Vision-Language-Action
VLA
Vision-Language-Action (VLA) is an AI architecture that fuses visual perception, language understanding, and action execution into an end-to-end system for embodied intelligence — essentially enabling robots to "see, understand instructions, and act immediately" . As described in a Frees Report piece, it extends traditional multimodal large models into physical action capabilities, mapping multimodal inputs directly to motor outputs for tasks like robotic grasping or warehouse logistics .
The architecture has become a foundational framework in embodied AI, though with recognized limitations: mainstream VLA models rely heavily on trajectory memorization, which can fail on abstract concepts, environmental generalization, and long-horizon tasks . This has spurred architectural variants — notably AgiBot's ViLLA (Vision-Language-Latent-Action), which introduces latent action tokens to bridge the semantic gap between visual-linguistic inputs and physical execution , and Astribot's CLAP framework, which aligns human video demonstrations with robot trajectories through cross-modal contrastive learning . Recent research directions also explore integrating VLA with world models for unified perception-planning-action loops .
AI-generated — may contain errors, please verify.
Coverage
Which Path Leads to the "World Model" Endgame? | A Conversation with Biwei Huang, Founder of Aether AI
It's 2026, and we're still evaluating World Models the way Edison tested filaments.
A Serious Discussion on World Models
Billions of years ago, living organisms built the first "world model."
How Far Are We From a World Model, Now That We've "Uploaded" a Fruit Fly Brain?
The "OpenAI Moment" for Pharma Labs: HeTan AI Raises 50 Million Yuan, Robot Scientists Step Onto the Lab Bench
The story of robot scientists has only just begun.
DeepRoute's Real-World Road Test: When AI Learns to "Fear," the "Black Box" of Assisted Driving Is Opened | Yunqi Capital
When Cars Begin to Understand the World
Yunqi Capital Quarterly | Upward, the Consistent Answer
New Growth, New Gains
Wang Qian, Invariant Robot: How Far Is Embodied Intelligence's Scaling Law? | Yunqi Capital Doers Series
Embodied Intelligence ≠ Stuffing DeepSeek into a Unitree Robot
Yunqi Capital | Yuanrong Qixing Surpasses 30,000 Units in Mass Production Deliveries for September, Setting Another Record
How VLAs Are Pioneering a New Future for Driving
AI-Powered, Accelerating Evolution! Five BlueRun Ventures Portfolio Companies Named to ChinaVenture's "Sharp 100" List
Show sharp edge, charge ahead all the way — or more naturally: **Sharp and unstoppable, full speed ahead.**
