Notes from the stack.
What we measure, what it means for your fleet, and what we ship next. Every number comes with the hardware and model it was measured on.
What GPU to GPU compression is worth on a $1 billion GPU budget
The biggest training runs are limited by the network between racks, not by the GPUs. Here is what sending 24.5% to 49% fewer bytes, bit for bit, is worth on a 100,000-GPU run like GPT-6 Astra and on a $1 billion GPU budget.
Read the post
One stack for every layer of AI serving and training
weights.zip started as lossless compression for model weights. Today it covers the whole path a model takes, from the bytes in storage to the tokens streaming back. Here is the stack as one product, and what each layer saves.
Read the post
Start with your bottleneck.
Tell us where your fleet spends its time and memory. We run the matching layer on your models and your GPUs and report the result.