AI
Corporate Employee Says Department's AI API Spending Has Reached 20% of Past Years' Payroll
A V2EX user posted that his company began large-scale purchases of the Claude API distributed through AWS in February with the release of Opus 4.6, and later purchased a Gateway service covering multiple model providers while setting monthly quotas for employees. The poster says that so far large model API spending has accounted for 20% of the department's total payroll spending in previous years and could exceed 30% for the full year, and he worries about quota cuts and layoffs next year; these are all based on his personal understanding and speculation.
The poster says he asked about the additional costs generated by the company's AI adoption this year while having dinner with the VP of his large department. According to his account, the company started AI adoption relatively late, only beginning large-scale purchases of the AWS-distributed Claude API in February this year, and later purchased an external Gateway service to achieve coverage of almost all model providers.
The company controls usage by giving employees monthly limits. The poster says that so far, overall large model API spending has already accounted for 20% of the department's total payroll spending in previous years, and based on the trend he estimates the full calendar year could reach over 30%.
The poster also says the company's annual profitability has not improved and may even be worsening, so he speculates that this spending will ultimately be passed on to employees. He believes quota cuts and layoffs may occur next year and estimates the layoff rate could exceed 20%. The above cost figures, business impressions, and layoff judgments all come from the poster's personal account; the material contains no official company data or confirmation.