$1,500 a month. That's the per-employee AI spending cap Uber has reportedly set, a figure that developer and blogger Simon Willison analyzed as one of the more useful public calibration points on what enterprise AI use actually costs.
Most people using AI tools professionally are on flat subscriptions in the $20-50/month range: ChatGPT Plus, Claude Pro, or a coding assistant like Cursor. Getting anywhere near $1,500/month requires something categorically different - direct API access (where you pay per token, and a token is roughly three-quarters of a word, so processing large volumes of documents or code adds up fast), or automated pipelines that run AI continuously without a human approving each task.
The Cap Reveals What's Actually Happening
For Uber to set a $1,500 ceiling, employees must have been approaching or exceeding it before the cap was introduced. Finance teams don't build guardrails for hypothetical behavior. That means the company was observing real spending patterns that justified the limit.
The gap between a $20 consumer subscription and a $1,500 enterprise ceiling reflects two fundamentally different use patterns. Interactive users - people chatting with Claude or generating content in a browser tab - barely register in API costs. The expensive cases are automated: processing contracts at volume, running AI over large codebases, generating and testing hundreds of outputs without a human in the loop. Those use cases scale in a way that individual subscriptions don't.
Pricing Implications for Everyone
For small businesses and freelancers deciding how much to budget for AI tools, Uber's number is useful context. If heavy enterprise use at a large tech company tops out at $1,500/month per person, and you're spending $100-300/month across a handful of tools, you're not underinvesting.
For teams evaluating AI platform costs, the $1,500 figure also frames what should feel reasonable versus expensive. An enterprise AI proposal priced at $800+/month per seat warrants scrutiny about which specific use cases justify that spend.
The broader signal in Willison's analysis is that enterprises are now treating AI spend like cloud compute - a cost center that needs budget controls, not a collection of SaaS subscriptions that mostly go unmonitored. That shift has already changed how model providers price their API tiers, and it's pushing the market toward flat-rate enterprise pricing over pure consumption models. Predictable costs win procurement approvals.