NVIDIA's annual GPU Technology Conference (GTC), nicknamed this year the “Super Bowl of Artificial Intelligence (AI) events,” took place this week. Each year, CEO Jensen Huang steps onto the stage in his signature leather jacket to unveil the latest breakthroughs in AI, accelerated computing, and deep learning.
This year’s keynote cements NVIDIA’s dominance in AI while demonstrating how strategic collaborations continue to shape the future of technology. Over 25,000 people filled the auditorium live, and the keynote has been streamed over 18 million times.
But is this the last year of being number one, with competition closing the gap? Think about what happened in January, when NVIDIA's stock took an 18% nosedive after Deepseek R1 was released.
It did recover afterward, but it shows investors are keeping a very close eye on who is delivering what and who will bring them more revenue. For now, the battle has only just begun.
Now, how does Jensen prepare before a keynote? Surprising the Acquired FM pregame show hosts and guests with some Denny’s Diner "NVIDIA breakfast bites." That was a fun moment for those of us watching the live feed.
The Next Leap in AI Computing
Huang opened the keynote with a bold vision, an era where AI accelerates at speeds we have not seen before, transforming every industry from healthcare and robotics to quantum computing. Although most of it had already been leaked online, he unveiled the latest NVIDIA’s next-generation AI chips as if it were the first time we were hearing about it.
NVIDIA introduced the Blackwell Ultra GPU, an evolutionary leap in AI processing. It boasts a massive boost in performance compared to its predecessor, Hopper, offering exponential increases in compute efficiency. The real game-changer was the introduction of the Vera Rubin chip, scheduled for release in 2026, and the Vera Rubin Ultra in 2027.
NVIDIA needs to work hard to stay ahead of the competition in AI hardware
These chips represent the next decade of AI acceleration, promising unmatched speed and power for training massive AI models. Huang was clearly ready to sell, sell, sell.
NVIDIA needs to work hard to stay ahead of the competition in AI hardware because companies like AMD and Intel are racing to innovate and catch up. One of NVIDIA’s biggest advantages for now is that it is deeply integrated into the AI software ecosystem.
AI Supercomputers for Everyone
Another announcement was the DGX Spark and DGX Station, NVIDIA’s latest personal AI supercomputers.
DGX Spark, the world’s smallest AI supercomputer, delivers up to 1,000 trillion operations per second. It is designed to make AI training more accessible to smaller businesses and researchers.
The DGX Station is a high-performance AI workstation that brings supercomputer capabilities to individual developers and enterprises, harnessing enterprise-level AI capabilities without the need for a data centre.
Robotics Revolution
I wouldn't be surprised if robotics was not on the table. One of the reveals was Isaac Groot N1, NVIDIA’s foundation model for humanoid robotics.
Inspired by human cognitive learning processes, Groot N1 is designed to train robots to interact with the physical world. The model integrates NVIDIA’s Omniverse platform, allowing roboticists to simulate real-world physics and refine AI-driven robotic movements before deployment.
NVIDIA’s approach, leveraging AI-powered simulations to enhance robot learning, has the potential to accelerate the adoption of autonomous robotics in manufacturing, logistics, and even personal assistance applications
The robotics industry has struggled with real-world adaptability. I questioned this when walking around at MWC25 in Barcelona earlier this month. NVIDIA’s approach, leveraging AI-powered simulations to enhance robot learning, has the potential to accelerate the adoption of autonomous robotics in manufacturing, logistics, and even personal assistance applications. The reality is that we are still in the early stages of robotics.
Quantum Computing
Quantum computing is quickly becoming a critical piece of the AI puzzle we find ourselves in. NVIDIA’s announcement of the Accelerated Quantum Research Center in Boston opens the doors to pushing the boundaries of computing. The new research facility focuses on solving major quantum computing challenges, such as qubit noise reduction and quantum error correction. By integrating AI with quantum computing, NVIDIA wants to accelerate breakthroughs in fields ranging from materials science and cryptography to medical research.
Quantum computing is still in its infancy, and sometimes it feels more like we are watching a TV show (remember Quantum Leap from the late 80s, early 90s?), but NVIDIA’s investment tells us that the future of AI is not just about faster GPUs anymore. It is about integrating multiple computing prototypes to unlock new potential.
Newton Physics Engine, A Game-Changer in AI Simulation
One of the most exciting collaborative efforts for me revealed at GTC was the introduction of the Newton Physics Engine. Developed in partnership with Disney Research (you know I'm a big Disney fan and their tech developments) and Google DeepMind, Newton is a next-generation physics simulation engine designed to revolutionise robotics and AI-driven simulations.
A little shoutout to the adorable AI-powered research robot Blue, inspired by Star Wars. You stole our hearts when you closed the keynote with Huang!
By being able to create more realistic simulations, researchers and engineers can better train AI systems in controlled environments before deploying them in complex physical spaces
Why is this important? Because the reality is that AI and robotics models struggle with real-world unpredictability. By being able to create more realistic simulations, researchers and engineers can better train AI systems in controlled environments before deploying them in complex physical spaces. This can and will accelerate the evolution of AI-powered robotics, self-driving cars, and industrial automation. I will follow up on this development closely.
Collaboration As the Magic Keyword
The common thread at NVIDIA’s keynote was one of my favourite words, collaboration! Time and again, we see that the most significant advancements in technology come from strategic partnerships between industry leaders.
Doing it alone will only take you so far, no matter how much money you have in the bank. NVIDIA's partnerships with DeepMind, Google, Cisco, Disney Research, and Oracle, to name a few, serve as a reminder of the value of working across industries. An ecosystem where innovation thrives is created when businesses with diverse strengths collaborate.
Competition may drive innovation, but collaboration amplifies it and takes it to the next level. NVIDIA’s partnerships push each other to discover what’s possible
Competition may drive innovation, but collaboration amplifies it and takes it to the next level. NVIDIA’s partnerships push each other to discover what’s possible. For example, NVIDIA’s joint efforts with storage leaders to develop AI-optimised enterprise infrastructure show how industries are evolving. By optimising AI storage solutions, businesses can unlock real-time insights from large data sets.
Final Thoughts
NVIDIA’s GTC 2025 keynote showed its vision for AI’s future regardless of the challenges it has had so far this year. From cutting-edge AI processors to groundbreaking quantum computing initiatives, the company, for now, remains at the forefront of the AI chip market.
Beyond the hardware and software announcements, the real message was clear, collaboration drives innovation. In a world where AI is already transforming everything we do as humans, working together is a normal evolution step.
AI should not be exclusive to the big players
And although we often focus on what the tech giants are doing, AI should not be exclusive to the big players. With AI models growing in complexity, the need for shared resources, open-source tools, and collective problem-solving is more key than ever. These collaborations are key to pushing the limits of what AI innovation can achieve, not just for business deals and profits but also to serve humanity.
As we move forward, my question to you is, how can we further create collaborative ecosystems that shape the next generation of AI breakthroughs?
