Physical AI for Materials Intelligence

Physical AI (Robotics · Digital Twin · Recipe Language Model)
for Materials Intelligence

Launch Exhibition Mode Robotics MIRecipe

Materials Intelligence

Open GitHub

We introduce a domain-specific RLM developed through seven AI layers and interconnected robotic boxes to drive the evolution of physical AI for defining a new paradigm - Materials Intelligence.

Materials Intelligence

Materials Intelligence Lab Overview

Materials Intelligence Lab Overview

Physical AI System

Physical AI System architecture overview
01Robotics
02Digital Twin
03MIRecipe
04RLM + MIOS

MIFactory: for Materials Intelligence and Manufacturing

RB

Robotics
DT

Digital Twin
MIR

MIRecipe
RLM

Recipe Language Model
OS

MIOS
XR

Media Wall

Physical AI Workflow

Open Architecture
Literature Mining
MIRecipe
Recipe Language Model
MIOS + Digital Twin
Robotic Execution

Key Publications

Open Publications
Material Intelligence Framework cover
Review · Nexus, 2025

Material Intelligence Framework

Reading-Doing-ThinkingAI + roboticsClosed-loop discovery
Robotic Nanocrystal Synthesis cover
Article · Nature Synthesis, 2023

Robotic Nanocrystal Synthesis

Data miningRobotic synthesisInverse design
Agentic Robotic Boxes + RLM cover
Article · Engineering, 2026

Agentic Robotic Boxes + RLM

11 robotic boxes50,764 PSCs578M+ tokens

MIRecipe

Open MIRecipe
Reading Literature
60000+papers
Screened Literature
40000+papers
Recipe Resource
20000+files
Recipe Report
12000+reports
Recipe Tokens
578M+tokens
Reading Literature
60000+ papers
Screened Literature
40000+ papers
Recipe Resource
20000+ files
Recipe Report
12000+ reports
Recipe Tokens
578M+ tokens