Transcript
Logan (0:00)
Foreign.
Podcast Host (0:05)
Welcome to Reshaping Workflows with Dell Pro Precision and Nvidia, where innovation meets real world impact in high performance computing.
Logan (0:19)
This is Logan from Reshaping workflows I've at GTC 2026. As you can see, we are roaming around the trade show floor looking for the most exciting happenings in AI that can only be found at GTC 2026. Speaking of great happenings, I've got my pal here, Alex Tilton from Dell Technologies, who is the product manager for the newly, I guess announced, soon to be launched, Delpro Max with GB300. So Alex, tell us more.
Alex Tilton (0:56)
Hey Logan. Well, first of all, thanks so much for having me. Really excited. Great show, great keynotes. Lots of people showing interest in Dell products. So that's always good. Loving it Here we're showing our GB300 today, Dell Pro Max with GB300, a brand new AI supercomputer. So it was announced it's going to be launching in a very short order here for the general availability. We are already delivering it to our first customers. So super excited about that. First to develop, first to ship, first to deliver, first to get it to work for our customers, solving their problems, working to get stuff done.
Logan (1:36)
So first, first, first, I hear a lot of firsts, so I'm going to ask you a question. So customers, you know, back in October received, you know, with the launch of the Dora Max, GB10, you know, a thousand AI tops. What makes the GB300 different?
Alex Tilton (1:51)
So GB300 is a totally different scale of work based on the name, of course you can say that they belong to the same family, right? But if I were to say it's an older brother, that would be a severe understatement. So GB300 is so much more than that. It is the first device of its kind that can handle much larger models, get stuff done much faster, much more memory on the system. Right. We are talking about supporting models that no other platform, no other notebook or desktop or even GB10 can do today for any of those AI workloads. So we're talking about capacity, we're talking about throughput. Everything is faster and better with GB300.
Logan (2:37)
I would agree. So, you know, for some context, GB10 is about a thousand AI tops. GB400 about 20,000. So 20x the size and model size up to 1 trillion parameters at FP4.
Alex Tilton (2:49)
You are right, Logan. That's. That's exactly right.
Logan (2:52)
