Fresh data-center capacity in several regions eased GPU availability, helping push regional inference cost indices lower this week.
Compute supply, energy and data-center capacity decide how cheaply AI can run. Infrastructure shifts show up in inference costs weeks later.
Summaries are aggregated for information only — follow the source link for the full story. Demo entries are illustrative.