This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
While Nvidia (NASDAQ: NVDA) brings in most of its revenue these days from its data centre graphics processing units (GPUs) used to train and run artificial intelligence (AI) workloads, it's still the overwhelming market leader in the gaming GPU market. The company's market share typically hovers above 80%, with AMD in a very distant second place.
One reason why Nvidia has maintained such a dominant lead for so long is that it continually pushes the envelope with new features. Popular games and game engines implement those features, which may only work on Nvidia's hardware, creating an incentive for gamers to stick with Nvidia to unlock the best graphics performance and quality.
At CES 2025, Nvidia finally unveiled its RTX 50 series graphics card which is built on the Blackwell architecture. The company is promising up to twice the performance of last-generation products thanks to its various AI-powered features that can upscale images, generate frames on the fly, and perform other performance-boosting tricks. Along with the announcement, Nvidia discussed some new AI capabilities that should help the company stay on top in the graphics card market.
AI in more places
Nvidia rolled out the fourth generation of DLSS, its suite of AI-powered technologies that are capable of considerably boosting frame rates, along with the RTX 50 series. While Nvidia has touted that its new graphics cards can deliver twice the performance of their predecessors, much of that performance gain will be delivered by improvements to these technologies rather than increases in raw computing power.
One feature of DLSS is the ability to generate frames on the fly. Instead of rendering a game at 100 frames per second, with each frame going through the full graphics processing pipeline, the game can be rendered at a lower frame rate, and in-between frames can be interpolated. This enables gamers to enjoy improved visual fidelity without tanking the frame rate.
DLSS 4 bumps up the maximum number of frames that can be generated at one time. The previous iteration of DLSS was limited to generating just one additional frame since generating more would lead to performance problems. Nvidia also improved the quality of the generated frames, claiming that DLSS 4 reduces the visual artifacts that are inherent with this type of frame generation.
Beyond improvements to existing features, Nvidia is bringing AI deeper into the graphics pipeline with RTX Neural Rendering. The idea is to bring small neural networks into the programmable shaders that do the work of translating triangles, textures, and other data into 3D scenes. Developers can do multiple iterations of training on their game data and shader code, with the end goal of accelerating certain tasks at runtime.
This feature could be a big deal. One application is to use AI to compress textures, allowing developers to store more high-resolution images in the same amount of video memory. Another is to use AI to compress highly complex shader code, boosting performance by as much as a factor of five. How well all this works remains to be seen, but there's a lot of potential.
Yet another new feature from Nvidia is RTX Neural Faces. This feature allows developers to use generative AI combined with input data to generate realistic faces for characters. These faces are trained on thousands of generated images that cover different angles, emotions, and lighting conditions, resulting in a natural and convincing result.
Maintaining its lead
There will be 75 games that work with DLSS 4 when the RTX 50 series graphics cards launch, with more coming later. The broad support for Nvidia-specific features is one reason why the company has been able to fend off AMD and Intel in the graphics card market. For those wanting the absolute best gaming performance and visual quality, in many games Nvidia is the top choice.
The RTX 50 series will soon face new graphics cards from AMD, although AMD hasn't yet disclosed much information. Once Nvidia fills out its lineup with mid-range graphics cards, likely to happen later this year, Intel's impressive B580 graphics card will also be in the crosshairs.
While Nvidia will have some stiff competition, the company's AI innovations that boost frame rates, cut down on video memory usage, and increase realism will make it tough to beat.
This article was originally published on Fool.com. All figures quoted in US dollars unless otherwise stated.
