Not content with folding laundry, making coffee or being occasionally kicked over by a YouTuber for clicks, Figure’s humanoid robot has now climbed a ladder.

Autonomously.

The video, posted by Figure founder Brett Adcock, shows Figure 03 gripping the rails, working out its feet and steadily climbing upwards using the company’s Helix AI system. No operator. No visible joystick. No helpful intern just outside shot with a very expensive remote control.

Just a robot going up a ladder.

And yes, it is a short clip. We do not have a published success rate, an explanation of how many takes it needed, or a detailed technical paper to tell us what happened when it reached the top. Sensible caveats all round.

But it is still one of those moments.

Because ladders are awkward. They require balance, spatial judgement, grip, foot placement and the ability to recover when something is slightly off. Humans mostly manage them because we have had years of practice, a healthy fear of falling and, in Britain, a cultural attachment to asking someone else to hold the bottom.

Robots have historically struggled with all of that.

Figure 03 does not look like it is auditioning for Mission: Impossible just yet. But it does look like a machine beginning to understand the physical world in the way we do.

And for property, the implications are not hard to see.

Today, a surveyor climbs into the loft. Tomorrow, perhaps a robot does. It checks the roof space, takes measurements, scans for damp, photographs insulation and flags the bit of wiring that has been “temporarily” held together with electrical tape since 1997.

Guttering. Roofs. Void spaces. Commercial inspections. Construction sites. Maintenance. The awkward, repetitive, potentially risky jobs that nobody particularly wants to do, but somebody absolutely has to.

That is where this gets interesting.

AI has spent the last couple of years living on our screens, writing emails, generating images and making some very confident mistakes about the contents of a leasehold pack. Now it is learning to operate in the actual world.

One rung at a time.