AI Supercomputers: From Pre-training to Reasoning Power! #NVIDIA #Microsoft

[00:00:00]We built the first AI supercomputer together. That was based on ampere.

[00:00:04]Of course, hopper was an incredible success. This first two generations were focused on pre-training. Grace Blackwell came along and all of the focus moved to post-training, uh reinforcement learning, which allowed us to have reasoning models. And these reasoning models based on mixture of experts were incredibly intelligent, energy efficient, um but it requires a giant systems. And so we created NVLink 72, and the entire rack became one computer. We had evolved from one node to now one rack. Well, Microsoft deployed the largest number of Grace Blackwells in the world today.

[00:00:47]The fastest and the largest number of Grace Blackwells in the world.

[00:00:51]Fairwater is just a magnificent system to look at.

[00:00:55]It's just a miracle of engineering. It's It's just an incredible feat. Uh it's completely liquid-cooled. You mentioned something earlier that I'm very proud of as well, that it's closed-looped.

[00:01:06]Basically uses almost no water, and it's incredibly environmentally uh friendly.

[00:01:11]It's energy efficient. We're able to increase the token generation rate and reduce the cost of token generation by an order of magnitude, some 30 times over hopper. So that was a huge achievement.