San Jose, California – March 2025
This week at NVIDIA GTC 2025, the global stage for AI innovation, NVIDIA unveiled a powerful vision for the future—introducing breakthroughs that will define the next era of AI infrastructure.
From the launch of the Blackwell Ultra AI Factory Platform to advancements in agentic AI models, humanoid robotics, healthcare, manufacturing, and simulation-first data centre design, this year’s announcements mark a pivotal shift toward more intelligent, scalable, and efficient AI systems.
During his keynote, NVIDIA CEO Jensen Huang emphasised that modern data centres have become “AI factories”—next-generation digital infrastructure designed to run AI workloads around the clock.
“AI has made a giant leap – reasoning and agentic AI demand orders of magnitude more computing performance,” Huang noted, highlighting the urgent need for new infrastructure to power the next wave of AI-driven services.
The spotlight was on enabling reasoning, agentic, and physical AI at industrial scale—powered by innovations like the Dynamo inference engine, 800G Spectrum-X networking, open-source simulation platforms, and Omniverse Blueprint for digital twin–based design. These technologies are not only reshaping the infrastructure behind AI factories—they're redefining how organizations build, test, and scale intelligent systems.
As an NVIDIA DGX-certified provider, NEXTDC is proud to stand at the forefront of this transformation—delivering the high-density, liquid-cooled, AI-ready infrastructure that empowers our customers to fully leverage these next-gen advancements.
Top Announcements from NVIDIA GTC 2025 – and What They Mean for You
1.0 Blackwell Ultra AI Factory Platform
NVIDIA’s next-generation AI platform is designed for reasoning, agentic, and physical AI workloads at industrial scale. Powered by the new open-source Dynamo inference software and 800G Spectrum-X networking, it delivers higher throughput, lower latency and jitter, and reduced total cost of ownership.
Benefit: Run large-scale AI services more efficiently, with faster response times and optimised infrastructure economics.
2. Open Reasoning AI Models: Llama Nemotron Family
A powerful new open model family designed for building agentic AI platforms—capable of reasoning, collaborating, and executing complex workflows.
Benefit: Accelerate the development of intelligent AI agents tailored to solve real business challenges autonomously.
3. World’s First Open Humanoid Robot Foundation Model – Isaac GR00T N1
A foundational model for humanoid robots with generalised skills and cognitive reasoning, fully customisable for industry use.
Benefit: Enable next-gen robotics solutions for logistics, services, and industrial automation.
4. Physical AI in Healthcare: GE HealthCare + NVIDIA Collaboration
A partnership aimed at advancing medical imaging systems and autonomous diagnostics through simulation and AI robotics with Isaac™ for Healthcare.
Benefit: Improve speed and accuracy in diagnostics, reduce clinical workloads, and bring next-gen AI to healthcare devices.
5. General Motors + NVIDIA: AI for Vehicles and Smart Manufacturing
GM will use NVIDIA platforms to build the next generation of AI-enhanced vehicles, robots, and factory systems.
Benefit: Align with proven, scalable AI architecture already being adopted by global industry leaders.
6. NVIDIA Cosmos World Foundation Models & Physical AI Tools
A set of AI models and toolsets to support physical-world reasoning, controllable world generation, and prediction in simulated environments.
Benefit: Build intelligent systems that can understand, simulate, and navigate real-world environments with precision.
7. Newton Physics Engine (Open Source)
A powerful simulation engine for robotics and physical AI, developed in collaboration with DeepMind and Disney Research.
Benefit: Accelerate robotics R&D, reduce prototyping costs, and bring new automation systems to market faster.
8. Vera Rubin Ultra (2027)
An enhanced version of the Vera Rubin platform for AI mega-scale deployments with unmatched processing power and data throughput.
Benefit: Future-proof your infrastructure for the most advanced AI workloads across industries.
9. Omniverse Blueprint: AI Factories, Built Smarter
NVIDIA’s new Omniverse Blueprint introduces a simulation-first approach to AI factory design and operations, enabling teams to model, optimise, and validate infrastructure at gigawatt scale—before it’s built. Using real-time digital twins and integrated tools like Cadence Reality, ETAP, and Spectrum-X, engineers can test cooling, power, and networking performance, streamline supplier coordination, and reduce risk.
Benefit: Design and deploy AI factories faster, lower your total cost of ownership, avoid infrastructure failures, and ensure readiness for tomorrow’s AI workloads.
What This Means for You
As AI infrastructure becomes more complex and critical, staying ahead means being ready to scale, simulate, and optimise before deployment begins. As an NVIDIA DGX-certified provider, NEXTDC is ready to help you lead with liquid-cooled, high-density infrastructure that supports these cutting-edge advancements in reasoning, robotics, and real-time simulation.
Want to learn more about what was covered at NVIDIA GTC 2025 and how to accelerate your AI readiness?
Get in touch with your NEXTDC sales team today.
NEXTDC – Where AI Thrives.
