JPMorgan Slams Brakes on Unchecked AI Spending with Claude Code Cap

JPMorgan Chase has implemented a $2,000 monthly spending limit on Anthropic's Claude Code for select engineers working within its internal secure development environment called Devspace.

This move signals that even the most aggressive AI adopters in finance are hitting the brakes on unchecked usage. The cap is part of JPMorgan's efforts to manage costs associated with AI investments, which are a significant chunk of its projected $20 billion technology budget for 2026.

Devspace serves as a controlled gateway that allows JPMorgan to monitor data flows and maintain compliance with internal security policies. This approach mirrors similar cautionary steps taken by Goldman Sachs in managing AI tool access across its engineering teams.