Total cost of ownership
The full cost of running a system over its lifetime — training + inference + operations + maintenance.
Think of it like
The cost of owning a car: upfront purchase, fuel, insurance, maintenance, depreciation.
Example
A model costs $100K to train, $10K/month to serve, and has a 2-year useful life. TCO = $100K + $240K = $340K.
How it actually works
TCO includes hidden costs: engineering time to maintain, retraining when the model becomes stale, handling support issues, compliance. A cheap model might have expensive operations (buggy serving code, poor monitoring).
For product teams
TCO is how you compare different solutions — a cheap model with expensive operations might cost more than a pricey model that's reliable.
For engineers
Instrument everything so TCO is visible; cheap training but expensive ops is a bad deal.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome