American chipmaker unveils new series of AI-powered graphics chips

ISTANBUL

Nvidia, the American chip giant, unveiled its new series of artificial intelligence-powered graphics chips during the opening day of CES 2025 in Las Vegas on Tuesday.

The GeForce RTX 50 Series graphics processing units, built on Nvidia’s Blackwell architecture for accelerated generative AI computing, will be available across multiple segments, following a similar approach to previous lineups.

Nvidia first introduced its GeForce RTX series in September 2018 as the world’s first graphics chips capable of real-time ray tracing — a technique that simulates how light behaves in the real world. The technology casts light rays from a source and accurately reflects them off objects and surfaces, producing photorealistic shadows and lighting effects.

The company said desktop versions of its most advanced model, the GeForce RTX 5090, along with the slightly less powerful GeForce RTX 5080, will launch on Jan. 30.

Nvidia also announced that the GeForce RTX 5070 Ti and RTX 5070 desktop versions will be released in February, while laptop variants of the new series will become available in March.

The latest GPUs feature Nvidia’s proprietary Deep Learning Super Sampling (DLSS) 4, including a newly developed multi-frame generation function.

DLSS improves gaming performance by lowering a scene’s internal rendering resolution and using AI to upscale the image in real time, allowing for smoother, faster gameplay without a noticeable drop in image quality.

With the new multi-frame generation feature, DLSS 4 leverages generative AI to create additional frames, enhancing visual fluidity for demanding real-time graphics.

In a news release, Nvidia said DLSS 4 and multi-frame generation will be exclusive to the RTX 50 Series and enable ray-traced gaming at 4K resolution with up to 240 frames per second, marking a major advancement in graphics performance.