Blog
- The failure modes nobody warns you about
You are deciding whether to own your inference stack or rent it. Renting moves the OOM kernel, the thread scheduler, and the context-window boundary onto someone else’s host; yo...
- What a sovereign AI box cannot do
You are deciding whether to buy a sovereign AI box—hardware and software stack you own and operate—or to rent inference time from a cloud API. This page names the hard limits yo...
- What sharing cores does to two models at once
H1: What sharing cores does to two models at once
- Where the time goes when a model loads
Where the time goes when a model loads
- Model loading time is mostly not loading the model
H1: Model loading time is mostly not loading the model
- Quantisation without wrecking the model
You are technical and you are deciding whether to own inference or rent it. The question this page answers is how much quantisation you can apply before the model stops being us...
- Sizing RAM for a machine that holds several models at once
You are choosing hardware for a sovereign AI system and deciding how much RAM to dedicate to keeping model weights resident, versus how much to leave for the operating system, p...
- Buying vs. building your own local AI server: an honest comparison
You need a powerful AI server to run models locally—whether for personal projects, a small business, or a creative studio. The decision between buying a pre-built system and bui...
- How we spec a custom Garnet box for your workload
We build custom Garnet boxes for teams and individuals who need a reliable, efficient, and scalable machine to handle their machine learning and data work. The right box isn’t j...
- $0/local vs. cloud AI: the real cost and privacy tradeoffs
When building AI systems, teams face a fundamental choice: deploy models on their own hardware—$0/local—or rely on cloud providers like AWS, Google, or Azure. The decision isn’t...
- What is a sovereign AI box, and who is it for?
A sovereign AI box is a self-contained, local AI system that runs powerful open-source models on dedicated hardware—your own machine, always on, always ready. It’s not cloud-dep...