Latest posts
-
The Model Is the Easy Part
What a technical session on Nutanix Enterprise AI says about running LLMs on your own infrastructure: the gateway, the benchmarks, and the rack that sets the limit.
-
A VM Matched Bare Metal for LLM Inference. Here Is Why, and Where It Stops.
A Nutanix performance engineer measured vLLM in a VM matching bare metal on one GPU. Here is what published results say, why the hypervisor costs so little, and where the claim stops.