What “hosted in Canada” actually means

Data residency is a property of the infrastructure, not a checkbox in a contract.

Many providers advertise a Canadian region. In practice that covers very different realities: data stored in Canada but processed elsewhere, a region operated from another country, or a subsidiary subject to foreign law regardless of where the servers physically sit.

The questions that settle it are concrete. Where does inference compute physically execute? Who has administrative access to the machines, and from which jurisdiction? Do logs, caches, and backups follow the same rule as production data? And above all: can the entity operating the service be compelled by foreign law to produce your data?

For an agentic workload the stakes are higher than for a chatbot. An AI coding agent reads your repositories, your configurations, sometimes your production data. The context sent to the model is often more sensitive than whatever the AI agent does with it afterward.

Then there's training. “Your data is not used to train our models” and “your data is not shared with any third party” are two distinct guarantees, and the first does not imply the second.

Our position is simple because the infrastructure is: compute runs on servers in Canada, operated from Canada by a Canadian company. There is no region to select, because there is only one.