NextPlatform estimates Google has double TPU3 chip performance and dense racks

Google announced the TPU 3.0 AI coprocessor and said it is 8 times faster than the TPUv2.

Above – Pods: TPUv2 (top) and TPUv3 (bottom)

Photos of the full TPUv3 pod show Google scaled its next-generation pods:

There are twice as many racks per pod
There are twice as many Cloud TPUs per rack

Scaling the TPUs and racks gives a 4X performance improvement over a TPUv2 pod. Most of the new hardware development appears to be happening at a system level around the TPUv3 chip.

Google claimed the TPUV3 :

* The TPUv3 chip runs so hot that for first time Google has introduced liquid cooling in its datacenters
* Each TPUv3 pod will be eight times more powerful than a TPUv2 pod
* Each TPUv3 pod will perform at “well over a hundred petaflops”

Google restated that its TPUv2 pod clocks in at 11.5 petaflops. An 8X improvement should land a TPUv3 pod at a baseline of 92.2 petaflops, but 100 petaflops is almost 9X.