Personally, if you ask me the best option for a PC with AI computing capabilities, My choice is very clear, RTX AI PC. And I begin by making this clear because it seems to me to be the most honest position to take in this conversation. Now, if I clarify this, then it is very clear that this choice is due solely to the objective data of what this platform is able to provide, if we compare it with what is offered today by systems in which the AI ​​computing capacity falls on the NPU.

To give a bit of context, let’s recall that one of the main trends related to AI in 2024 is the shift of certain tasks that until now depended on data centers, on the very systems from which they are requested. This means that if a user wants to use a certain AI-based function, the entire process takes place locally without the need to establish a connection to an online service.

In this regard, perhaps the best-known proposal is the Copilot+ PC from Microsoft, which, as you may remember, establishes a minimum of 40 TOPs that can be processed exclusively by the NPÚ. A capability that certainly already provides room for some AI workflows to be implemented. Although the NPU is a highly specialized chip, by being an element integrated into the processor, its capacity is much more limited than that provided by the GPU.

In this case, we also have to take this into account If we’re talking about RTX AI PCs, we’re focusing on systems that integrate the GeForce RTX graphics adapterwhether dedicated or integrated. And this is not an insignificant detail, because for years NVIDIA has specialized (among other technological areas) in artificial intelligence, which means that its hardware has a very high level of specialization for this type of tasks, so its performance in performing them is first class.

This week’s installment of the AI ​​Decoded series, in which NVIDIA tackles artificial intelligence from multiple angles, is dedicated today to the RTX AI PC and why this specification is so important in this regard. And the reason for solving this problem this week is that ASUS and MSI will introduce new laptops this week that complement the NVIDIA specification which is committed to GPU-based AI.

In terms of universal AI capabilities, the systems ASUS and MSI will announce at IFA this week They integrate the GeForce RTX 4070 graphics adapter and offer up to 321 AI TOPS of performancethat is, they multiply the performance of the Microsoft specification by seven, making them exceptionally powerful systems for performing all types of operations, from using a personalized chatbot to generating images of various types, extremely quickly and with complex systems. .

Also of course we have to consider what NVIDIA has been working on technologies that rely on the computing power of their GPUs for AIand here we can talk about everything grouped under the heading of DLSS to provide an incredibly optimized gaming experience, up to solutions like Chat with RTX, which puts a chatbot at our fingertips without having to worry about complex configuration settings.