The annual developer conference of NVIDIA Corporation (NASDAQ:NVDA), known as NVIDIA GTC (GPU Technology Conference), is set to commence on Monday.
The conference, which will be held between March 16-19, has become a significant event for AI industry enthusiasts and will take place at the SAP Center, the home ground of the San Jose Sharks.
The event is expected to draw a whopping 30,000 attendees from 190 countries. The conference will span across ten venues in downtown and will also be streamed free on nvidia.com for virtual attendees.
This year’s GTC will cover a wide range of topics, including physical AI, AI factories, agentic AI, and inference. NVIDIA’s founder and CEO, Jensen Huang‘s keynote is expected to touch upon the full stack: chips, software, models, and applications.
With over 700 sessions planned, the conference promises to provide comprehensive details on the latest developments in the AI industry.
Here’s What To Expect
Pregame Show: The pregame show will feature the CEOs of Perplexity AI, LangChain, Mistral AI, Skild AI, and OpenEvidence three hours before Jensen Huang takes the stage.
Open Vs Closed Models: Harrison Chase and leaders from Andreessen Horowitz, Allen Institute for AI, Cursor, and Thinking Machines Lab, discuss with Huang how open models compare with frontier closed models and what it means for developers building on them.
Physical AI Systems: Experts to demonstrate practical, end-to-end workflows for physical AI development using NVIDIA Isaac and NVIDIA Omniverse technologies.
AI Factories: Discussion on designing and scaling enterprise AI factories for LLMs, agentic AI, physical AI, and HPC, highlighting the infrastructure, software stack, and NVIDIA‘s NV-Certified systems and reference architectures that help partners deploy AI systems efficiently and consistently.
What Do Analysts Say?
Economic Framing
Chris Smith from McKinsey said, “What’s changed this year is the economic framing.” He explained that Huang’s latest framing of AI as the rise of “AI factories,” where data and electricity produce tokens and insights, will be key to this GTC. In this model, data centers shift from cost centers to revenue-generating production facilities, with compute directly tied to earnings.
Rubin Power Demand
Sebastien Naji of William Blair is watching how Nvidia will handle the high power demands of its upcoming Rubin chips and whether major cloud providers will support them, while also looking ahead to the Feynman generation, which may introduce copackaged optics to move data with light instead of electricity, reducing power use and enabling larger AI clusters, reported Business Insider.
Beyond GPU And Racks
UBS analyst Timothy Arcuri said Nvidia may expand its AI systems beyond GPUs and racks, with its SuperPod architecture—introduced at the Consumer Electronics Show—combining chips, networking, storage, and software to help customers build and manage large-scale AI workloads, reported MarketWatch.
New Chip Tied To Groq
A Truist Financial analyst, William Stein, said the GTC could modestly boost the stock, though not in a “forceful” way, with investors already optimistic about its roadmap from Blackwell Ultra to the Feynman architecture, and a potential new chip tied to its Groq acquisition expected to be unveiled.
Huang Outlines Five-Layer AI Stack
The GTC conference comes on the heels of significant developments at NVIDIA. Earlier this week, Huang stated in a blog post that the AI infrastructure buildout, currently costing a few hundred billion dollars, could eventually reach trillions, emphasizing that the effort is still in its early stages.
Huang wrote that AI development relies on five “layers”—energy, chips, infrastructure, models, and applications—all of which must scale together for widespread adoption, with Nvidia positioned at the center linking much of this ecosystem.
These developments underscore NVIDIA’s commitment to advancing AI technology and its strategic investments in the sector. The GTC conference is expected to further highlight the company’s vision and future plans for AI.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by a Benzinga editor.
Image via Shutterstock
Recent Comments