Nvidia’s 2025 GPU Technology Conference (GTC) took place from March 17 to 21 in San Jose, California, attracting over 25,000 attendees. The event featured significant announcements in AI technology, hardware, and strategic partnerships

Key Highlights:

• Next-Generation Architectures: Nvidia has unveiled plans for its next-generation Vera Rubin architecture, slated for release in 2026. This highlights Nvidia’s commitment to continuous innovation in AI infrastructure. Rubin AI Chips: CEO Jensen Huang introduced the next-generation Rubin AI chips, expected to surpass the performance of current models like Blackwell and Hopper.
• Blackwell Platform Evolution: Nvidia’s Blackwell platform is a central focus, with updates on its production and deployment. The announcement of Blackwell Ultra AI hardware, set for release in the second half of 2025, indicates a rapid progression in performance capabilities designed to enhance AI workloads and bolster cloud service offerings. ​. This includes increased memory, speed, and overall performance.
• AI Software and Tools: The introduction of Dynamo, an open-source software designed to enhance AI inference efficiency and reduce costs, is a significant development. This software aims to optimize GPU resource allocation for large language models.
  There is also alot of focus on agentic AI, and the software tools needed to create those systems.

• DGX Personal AI Computers: The company announced DGX AI computers powered by Blackwell Ultra chips, aimed at assisting developers in running large AI models on desktops.

• Strategic Industry Partnerships: Major collaborations were highlighted, including extended partnerships with Microsoft, Google, and Oracle to integrate Nvidia’s AI hardware into their cloud services. Additionally, General Motors plans to utilize Nvidia’s technology for advanced driver-assistance systems, and GE Healthcare will leverage AI for automated medical image diagnoses. 

Server manufacturers with Cisco, Dell Technologies, Hewlett Packard Enterprise, Lenovo, and Supermicro.

Partnerships with companies like Disney, to create advanced robotics.
Partnerships with companies like Google deepmind, to improve AI generated content watermarking.

These partnerships highlight Nvidia’s role in driving AI adoption across various sectors.

• Data Center Infrastructure: Emphasis has been placed on the growing demand for AI data center infrastructure, with projections of substantial capital expenditures in the coming years.
• Robotics: Nvidia is heavily investing in the robotics field. With demonstrations of advanced robots, and software platforms to create them.

• Developer Engagement: The conference featured over 1,000 sessions and 400 exhibits, with hands-on training and networking events, emphasizing Nvidia’s commitment to supporting the developer community
  

Despite these advancements, Nvidia’s stock experienced a 3.4% decline following Huang’s keynote, reflecting investor concerns about future chip demand and revenue growth