At Computex Taipei 2025, Nvidia announced several major developments, including progress in humanoid robotics with its Isaac GR00T-Dreams platform, which helps generate training data for robots. The company also introduced NVLink Fusion, enabling customers to create custom AI servers and data center infrastructure using both Nvidia and third-party components, like Qualcomm’s CPUs. Nvidia’s new RTX Pro Blackwell servers, running on its advanced Blackwell GPUs, are aimed at making enterprise workloads more efficient by transitioning from CPU-based to GPU-accelerated systems. These moves come as Nvidia benefits from the recent removal of US AI chip export restrictions and as it expands global partnerships, including new deals with Saudi-backed startups.
Related articles:
Computex 2024 highlights: Emerging AI hardware trends
Qualcomm expands into server market with new AI-focused CPUs
Nvidia in talks with Asian startups on AI partnerships
AI data center energy consumption and the rise of GPU acceleration
