Infrastructure & Compute
Cold start
The extra delay when a service or model must initialize before handling its first request.
Example
The first request is slow because the serverless model must start and load its weights.
Why it matters
It helps teams explain delays that steady-state speed figures may hide.
You'll hear it
“The first request was slow, but the next ones were quick.”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Does a cold start happen on every request?
- No. A running, ready service may handle later requests without the same setup work.
- Can a large model increase startup delay?
- It can. Loading more information and preparing resources may take longer.
- Does keeping a service ready remove every delay?
- No. Busy periods, network conditions and processing work can still cause waiting.