NVIDIA Unveils Blackwell Ultra, GM AI Partnership & Two Personal AI Supercomputers at GTC 2025
March 19, 2025 | San Jose, CA
At the highly anticipated NVIDIA GTC 2025, CEO Jensen Huang unveiled a wave of major announcements that further cement NVIDIA’s position at the core of the global AI revolution. The conference, held at the San Jose McEnery Convention Center, introduced new products, AI partnerships, and innovations aimed at transforming industries from automotive to robotics.
Blackwell Ultra and Rubin: NVIDIA’s Next-Gen AI Powerhouses
Leading the announcements was the Blackwell Ultra, NVIDIA’s most advanced GPU series to date, set to dominate data centers and AI labs globally. Huang also revealed the forthcoming Vera Rubin chip architecture, scheduled for release in 2026, promising exponential boosts in memory and compute capabilities. The Rubin Ultra lineup, expected in 2027, will bundle four GPUs into a single powerhouse package.
Huang quipped, “The more Blackwell you buy, the more you make,” taking aim at older Hopper architecture chips.
GM Expands AI Partnership with NVIDIA
In a major strategic shift, General Motors (GM) and NVIDIA announced an expanded partnership to leverage AI in autonomous vehicles, manufacturing, and enterprise operations. GM, which recently scaled back its commercial robotaxi efforts, will now integrate NVIDIA’s AI for Super Cruise driver-assistance systems, factory automation, and design simulations.
Groot N1: A Leap in Humanoid Robotics
NVIDIA also introduced Groot N1, an open-source foundation model designed for humanoid robotics, co-developed with Google DeepMind and Disney Research. The model, inspired by dual-system human cognition, will power next-generation Disney BDX droids, which were showcased live on stage next to Huang.
AI Supercomputers for the Edge
The tech giant launched two “personal AI supercomputers” — the DGX Spark and DGX Station — aimed at researchers and enterprises seeking to fine-tune AI models on the edge. These systems will support the growing need for localized AI deployments, echoing trends in edge computing and AI inferencing.
NVIDIA Fires at DeepSeek in AI Model Race
Huang took direct aim at Chinese AI firm DeepSeek, stating that NVIDIA’s new Llama Nemotron Reasoning model outperforms DeepSeek’s popular R1 model in speed and accuracy. Partnerships with Accenture, BlackRock, and Sequoia Capital were also announced to accelerate enterprise adoption.
Newton Engine: AI for Disney’s Next-Gen Robots
In a surprise move, NVIDIA confirmed its collaboration with Disney and Google DeepMind on Newton, a cutting-edge physics engine designed to simulate realistic robotic movement. Disney plans to deploy Newton in its next wave of park robotics, particularly in Star Wars-themed attractions.
Wall Street Reacts
Despite the flurry of innovations, NVIDIA’s stock (NASDAQ: NVDA) dipped 4% in after-hours trading, as investors expected additional surprises beyond what was leaked ahead of the event.
Bigger Picture: AI as an Industry OS
In an interview with CNBC, Huang downplayed the impact of Trump-era tariffs on NVIDIA’s supply chain, emphasizing partnerships with TSMC and a commitment to U.S.-based AI chip production. “AI is now the operating system of every industry,” Huang told CNBC, signaling NVIDIA’s ambitions beyond silicon.
Conclusion
As Silicon Valley sharpens its competitive edge against global challengers like DeepSeek and Huawei, NVIDIA’s GTC 2025 has set the tone for an AI-driven future where industries, from entertainment to transportation, are reimagined by intelligent systems.
Leave a comment