# Goldman Sachs limits Claude Code to engineers, Bloomberg reports

> Bankers and traders lose access to Claude Code and GitHub tools after months of use, according to Bloomberg's November 2 report.

URL: https://claudegpt.si/news/goldman-sachs-claude-code-restriction  
Published: 2026-11-09T02:53:21.142Z  
Author: ClaudeGPT Si Newsroom  
Section: AI Tools  
Topics: Goldman Sachs, Claude Code, Anthropic, GitHub, Coding agents  
Publisher: ClaudeGPT Si

## Key takeaways

- Bloomberg reported on November 2 that Goldman Sachs limited Claude Code and GitHub tools to engineers under October guidelines.
- Traders and bankers had used the tools for months; a Goldman spokesperson cited productivity and risk management.
- We could not open Bloomberg's original, and the coverage gives no count of affected staff.

Goldman Sachs has told most of its bankers and traders to stop using AI coding tools, including Anthropic's Claude Code and Microsoft's GitHub tools, according to Bloomberg's report of November 2, as summarized by [The Decoder](https://the-decoder.com/goldman-sachs-pulls-ai-coding-tools-from-traders-and-bankers/). The story is back on the Hacker News front page this weekend, and people are searching for what the Goldman Sachs Claude Code restriction covers.

## What Goldman changed

Bloomberg cites internal guidelines from October. Under them, only engineers keep access to the tools. Traders and bankers had used them for months, according to the report, to track trading activity and to produce written and visual material.

The change is a reversal. In July, Goldman's chief information officer, Marco Argenti, praised "vibe coding" by non-technical staff. Vibe coding means describing what you want to an AI agent and letting it write the code. The plan then was for a trader or asset manager to build a prototype with AI and hand it to an engineer to scale. That plan appears to have been dropped.

## What Goldman says

Spokesperson Nick Carcaterra told Bloomberg the decision reflects a "variety of factors including productivity and risk management." He said the experimentation phase is over and that access is limited to people best suited to use the tools for the long term.

## What we don't know

The reports give no count of affected employees. They do not say which specific risks drove the decision, and neither Anthropic nor Microsoft is quoted in the coverage we could read. We could not open the Bloomberg original, so every detail above comes from The Decoder's account of it.

## Why it matters

Banks are among the biggest prospective buyers of coding agents, and they also carry the strictest rules on code, data and audit trails. A top bank narrowing access after months of open use is a signal that firms are moving from trying agents to deciding who may use them. It is one company's internal policy, not a ban across Wall Street, and Goldman has not said it will stop using AI coding tools overall.

For Anthropic, the timing is awkward, since it is preparing a public listing. Our coverage of that is in [Anthropic IPO faces investor pushback on a $1.5T valuation](https://claudegpt.si/news/anthropic-ipo-investor-pushback).

## Sources

- [Goldman Sachs pulls AI coding tools from traders and bankers](https://the-decoder.com/goldman-sachs-pulls-ai-coding-tools-from-traders-and-bankers/), The Decoder
- [Goldman restricts traders and bankers from using some AI tools](https://www.bloomberg.com/news/articles/2026-11-02/goldman-restricts-traders-and-bankers-from-using-some-ai-tools), Bloomberg
