"Amazon has deployed a ton of Nvidia's GB200 and GB300 NVL72 racks, and given the visual similarities to its Trainium3 Ultra Servers, we wouldn't be the least bit surprised to find out that large portions of the racks are shared between the two. In fact, with the launch of Trainium4, Amazon has already announced it'll be able to slide its custom compute blades directly into the same MGX chassis used by Nvidia's GPUs, so clearly we're headed in this direction."
"This makes business sense: At the scale AWS operates, the fewer one-off parts the cloud titan has to wrangle the better. For Amazon, it's better to have one modular rack architecture than one for each chip in the datacenter. This is one of the reasons that hyperscalers like Amazon and Meta founded standards bodies like the Open Compute Platform (OCP) in the first place."
"But it's not just the racks that look the same now - the compute and network fabrics do as well. Speaking at Re:Invent on Thursday, Peter Desantis showed off the Trainium3 compute blade, which pairs a Graviton CPU with four Trainium3 accelerators and a pair of Nitro data processing units. Up until now, AWS' Trainium systems had used x86 CPUs from Intel."
Amazon's Trainium3 UltraServer racks closely resemble Nvidia's GB200 and GB300 NVL72 racks, indicating shared components and converging rack designs. Trainium4 blades will fit the MGX chassis used by Nvidia GPUs, enabling blade interchangeability across accelerator types. Hyperscalers prefer modular rack architectures to minimize unique parts and simplify operations at large scale. Industry standardization efforts such as the Open Compute Platform have incorporated contributions like Nvidia's MGX designs while AMD and Meta introduced a Helios-based double-width rack. Compute and network fabrics are aligning, with Trainium3 blades pairing a Graviton CPU, four accelerators, and Nitro DPUs, mirroring multi-GPU plus CPU plus smartNIC configurations.
Read at Theregister
Unable to calculate read time
Collection
[
|
...
]