Steve Jobs' successor: Nvidia CEO Jensen Huang’s sensational presentation at GTC2025
Huang wowed with his captivating keynote on the future of Artificial Intelligence at what he dubbed the “Super Bowl of AI”.


Nvidia CEO Jensen Huang has long been touted as the ‘next’ Steve Jobs and he went some way to proving it at the company’s 2025 GPU Technology Conference (GTC), which he deemed the “Super Bowl of AI”.
Jobs' creative vision for the future and refined management style, which initially led to him being fired from Apple in his younger days, led to iMacs, iPhones, iPads and a company valued at $3 trillion.
Huang, previously a microprocessor designer at AMD and an incredibly high-rated engineer, co-founded Nvidia in 1993, since when the California-based company has outgrown even Apple in terms of its valuation, currently estimated at $3.5 trillion.
Jensen Huang’s obsession with the future and innovation
Like Jobs, Huang is obsessed with making long-term visions a reality through constant innovation, which was clear for all to see during his hugely impressive GTC keynote (which he gave without a script, no less).
Nvidia brands itself as the “World Leader in Artificial Intelligence Computing,” with Huang revealing what’s next in the ever-expanding world of AI in his lengthy address. The 62-year-old spoke for two-and-a-half hours, which gave him time to rattle through a number of topics, although not in as much detail as he would’ve liked, he admitted. He also had the opportunity to show off plenty of new tech, as you might imagine, as well as his considerable communication skills.
Huge steps forward for humanoid robots
Arguably the most eye-catching part of Huang’s keynote was right at the end, when he revealed Nvidia had partnered with Disney Research and Google DeepMind to create Newton, the world’s first open source physics engine for humanoid robotics simulation.
The project aims to advance robot learning and development, teaching them how to handle complex tasks with greater precision. One of the most difficult issues with robotics simulation to date has been the “sim-to-real” gap, which essentially means simulations can often be vastly different to reality.
Huang plans to play a key role in eliminating that gap and showcased the smooth-moving Blue, a Star Wars-esque robot that made its way from shuffling through sand on the giant screen to the stage. Blue, with two Nvidia computers inside, is powered by the Newton physics engine and, as the CEO explained, is an example of “complete real-time simulation”.
Check out Blue for yourself:
Nvidia showcases Blue, a cute little robot powered by the Newton physics engine.
— Chubby♨️ (@kimmonismus) March 19, 2025
This was the „wow“-moment for me. Robots are coming. This time for real.
And I am all in for it! pic.twitter.com/rjlmX275Ad
GTC 2025 keynote: other main takeaways
Blackwell AI chip in full production
Elsewhere, Huang announced Nvidia’s Blackwell AI chip at last year’s GTC and has now revealed it is in full production. Blackwell is said to be 2.5 times faster than the previously used H100 chip.
And Nvidia won’t be standing still. Blackwell Ultra, which will have one-and-a-half times more memory and two times more bandwidth, will be available in the second half of 2025 and will be used to accelerate building AI agents, physical AI and reasoning models.
Next-generation GPUs
Similarly, Huang announced the creation of next-generation GPUs (graphics processing units) for 2026, 2027 and 2028, each of which will be significant upgrades on the previous one. The last two of those will contain multiple GPUs connected together.
Reasoning models advancements
Huang unveiled Nvidia Dynamo and the Llama Nemotron family of reasoning models, which will allow developers and enterprises to build AI agents.
A reasoning model is designed to solve complex problems by breaking them down into smaller, manageable steps using explicit logical reasoning. That includes tasks related to critical thinking and decision making.
Personal AI supercomputers
Anyone from developers to students will be able to run their own AI models from their personal desktop thanks to the Nvidia DGX personal AI supercomputers, which can already be reserved but don’t yet have a price tag.
The Mac mini-sized DGX Spark, the “world’s smallest AI supercomputer” was previously announced with the name “Digits” earlier this year, at a starting price of $3,000.
Get your game on! Whether you’re into NFL touchdowns, NBA buzzer-beaters, world-class soccer goals, or MLB home runs, our app has it all.
Dive into live coverage, expert insights, breaking news, exclusive videos, and more – plus, stay updated on the latest in current affairs and entertainment. Download now for all-access coverage, right at your fingertips – anytime, anywhere.
Complete your personal details to comment
Your opinion will be published with first and last names