NVIDIA personal computing systems are about to get a serious power boost. At Computex 2025, NVIDIA unveiled two new AI computers: the DGX Spark and DGX Station. These devices bring supercomputer-level performance straight to your desk.

NVIDIA’s AI personal computing systems

While we’ve seen AI-powered PCs, like Microsoft’s Copilot+ computers, NVIDIA’s offerings aren’t your average PCs. These are built for developers, researchers, and AI labs, and they’re designed to handle intense workloads like training massive models or running complex simulations. The best part is that all of this can be done without needing a full server room.

Both systems are powered by NVIDIA’s new Grace Blackwell chips. These chips are built for the next wave of AI, especially “agentic” AI. These are models that can make decisions and take action on their own. With these chips, users can prototype, fine-tune, and run AI models faster than ever.

What is DGX Spark?

DGX Spark is designed for people who want serious AI computing without needing to rack up servers in a data center. This might is deal for developers or researchers working on smaller budgets or fewer resources. It’s powered by the NVIDIA GB10 Grace Blackwell Superchip, making it fast, efficient, and compact enough to sit under your desk.

We’re talking about up to 1 petaflop of AI compute here. For context, that’s roughly a thousand trillion operations per second. This is something you’d expect from massive cloud servers, not a workstation. It also comes with 128GB of unified memory. This helps when working with large datasets or complex neural networks without constantly juggling memory across components.

What really makes DGX Spark stand out is how it connects to a larger AI infrastructure. You can prototype and test models locally, then export everything directly to NVIDIA DGX Cloud or any other accelerated cloud or on-premise setup. That’s a big deal for researchers and startups. It means you don’t have to change your entire workflow just to scale

What about DGX Station?

As for DGX Station, that is the big one. It uses the ultra-powerful GB300 chip and offers up to 20 petaflops of AI performance. It also supports 784GB of unified memory and 800Gb/s networking speeds, which is very powerful for a desktop setup. You can run massive AI projects locally or split the system into multiple mini-computers using NVIDIA’s Multi-Instance GPU tech.

Both machines come with the full NVIDIA software stack, including the DGX OS, NIM microservices, and tools like PyTorch and Jupyter preloaded. Basically, they’re plug-and-play AI powerhouses.

Nvidia personal computing systems are also being picked up by major partners. Companies like Dell, HP, Acer, MSI, and GIGABYTE will offer these systems starting in July.

This article first appeared on Android Headlines

📰 Crime Today News is proudly sponsored by DRYFRUIT & CO – A Brand by eFabby Global LLC

Design & Developed by Yes Mom Hosting