NVIDIA CEO Jensen Huang introduces new AI-based chip technologies at this incredible CES 2025 | On the way to the future

There is no doubt that one of the most important events in the world of technology is CES (Consumer Electronics Show). In this new 2025 edition, many innovations are coming, mostly focused on AI, but also other developments that could impact the medium-term future. Some of these developments we have seen in the past edition, however, they have matured enough to surprise us this year.

This CES 2025 is once again held in Las Vegas, with four days of pure exhibitions of the latest technology built by human beings. A place where you can find your favorite brands showcasing their most advanced technological power, reflecting in some way that the future is getting closer.

---

CES 2025 official site

---

Probably one of the most amazing conferences has been provided by the genius Jensen Huang, CEO of NVIDIA. This time, a conference full of cutting-edge technology, where the common denominator is the advancement of artificial intelligence.

We have reached a point where graphics chips have advanced immensely. Something that Jensen highlights is that now, the RTX graphics card family is capable of making large-scale predictions. For example, the computational engine can commonly calculate 2 million pixels and the AI can calculate the other 33 million. With this in mind, rendering gets such a boost that affects performance after AI has been substantially trained. On the other hand, one of the important things he has mentioned is the way graphic models are rendered, reducing them to lower resolution, making their weight considerably lighter, and then intelligently leveraging a neural network to scale their resolution to a higher level, with less resource consumption.

---

---

This time, the RTX Blackwell was presented to us, maintaining Ray Tracing and new AI technologies. In terms of architecture, it is a graphic beast, with 92 trillion transistors and a memory bandwidth of 1.8 TB/s. While there are more technical details, these are the simplest to understand. What we can point out is the technology, which can compress neural textures and neural material shading. Once again emphasizing what AI technology can achieve. The RTX 50 series from NVIDIA will go on sale in March, focusing much more on a new power feature: AI TOPS.

Another striking thing is the training of AI for the future. What we used to see so far was training in real places, requiring laboratories or large spaces to train a robotic model. NVIDIA has come up with a solution that takes us to the so-called "Omniverse." A digitally simulated space where AI can be trained, and obviously will have access to a lot of tests and simulations with practically infinite possibilities. This is a big step in robot development because they will gather information in larger quantities and undergo all kinds of tests in considerably less time. This is what is known as physical AI, and it will accelerate the development of AI in physical spaces, such as the use of robots in warehouses, functioning autonomously. Then Cosmos takes care of giving it the final look, simulating reality using photorealistic synthetic data that simulates physics. This technology can achieve real-time token generation, allowing AI to select new ways to approach a path in a real job.

---

---

All this technological revolution follows the development of mechanisms that allow autonomous vehicles to exponentially improve their performance. For this purpose, the aforementioned technologies will also be used, with the mention that the vehicles will carry a computer with chips that process vital data, called Thor, which has 20 times the power of its previous version. What are the sensors, AI training, and other factors that make driving safe will be based on the power of this artifact. As expected, training will also occur in a simulated manner, creating 3D graphic environments represented with high fidelity to reality. NVIDIA has already created its own operating system called "NVIDIA DRIVE OS," which has been awarded for its high performance and performance.

There is no doubt that the development of robotics would not make sense without the advancement of artificial intelligence. But the development of technologies that allow intensive training for each of the tasks to be performed is also necessary. This would not be possible without new chips that allow the optimization and calculation of a large amount of data. The truth is that every year everything seems to improve even more, and we wonder, what amazing things will come next year? Technology is incredible!