At the GTC 2025 conference, NVIDIA CEO Jensen Huang unveiled key innovations set to shape the future of artificial intelligence (AI) and computing.
New Chip Announcements:
- Blackwell Ultra: The next generation of graphics processing units (GPUs), scheduled for release in the second half of 2025.
- Vera Rubin: Named after renowned astronomer Vera Rubin, this chip is slated for a 2026 release, offering much higher performance of the current Blackwell chip.
- Vera Rubin Ultra: Planned for 2027, this chip will further enhance capabilities and energy requirements.
Advancements in Robotics:
- Isaac GR00T N1: An open-source model designed for the development of humanoid robots, enabling faster and more efficient learning and adaptation to various tasks.
- Cosmos AI Model: An updated model that facilitates the generation of synthetic data for robot training, reducing the costs and time associated with collecting real-world data.
Partnerships and Infrastructure:
- Collaboration with General Motors (GM): NVIDIA and GM are partnering to develop systems for autonomous vehicles and integrate AI into manufacturing processes and future vehicles.
- NVIDIA Dynamo: A new open-source software platform designed to optimize data center operations and improve efficiency in executing complex AI models.
Attention was given to two new personal AI supercomputers: DGX Spark and DGX Station. DGX Spark, dubbed “the world’s smallest AI supercomputer,” is powered by the Grace Blackwell chip and designed for researchers, students, and developers to build advanced AI models locally. Meanwhile, DGX Station, equipped with the Blackwell Ultra chip and 784 GB of memory, is tailored for large-scale AI tasks on a desktop, making cutting-edge technology more accessible.
These innovations underscore NVIDIA’s commitment to advancing artificial intelligence, robotics, and computing infrastructure, opening new possibilities for industries worldwide.
View more on GTC 2025 – Announcements and Live Updates | NVIDIA Blog.
