AI is leaving the screen. Tokyo spent two days with it.

This week the Humanoids Summit came to Tokyo. Two days at the Takanawa Convention Center. Toyota, Honda, Disney Research, Google DeepMind, Boston Dynamics, McKinsey, Nvidia, and Japan's METI in the same rooms. Brittany Arthur, our CEO, was the MC. She opened the summit and gave the closing remarks both days.

So we had a seat at the front for all of it.

A few things kept surfacing. Hands, again and again. Touch. Whether people will accept these machines in a shared space, and whether they hold up once they leave a controlled floor. Capability was everywhere in that room. Most of the talking was about everything around it.

In our Future Signals 2026 report, we called this Signal 4. Beyond the Screen. The screen is no longer the boundary of what AI can affect. For years AI read text and made images and stayed behind glass. That boundary is dissolving, and Tokyo was two days of watching it dissolve in person.

The report is about the architecture of intent. When AI can execute almost anything, the scarce thing is knowing what is worth executing. The hand is solvable. The reason to build the hand is a human question, and it stays one no matter how good the hand gets.

AI is getting a physical vocabulary. Tokyo made that plain. The work that matters next is deciding what we point it at.