Nvidia unveils the key ingredient behind its DLSS technology

At Editors Day at CES 2025, Nvidia disclosed one of the factors contributing to the success of its DLSS AI upscaling technology: an enormous supercomputer. This supercomputer is the essential component behind DLSS, the most sophisticated AI upscaling technology for PCs.

With the introduction of DLSS 4, Nvidia transitioned to a new transformer-based AI model, moving away from Convolutional Neural Networks (CNNs). This change allows the new AI model to contain twice the number of parameters as the previous DLSS versions. Consequently, this has led to improved image stability, reduced ghosting, smoother edges, and better detail during motion, all made possible by the substantial GPU resources Nvidia has dedicated to its new AI models.

For six years, Nvidia has operated a supercomputer solely focused on refining DLSS. With DLSS 4, it is evident that this commitment has paid off.

However, Nvidia's DLSS Supercomputer isn't the sole reason for the high quality of their AI upscaler. Nvidia has been proactively identifying issues within DLSS so their AI models can address them. They have created a database containing examples of high-quality graphics alongside instances of problematic DLSS upscales.

Through their continuously expanding dataset and an AI that is always working to overcome DLSS-related challenges, Nvidia has rapidly improved the quality of their AI upscaling solution. This indicates that we can anticipate even better performance from DLSS in the future, which is excellent news for PC gamers. Nvidia has invested significant resources into the ongoing development of DLSS. DLSS 4 is the first iteration to employ Nvidia's new transformer-based AI models, suggesting this new AI model will continue to evolve over time.

And as always, while waiting for more exciting news, remember to use our comparison tool to find the best deal for your next favorite game.