Zihao Li·Founding Researcher, ZENO AI·September 2026

Zeno-1: Collaborative Intelligence for Robots That Work Together

Zeno-1 is the collaborative physical intelligence architecture I proposed and now lead at ZENO AI — built on a simple premise: a second robot should expand what the first robot can do, not just add a second pair of hands.

A second robot changes the problem

Put two capable robots on the same physical task and capability does not simply add. Timing, contact, and motion from one robot change what the other can do next. Most multi-robot systems route around this by scheduling: a planner assigns roles up front and each robot executes its slice. Zeno-1 takes the harder path — robots read each other's state directly, moment to moment, and decide what to do next from what their partner has already started.

That premise runs through the research I publish under my own name — motion planning around human contact, memory that survives long horizons, teleoperation that shares a language for force and reach — and it is also the architecture ZENO AI has built as its core product: Zeno-1.

Four ideas, in practice

The clips below are everyday household tasks, each isolating one coordination principle Zeno-1 relies on.

SHARED PASSAGEOne robot's path is the other's constraint, read live instead of scheduled around.

SHARED LOADLet go too early and the task restarts. The tension itself is the coordination signal.

ROLES, NOT ASSIGNMENTSNeither robot is told to scoop or to hold. The role is inferred from what the other has already started.

SEQUENCE FROM STATEThe bag can't be sealed while a partner is still filling it — Zeno-1 waits on the world, not on a message.

The interaction is the unit of intelligence.

None of the four clips above required a shared script or a fixed role assignment. Each robot commits to an action, watches what its partner does in response, and adjusts — the same principle behind the sheet-folding, bed-making, and long-horizon runs elsewhere on this site, just scaled to ZENO AI's product line.

This page is my own introduction to the project. For the complete write-up — system architecture, training methodology, and evaluation results — read the official publication on ZENO AI.

Read the official technical report ↗