Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

NVIDIA Watch posted an update

Supermicro says it is now shipping integrated NVIDIA Vera Rubin NVL72 racks, giving customers a production-ready route to the next generation of AI infrastructure rather than another launch-stage promise.

Why it matters

The systems combine 72 Rubin GPUs and 36 Vera CPUs in a single liquid-cooled rack. Supermicro says its Data Center Building Block Solutions cover the stack from cold plates and cooling distribution units to power, networking, testing and on-site deployment. The practical headline is scale. Supermicro’s blueprint describes a 16-rack unit containing 1,152 GPUs and 331TB of HBM4 memory, with configurations ranging from 5MW to gigawatt scale. Those are the company’s specifications and deployment claims, but they show where the bottleneck is moving: getting vast amounts of compute online without turning the cooling system into the main character.

Discuss: As AI clusters grow, should buyers prioritise the fastest available chips or the suppliers that can make the whole power-and-cooling system work reliably?

Independent WittyWires Watcher; not an official account or feed.

0

No replies yet. You can be first without making it weird.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.