Artificial intelligence has left the screen

Until very recently, artificial intelligence was a bodiless entity. It lived trapped behind the glass of our smartphones or in the cold silence of data centres. We could interact with it, marvel at its prose or let it solve complex equations — but always under one condition: for AI, the physical world did not exist. The "world" was nothing but data.

But something has changed in 2026. We saw it at CES in January, where "Physical AI" was one of the central themes, and we are now confirming it with the recent alliances between giants like Nvidia and ABB. AI has broken through the glass. It has stopped being pure mind and started having hands, legs and, above all, presence.

From the screen to the street
For years we grew used to viral videos of robots performing somersaults or rehearsed choreographies. Spectacular, yes — but ultimately a technological magic trick: lots of hardware, very little real autonomy. These were machines following a script, not agents understanding an environment.

However, 2026 will not be remembered for robots that jump, but for robots that understand space. The key difference in the new Physical AI models is that they no longer rely solely on rigid lines of code. Robots now "wake up" in physical bodies having already trained in massive digital simulations — a thousand years of learning compressed into an afternoon.

What we see in factories today is not just advanced mechanics; it is embodied intelligence that knows how to interact with the unpredictable.

The human factor: companion or tool?
The most fascinating aspect of this transition is not happening on assembly lines, but in the intimacy of homes. The deployment of systems like ElliQ in New York to combat loneliness among the elderly holds up an uncomfortable mirror.

I recently reflected on how "simulating is not feeling", and the technical distinction still holds: these systems have no inner experience of affection. Yet when AI leaves the screen and sits on the sofa, the logical boundary blurs. If a robot can reduce a person's cortisol levels or offer assistance that feels human, the debate about its "inner life" becomes secondary to its real-world impact.

Current robotics is forcing us to redefine our relationship with technology: we are moving from using tools to coexisting with agents.

The atom and the void: two sides of the same coin
The question comes up regularly: why does space robotics occupy such a central place in this blog? The answer is that the conquest of space and the robotic revolution on Earth are, at their core, the same technological challenge.

Every time a robot learns to navigate a messy warehouse or manipulate a fragile object down here, we are refining the algorithms that will operate tomorrow in the solitude of Mars or the moons of Jupiter. Earth is our most demanding training ground; if we can get AI to master the complexity of atoms and gravity in everyday life, the leap into the void of space will be a natural extension of its capabilities.

Conclusion
Physical AI is the definitive bridge. We are leaving behind the era of digital oracles and entering the era of synthetic companions.

Perhaps the greatest challenge of this year is not perfecting humanoid motors or increasing the power of vision chips. The real challenge will be our capacity to adapt to a world where, for the first time in history, we are no longer the only intelligence walking through the room.

And when that happens, the question stops being what it can answer — and becomes what role it will play in the future we are building together.

Read original article