Join us as we revisit Ultralytics' experience at NVIDIA GTC 2025 in San Jose, California, where we had the chance to explore groundbreaking AI innovations.
.webp)
From March 17th to March 19th, Ultralytics attended the NVIDIA GPU Technology Conference (GTC) 2025 at the San Jose Convention Center in California, USA. Our team engaged with industry leaders, researchers, and innovators. It was a great opportunity to connect, share insights, and explore the latest advancements in AI and accelerated computing.
The event was a fascinating showcase of the cutting-edge technologies that are shaping the future of various industries, from robotics to AI-powered solutions. We took part in inspiring discussions, attended insightful keynotes, and learned about the innovative breakthroughs driving the future of technology.
In this article, we’ll recap our time at NVIDIA GTC 2025, focusing on the exciting innovations, meaningful conversations, and the road ahead for AI. Let’s get started!
NVIDIA GTC aims to bring together some of the brightest minds in AI, machine learning, high-performance computing (HPC), and deep learning. Launched in 2009, the conference began as a relatively modest event at the Fairmont Hotel in San Jose, California.
Those who attended have mentioned that the event had the feel of a small science fair, with academics and researchers presenting their work on poster boards. The focus at the time was largely on GPUs (Graphics Processing Units) and their applications in gaming, with discussions centered around improving graphics and processing power for video games.
The event drew around 1,500 attendees and was focused on showcasing the potential of GPU technology. Today, GTC covers a much broader range of topics, including AI, autonomous systems, and data science, highlighting how much the conference has grown over the years.
NVIDIA GTC 2025 attracted around 25,000 attendees, including industry leaders, researchers, developers, and innovators. This five-day event has become a major highlight in the AI industry, earning the title of “The Super Bowl of AI” this year, following its recognition as “The Woodstock of AI” in 2024. As NVIDIA Founder and CEO Jensen Huang put it, “Everybody wins at the Super Bowl,” capturing the excitement and success of the event.
With over 1,000 sessions covering a wide range of topics, including generative AI, autonomous systems, virtual reality, high-performance computing, robotics, data science, and AI in healthcare, GTC expanded to two exhibit halls and additional venues like the PwC building for vendor meetings and sessions. Huang’s keynote at the SAP Center alone drew more than 17,000 attendees.
The event also featured several other notable speakers and discussions. Yann LeCun, Chief AI Scientist at Meta, and Bill Dally, Chief Scientist at NVIDIA, shared their insights on the future of AI models and hardware. On top of that, panels on the rise of humanoid robotics and the impact of AI agents in business provided valuable perspectives on the cutting-edge technologies shaping industries today.
Amidst all the excitement, Ultralytics had the chance to connect with the tech community and engage in interesting conversations. From the moment we entered the bustling halls of the San Jose Convention Center, we encountered familiar faces, new connections, and potential partners all eager to explore the latest discoveries in AI. The atmosphere was lively and welcoming, with attendees freely exchanging ideas and insights.
In particular, there was a lot of enthusiasm surrounding Vision AI, a branch of artificial intelligence focused on enabling machines to interpret and understand visual information, such as images and videos. This technology is quickly transforming industries, from autonomous vehicles to healthcare and robotics, with the potential to drive smarter, more efficient systems across a wide range of applications.
Many of the people we spoke to were interested in discussing how to integrate advanced YOLO pipelines and leverage Ultralytics solutions, as well as exploring innovative collaboration strategies.
Our team also spent time sharing how our Ultralytics YOLO models, built for computer vision tasks like object detection and tracking, are fast, lightweight, and efficient. We compared our models to larger, more complex AI systems, stressing that in the fast-paced AI world, agility and efficiency are essential.
One of the main highlights of the event was Jensen Huang’s keynote, where he shared an exciting vision for the future of AI with some major announcements. Let’s take a closer look at some of the key points that stood out.
Jensen Huang spoke about how NVIDIA’s GPUs are driving the future of deep learning, starting with the upcoming Blackwell Ultra chip, which is set to be released in the second half of 2025. This next-gen GPU will offer larger memory and improved performance, making it capable of supporting the increasingly complex AI models that are becoming the norm in the industry.
After the Blackwell Ultra chip, NVIDIA plans to release the Vera Rubin chip in late 2026, named after the famous astronomer Vera Rubin. The custom Vera design will be twice as fast as the CPU used in last year’s Grace Blackwell chips. Also, the even more powerful Vera Rubin Ultra is set to follow in 2027. These advancements will help customers keep their AI infrastructures up to date and scale effectively as deep learning continues to evolve.
Jensen Huang also spotlighted NVIDIA's efforts in robotics, unveiling the ISAAC GR00T N1, a foundational model designed for humanoid robots. This model is paired with a new physics engine called Newton, developed in collaboration with Google DeepMind and Disney Research.
The GR00T N1 uses a dual-system architecture, where System 1 handles fast, reflexive actions and System 2 focuses on more thoughtful decision-making. This combination allows the robot to perform tasks like object manipulation and material handling with ease and precision.
The GR00T N1 is fully customizable, so developers can train it with real or synthetic data to fit specific needs. This makes it adaptable across industries such as autonomous vehicles and smart manufacturing. With its advanced capabilities, the ISAAC GR00T N1 is set to help tackle global labor shortages and push forward the next generation of automation, creating more intelligent and capable humanoid robots.
While Jensen Huang mainly focused on NVIDIA’s advancements in hardware at GTC 2025, he also highlighted the company’s progress in AI software. Along with introducing next-generation GPUs like the Blackwell Ultra and Vera Rubin chips, Jensen Huang announced the launch of the Dynamo software platform.
Dynamo is designed to optimize the use of multiple GPUs, allowing them to work together like one massive AI factory, significantly improving inference performance and enabling better scalability for AI applications. It is open-source and modular, giving developers the flexibility to scale AI workloads efficiently across different environments.
The platform includes features like disaggregated serving and smart request routing, which are particularly useful for running large language models (LLMs) and generative AI. Disaggregated serving refers to splitting different tasks, like processing input data and generating output, across separate GPUs. This helps make the process more efficient and speeds up performance. Similarly, smart request routing directs incoming tasks to the right GPU resources, reducing the need to repeat calculations and lowering overall processing time.
The keynotes and discussions at NVIDIA GTC 2025 made it clear that the future of AI is moving fast. For Ultralytics, the event was more than just a conference - it was a chance to see the excitement and potential of AI in action. Every conversation we had inspired us and fueled our enthusiasm for what’s possible with Vision AI.
The collaborative atmosphere left us feeling motivated and excited. We’re eager to turn what we’ve learned at GTC into real-world innovations and continue pushing the boundaries of computer vision and AI technology.
We had an amazing time at NVIDIA GTC 2025, where we had the opportunity to connect with industry leaders, learn about the latest advancements in AI, and explore new ideas in the field. The event showcased incredible developments in both AI hardware and software, reinforcing the growing impact of AI across industries.
Leaving GTC, we were inspired and motivated by the innovative technologies on display. The conference highlighted the rapid growth of AI and its potential, and we’re excited to continue exploring new opportunities in AI.
Join our community and explore cutting-edge AI projects on our GitHub repository. Discover the impact of AI in agriculture and the role of computer vision in manufacturing through our solutions pages. Check out our licensing plans and start your Vision AI journey today!
Begin your journey with the future of machine learning
EN
ZH
KO
JA
RU
DE
FR
ES
PT
IT
AR
TR
VI