NVIDIA Unveils Alpamayo 2, Taps Taiwan Supply Chain for Robotics Push

NVIDIA is positioning physical AI as the next major computing wave after generative AI, with CEO Jensen Huang describing an autonomous vehicle as a robot on four wheels. The Alpamayo platform combines perception, reasoning, planning and action to help developers address rare, complex road scenarios and advance Level 4 robotaxis. The push matters to Taiwan because its ecosystem spans advanced chipmaking, electronics manufacturing, edge computing and precision machinery, supporting the infrastructure and hardware needed to move AI from data centers into vehicles, factories and robots.

At GTC Taipei on June 1, NVIDIA unveiled Alpamayo 2 Super, an open 32-billion-parameter vision-language-action model designed to reason across the full autonomous-driving stack. It can generate driving trajectories and chain-of-causation traces while predicting high-level actions including yielding, lane changes and stops. NVIDIA also introduced AlpaGym, an open-source reinforcement-learning framework for closed-loop training. Taiwanese companies including TSMC, Hon Hai Precision Industry, Quanta Computer, Wistron and Inventec span the chip, systems-manufacturing and edge-hardware supply chain supporting NVIDIA’s physical-AI expansion.