Nvidia Reveals AI-Focused RTX 500, 1000 GPUs For ‘Highly Portable’ Mobile Workstations

It’s the latest move by Nvidia to promote its view of the AI PC category, which differs from rivals like Intel and AMD due to the emphasis on using discrete GPUs to power AI tasks rather than CPUs that come with an integrated GPU and neural processing unit (NPU).

Nvidia has revealed a pair of low-end RTX Ada Generation GPUs that are designed to boost AI workflows for ultrathin laptops used by professionals and content creators.

The AI chip giant is pitching the RTX 500 and 1000 GPUs, announced Monday, as a significant step above CPUs for AI performance in “highly portable” mobile workstations but not as powerful as higher-end GPUs like the RTX 2000, 3000 and 5000.

It’s the latest move by Nvidia to promote its view of the AI PC category, which differs from rivals like Intel and AMD due to the emphasis on using discrete GPUs to power key AI tasks rather than CPUs that come with an integrated GPU and neural processing unit (NPU).

[Related: Gartner, IDC Think AI PCs Will Rule The PC Market In A Few Years]

The Santa Clara, Calif.-based company said mobile workstations from Dell Technologies, HP Inc., Lenovo and MSI will arrive in the spring with the RTX 500 and 1000 GPUs along with a CPU that includes an NPU. While Nvidia didn’t specify which vendor will provide these CPUs, both Intel and AMD currently offer CPUs with NPUs.

While the NPU can handle “light AI tasks,” the RTX 500 and 1000 GPUs are built for more “demanding day-to-day AI workflows,” such as removing background noise in video editing software, blurring images with upscaling in photo editing software or even “using AI for higher-quality video conferencing and streaming experiences,” according to Nvidia.

When it comes to what level of AI performance each GPU can provide, Nvidia said the RTX 500 and 1000 GPUs are capable of delivering 154 tera operations per second (TOPS) and 193 TOPS, respectively. These figures are lower than what more powerful RTX models can achieve, but they are significantly higher than the 34 total TOPS advertised for Intel’s new Core Ultra chips and the 39 total TOPS marketed for AMD’s Ryzen 8040 series.

Compared to a CPU-only laptop configuration, the RTX 500 can provide up to 14 times greater performance for generative AI models like Stable Diffusion, as much as three times faster AI-powered photo editing and up to 10 times greater performance for 3-D rendering, Nvidia said.

“AI is rapidly being adopted to drive efficiencies across professional design and content creation workflows and everyday productivity applications, underscoring the importance of having powerful local AI acceleration and sufficient processing power in systems,” wrote John Della Bona, a senior product marketing manager at Nvidia, in a blog post.