Computing has changed from being written by humans to being generated by artificial intelligence.
If “GPU” sounds familiar, you probably associate it with graphics cards and video games, but the revolution NVIDIA CEO Jensen Huang discusses goes far beyond graphics! In a recent interview with the NoPriors channel, coinciding with CES 2025, Huang delved into how Artificial Intelligence (AI) transforms the world of computing, science, and even how we will work.
So that you don’t miss the most awarded products in this edition of CES 2025, I leave you with an interesting video from CNET.
In this post, I will tell you the fascinating keys to his vision, with simple examples so that we can all understand where the technology is headed and how it will impact us.
From Traditional Computing to AI-Driven Computing
Until recently, the software we knew was “programmed” by people, line by line of code. Now, AI has become the author of that software, learning from huge volumes of data and running ultra-fast on specialized GPUs.
- Paradigm shift: Huang explains that it is no longer just a matter of creating code step by step but of training an AI model to “generate” the best possible approach. This opens the door to solving problems of previously unattainable complexity.
- New Moore’s Law (or “Hyper Moore’s Law”): According to Huang, “Doubling or tripling performance every year is our mission.” The idea is not just to reduce the size of transistors but to leverage AI to improve every piece of the development process, lowering costs and consuming less power. In fact, he said that in 18 months, they have managed to reduce the cost of processing one million tokens by up to 240 times, making AI much more affordable.
Remember that a token is the basic unit that LLMs read, understand, and generate.
The Data Center: The New Computing Unit
New basic computing unit is not a chip, but a data center
We used to evaluate at the power of a single chip or processor. For Huang, the entire data center has become the fundamental computing unit. So, instead of just designing chips, NVIDIA creates integrated systems that function as true “AI factories.”
- Hardware-software co-design: “Hardware-software co-design is essential to reach new levels of efficiency,” says Huang. Every advance in chip architecture connects to AI library optimizations, exponentially maximizing performance.
- AI factories: Imagine your data center not just storing information but “manufacturing” it in AI models that improve processes, analyze data, and generate valuable results in record time. This integrated approach saves cost, energy, and time.
Humanoid Robots and AI at Work
One of the most surprising topics of the interview was Huang’s prediction that humanoid robots will become more prevalent by 2025. Why “humanoid”? Because these robots will be able to take advantage of environments designed for people without the need for structural changes, accelerating their adoption.
- Tokenization for movements: Just as a language model works with “words” (tokens), a humanoid robot could process “motions” as tokens. Every action, from opening a door to lifting an object, is a piece of data that the robot learns and predicts.
- AI collaborators: Huang envisions a future where we work alongside AI “employees” who can take on specific tasks and train themselves in minutes with company information. For him, this fusion between people and AI systems will increase productivity and accelerate innovation.
Impact on Science, Health and More
AI will not remain a business or robotics issue; it is also changing the way we discover scientific solutions:
- Quantum chemistry, digital biology, advanced mathematics: Huang highlights how AI is “transforming every industry” and mentions that models can now explore and optimize designs that would surpass human capability.
- AI chip design: “AI will help us design our chips with never-before-seen efficiency.”Advances feedback on each other: AI designs new chips, which power more and better AI models.
Towards a Global AI Economy
Every company and nation will need its own artificial intelligence factory
NVIDIA’s CEO predicts that both countries and companies will need to build their “AI infrastructure” to avoid being left behind. This will involve significant investments in data centers and a rethinking of the workforce, where humans and AI systems will work side by side.
- Democratization of AI: Thanks to cost reductions and efficiency improvements, it will be increasingly easy for small and medium-sized companies to adopt AI solutions without going broke.
- Compatibility and scalability: For Huang, NVIDIA’s architecture is designed so that “old” hardware is not immediately obsolete. This ensures that today’s technology investments will still be valuable years from now.
An exciting and challenging future.
Reflecting on Huang’s approach, it is clear to me that we are in the midst of a great revolution.
AI is no longer a laboratory curiosity but the center of technological transformation. From the way we design chips and data centers to the integration of humanoid robots and AI “collaborators” into the workforce, everything indicates that we are entering a new era.
Artificial intelligence is our greatest opportunity to transform all industries
If this sounds like science fiction, look at the advances in robotics, language modeling and cost reduction in neural network training. The future is being written (or, rather, generated) at the pace of algorithms capable of learning and improving endlessly.
For those of us who follow AI closely, it is clear that we will see massive adoption in the coming years. What does it mean for your company or your children’s education?
The best time to learn how to use AI is NOW!
And you, can you imagine working side by side with an AI in your day-to-day life? Do you think humanoid robots will be able to integrate easily into our environment, or will they require radical changes in infrastructure? What advances in AI do you find most beneficial for your personal or professional life? How will transforming data centers into “AI factories” affect companies’ competitiveness? Where do you see the most significant opportunity (or the biggest challenge) in the mass adoption of AI?
Let me know what you think in the comments!
Good week
