Asus Crams Nvidia Superchip Into Dull Desktop

The Desktop That Hides A Supercomputer

Most high-powered machines look the part, with blinking lights, giant fans, and loud hums. But Asus just introduced a desktop PC that looks normal while secretly delivering the kind of performance you’d expect from a full-on supercomputer.

It’s called the ExpertCenter Pro ET900N G3. On the outside, it has the look of a basic office tower, complete with a DVD drive. But inside, it’s packed with hardware designed for high-level artificial intelligence tasks.

Inside, One Chip, Two Brains

The real magic starts with what’s inside, the Nvidia GB300 Ultra chip. It’s actually a combo of two parts, one is a Grace CPU and the other is a Blackwell GPU, connected by a lightning-fast bridge called NVLink-C2C.

This high-speed link lets them share tasks and memory as if they were one super-powerful brain. That means less lag and better performance when handling huge AI jobs like training language models or running simulations.

More Memory, More Muscle

One of the biggest challenges with AI computing is memory. If your system doesn’t have enough, it simply can’t run the models you want. The ExpertCenter Pro ET900N G3 solves this with up to 784 GB of unified memory.

That’s not just more memory, it’s a whole new scale. It has more than twice the memory of high-end workstations with multiple GPUs. This huge pool of shared memory helps load massive models directly, reducing the need for slow transfers and making everything faster.

AI Performance Measured In Petaflops

The performance numbers on this desktop are wild. The ExpertCenter Pro ET900N G3 delivers up to 20 petaflops of AI power, that’s 20 million billion operations per second.

That kind of computing strength isn’t just impressive, it’s transformational. It means scientists, developers, and researchers can now run demanding AI workloads like large-scale language training or video analysis without needing access to a server farm.

A Look That Throws You Off

What’s wild is that the ExpertCenter doesn’t even look high-tech. It’s shaped like a basic black tower PC and even includes a DVD drive and old-school slots you haven’t seen in years.

This “normal” look helps it fit into places high-end machines usually can’t go, like office spaces, labs, or classrooms. No flashing lights. No special cooling rigs. Just an understated machine with massive computing power that hides in plain sight.

Built For AI, Not Just Business

Unlike general-purpose PCs, this one was built from the ground up for AI work. Every component is designed to handle complex algorithms, deep learning models, and massive data sets.

It runs on a specialized operating system called Nvidia DGX OS, which is based on Ubuntu Linux but optimized for machine learning and analytics. This OS includes the full Nvidia software stack, so users get tools like CUDA, TensorRT, and cuDNN pre-installed and ready to roll.

No Data Center? No Problem

You’d expect this level of performance to require a data center setup, racks, specialized cooling, and tons of wiring. But the ExpertCenter Pro ET900N G3 breaks that rule.

It runs cool and quiet on your desk, just like a normal PC. It doesn’t need extra fans, liquid cooling, or heavy-duty wiring. This is a huge win for schools, startups, and researchers who want serious AI power without the infrastructure headaches.

Fast Networking For Bigger Projects

The ExpertCenter isn’t just powerful, it’s also connected. It includes support for Nvidia’s ConnectX-8 SuperNIC, a high-speed networking solution that lets multiple systems communicate quickly and efficiently.

This is great for bigger AI projects, where multiple machines need to work together. Instead of sending files back and forth and waiting for uploads, you can move huge data sets instantly. It also means you can scale up your projects as needed.

Meet the Ascent GX10, AI In Your Palm

If the ExpertCenter is a desktop-sized supercomputer, the Ascent GX10 is the same idea packed into a mini-PC. It’s small enough to hold in one hand, but it’s no toy.

Inside is the Nvidia GB10 Grace Blackwell Superchip, which links a 20-core CPU to a powerful GPU with the same high-speed NVLink-C2C technology. The GX10 can hit up to 1,000 AI TOPS (trillions of operations per second), putting it in the same class as enterprise-level systems.

Tiny Size, Huge Power

The GX10 offers up to 128 GB of unified memory, enough to run 200 billion-parameter AI models locally. That’s miles beyond what your average desktop or laptop can manage.

Its compact size makes it perfect for edge computing, field research, or small labs where space is tight but performance still matters. Developers can now run serious AI tools without relying on the cloud or massive GPU rigs.

Two Mini PCs, One Giant Leap

Need to go bigger? The Ascent GX10 can pair with another unit using built-in ConnectX-7 networking, creating a powerful AI cluster without needing racks or cables everywhere.

This setup allows you to double your AI performance and run even more complex models, like Llama 3.1 with over 400 billion parameters. That’s something very few systems can do, and not in something this small. It’s a creative way to scale power without scaling up your space needs.

AI At Your Fingertips

These systems aren’t just about specs, they’re about accessibility. By bringing serious AI capabilities into smaller, simpler devices, Asus is helping more people take part in the AI revolution.

Students, creators, and small teams now have access to tools that used to be locked behind big budgets and specialized facilities. It’s like giving someone a racecar that drives like a regular sedan, easy to use, but built to win.

From Gaming To Global Research

While most people think of GPUs for gaming, these machines show how that same technology can power breakthroughs in health, science, and education.

You could use the ExpertCenter to develop a medical AI tool, or the Ascent GX10 to create a language app that runs offline. These systems bridge the gap between entertainment tech and serious research tools, bringing them into the hands of more people in more places.

Easy To Use, Hard To Outgrow

Both machines are designed to be user-friendly. You don’t have to be a system administrator or Linux expert to get started.

At the same time, they’re powerful enough to keep growing with you. As your projects get more complex, the hardware can keep up without needing constant upgrades. That makes them a smart long-term investment for anyone diving into AI work.

A Complete Lineup For All Needs

Asus is building a whole lineup of AI-ready hardware, from consumer desktops to enterprise servers. The ExpertCenter and GX10 are just two pieces of a much bigger picture.

No matter if you’re building a PC yourself, looking for an AI laptop, or planning a full AI deployment at a company or university, Asus has options. They’re working to cover every step on the path to AI adoption, making it easier to find the right fit for your goals.

Want to see how Asus is protecting those powerhouse GPUs? Check out how their smart monitoring feature keeps things cool and under control.

The Future Of AI Is Getting Personal

AI used to be something only giants like Google, Microsoft, Amazon, and Meta could afford to play with. But the latest systems from Asus show that’s changing, and fast.

By making AI computing more personal, portable, and powerful, they’re opening the door for everyone, from students to startups, to build the future. All you need is a desk, a power outlet, and a big idea.

Curious how Nvidia is keeping its cutting-edge chips cool under pressure? Take a look at how they tackled AI overheating issues.

Curious what you’d do with a supercomputer on your desk? Drop your thoughts in the comments and hit like if this blew your mind.

Read More From This Brand:

• Nvidia Set to Shake Up PC CPUs
• Nvidia or AMD Leading the AI Revolution
• Nvidia and Netherlands Unite for AI Innovation

Don’t forget to follow us for more exclusive content right here on MSN.

If you liked this story, you’ll LOVE our FREE emails. Join today and be the first to get stories like this one.

This slideshow was made with AI assistance and human editing.