Georgia Tech Lands $20M NSF Grant to Build AI Supercomputer ‘Nexus’

New system aims to democratize access to AI research tools and boost breakthroughs in medicine, energy, and neuroscience

Georgia Tech has just clinched a $20 million award from the National Science Foundation to develop something big — really big. It’s called Nexus, and it’s not your average supercomputer. This AI-powered behemoth is built to help scientists everywhere crack the hardest scientific puzzles, from curing diseases to understanding the human brain.

Once operational, Nexus will be one of the most advanced research-focused AI supercomputers in the U.S. The best part? It’ll be accessible to researchers across the country — not just the elite few.

A New Kind of Machine, Built for the AI Era

This isn’t your old-school, room-sized mainframe humming away in a locked lab. Nexus is different. It’s fast, powerful, and purpose-built for artificial intelligence.

Katie Antypas, who heads the NSF’s Office of Advanced Cyberinfrastructure, explained it this way: “Nexus brings a fresh mix of AI capabilities and high-performance computing. It’s going to support persistent scientific services alongside traditional research methods, which means we’ll see faster breakthroughs in more fields.”

Ángel Cabrera, president of Georgia Tech, was just as enthusiastic. “It’s fitting we’ve been selected to host this new supercomputer,” he said in a statement. “It will support a new wave of AI-centered innovation across the nation.”

What Makes Nexus So Special?

It’s hard to wrap your head around the raw power Nexus is packing — but let’s try.

At peak capacity, it will be capable of over 400 quadrillion operations per second. That’s like every person on Earth performing 50 million calculations a second — at the same time.

And that’s not all. Here’s what else it brings to the table:

• Memory: 330 trillion bytes — enough to juggle massive datasets without breaking a sweat

• Storage: 10 quadrillion bytes of flash memory, the equivalent of 10 billion reams of paper

• Connectivity: Blazing-fast data transfers for real-time computation

• Access: Open to all U.S.-based researchers, no matter where they work

It’s a scientific playground. And it’s being built from the ground up with user-friendliness in mind — researchers won’t need to be computing wizards to use it.

The Timing? Couldn’t Be Better

AI isn’t just a buzzword anymore. It’s how modern science works.

From drug discovery and genomics to climate modeling and space exploration, researchers now rely on AI to test theories, run simulations, and spot patterns in oceans of data. The trouble is, most institutions don’t have the resources to run these computations — not even close.

That’s where Nexus steps in.

Suresh Marru, the project’s principal investigator and the director of Georgia Tech’s new Center for AI in Science and Engineering (ARTISAN), put it bluntly: “This supercomputer will help level the playing field.”

One Supercomputer, Many Fields

Georgia Tech and its partners aren’t thinking small.

Nexus is expected to play a major role in accelerating research in at least five areas:

1. Health and Biotech: From new vaccines to personalized medicine

2. Clean Energy: Speeding up the search for better batteries and greener fuels

3. Neuroscience: Modeling brain activity with greater precision

4. Manufacturing: Optimizing production with AI-based simulations

5. Climate Science: Analyzing long-term models and environmental data at scale

And that’s just the shortlist.

“This isn’t just a Georgia Tech win,” said Marru. “It’s a national win.”

Nexus vs. Other U.S. Supercomputers

To give a sense of where Nexus fits in the landscape of American supercomputing, here’s a quick comparison:

While Nexus may not be the biggest in raw numbers, its architecture is optimized for AI workflows — something most supercomputers weren’t originally built for.

Designed to Be Used, Not Just Admired

One of the recurring criticisms of large-scale research computing is how inaccessible it can feel — especially for newer scientists or smaller institutions.

Georgia Tech says they’re fixing that problem head-on.

Unlike systems that require deep technical know-how, Nexus is being built with a simplified, user-friendly interface. Scientists will be able to focus on their work — not get bogged down in code or permissions.

Just one sentence here: That’s a big deal.

It means more institutions, more researchers, and more students can access state-of-the-art tools — without the intimidation factor.

Big Impact, Small Barriers

This $20 million investment by the NSF is more than just a grant — it’s a vote of confidence.

Georgia Tech isn’t going at it alone. The project brings in a network of collaborators, including Indiana University, the Texas Advanced Computing Center, and the San Diego Supercomputer Center.

Together, they’re building not just a machine — but an ecosystem. Nexus will plug into the NSF’s broader cyberinfrastructure, meaning it won’t operate in isolation. It’ll connect with other national research tools, software platforms, and data repositories.

That connectivity is key. It turns Nexus from a standalone powerhouse into a collaborative engine for national research.