1) Amused that after the overwhelming conviction in material datacenter spending reductions:
META will still grow. GOOGL will grow their “technical infrastructure spending.” MSFT guided next q up. AMZN ahead on the q and no guide. Capex only commentary.

2) Interesting that the effective sum of the above decisions in terms of YoY growth is ahead of consensus data center revenue growth for the x86 share gainer stripping out El Capitan.
Also: Pensando/Mi300 impacts are not well understood.
Nothing to see here, move along.

3) And the Green Goblin goes from strength to strength. The silicon and operational execution has been outstanding.
The Genoa DDR5 launch misfire (only 1 channel per DIMM - will be fixed) was first real silicon hiccup under new regime.
Ideally will be the last for a while.

4) Still curious to see the Grace/Hopper vs. MI300 showdown.
First real challenge to the incumbent *ever*
Software still weak for MI300, but improving & MSFT/META working hard to help with Triton.
+ The world ex GOOG standardizing on PyTorch makes it easier for challengers.

5) Also TensorFlow was the worst mistake GCP has ever made IMO and that is really saying something given the track record there. Maybe great for internal workloads, but just cripples them externally in what should be their greatest strength.