixed pile and changes its folding strategy according to each item’s starting configuration. When it picks up more than one article, the robot can recover by returning the extra pieces before continuing.</p><p>The sequence includes several precise movements: tracking an edge with its thumb, pinching corners, and untangling towels before folding them. Those steps are significant because soft objects change shape and position as they are handled. In the filmed task, Helix is therefore responding to the arrangement in front of it rather than visibly repeating one fixed sequence.</p><h2>Figure says the same architecture moved from logistics to laundry</h2><p>According to <a href="https://www.figure.ai/news/helix-learns-to-fold-laundry">Figure’s account of the demonstration</a>, Figure applied the same Helix architecture used for logistics tasks directly to laundry folding. The company says the application required no change to the model and no change to its training hyperparameters. The only addition identified was a dedicated dataset.</p><p>That transfer is the broader point of the announcement. Figure is presenting one general architecture and one physical platform as capable of moving from industrial logistics to a household chore. In this example, the company’s approach suggests that task-specific data can add new behavior without requiring a separate model redesign.</p><h2>A filmed capability is not a deployment</h2><p>The distinction matters for anyone following humanoid robotics. The video shows autonomous towel folding under the conditions captured in the demonstration. It does not establish that Helix has been deployed in a home, installed in a production environment, or made available for routine use.</p><p>The material also does not provide a measured reliability rate, operating duration, or safety validation for deployment. A successful sequence can show that a robot performs a task; it cannot by itself show how consistently that performance holds over repeated operation, different environments, or unsupervised use.</p><h2>Why the demonstration matters</h2><p>Helix operates without explicit object-level representations, processing visual and language inputs directly to produce precise motor control. That design is central to Figure’s interpretation of the laundry clip: the system uses perception and language to select movements while adapting to the towels’ initial arrangement.</p><p>The practical signal is the combination of dexterity and recovery in one task. The robot is not only folding individual towels; it is also shown correcting a multiple-item pickup and dealing with tangled fabric. The practical limit is just as important: the clip demonstrates behavior, while deployment would require additional evidence about repeatability, safety, supervision, and performance outside the filmed conditions.</p><p>Figure’s Helix laundry video is a meaningful demonstration of flexible humanoid manipulation, particularly because the company says it reused its logistics architecture without changing the model or training hyperparameters. The claim should remain at the level shown: autonomous towel folding in a published demonstration. Until operating and safety evidence accompanies that capability, the video marks progress in robot behavior rather than proof that Helix is ready for dependable household or industrial service.</p><section class="media-fleet-sources"><h2>Official sources</h2><ul><li><a href="https://www.figure.ai/news/helix-learns-to-fold-laundry">Official source: figure.ai</a></li></ul></section><aside class="media-fleet-related"><h2>Related reading</h2><ul><li><a href="https://rentbuyrobot.com/article/hirebotics-smaller-auto-suppliers-welding-cobots">Hirebotics Smaller Auto Suppliers Welding Cobots</a></li><li><a href="https://rentbuyrobot.com/article/omron-sydney-poc-centre-industrial-robot-validation">Omron Sydney Poc Centre Industrial Robot Validation</a></li></ul></aside>