A new fault line has emerged inside Google: The Claude haves and have-nots

• Some Google DeepMind employees use Anthropic's Claude for coding internally.
• Most Google employees are only permitted to use the company's internal Gemini AI tools.
• The divide is causing tensions as Google pressures more employees to adopt AI.

A new divide has emerged within Google: those allowed to use Anthropic's Claude and those not.

Some Google DeepMind employees in recent months have been given access to the Claude AI tool for coding purposes, according to three people familiar with the matter.

Googlers are generally prohibited from using tools that aren't either made by Google or tailored specifically for internal use. Meanwhile, Claude has become one of the most popular AI coding tools across the tech industry.

The decision to let Google DeepMind employees use Claude has irked engineers in other parts of the company, who are only allowed to use Google's internal Gemini AI models for coding.

Some employees feel that Google's internal models aren't as good as Claude for coding, two of the people said. Giving some staff access to Claude has created tensions, as Google is telling more employees they are expected to use AI.

Some engineers have been given specific AI goals that will be factored into performance reviews this year. In some cases, these employees are being told they're not only expected to use AI to generate code, but to build tools that make their processes more efficient.

A Google spokesperson did not respond to a request for comment from Business Insider.

Google keeps its engineers on internal tools for various reasons. Much of its internal infrastructure is custom-built. It also believes dogfooding — where employees test and use products they launch to customers — can help them improve faster.

Other tech companies allow their employees to use external AI models. For example, Meta employees can use Claude internally, Business Insider previously reported.

The Google DeepMind CEO snaps back

Last week, Steve Yegge, a computer programmer and blogger, posted on X about Google's AI adoption. He said he spoke to a Google director, who told him the company's internal AI adoption was lagging.

"The TL;DR is that Google engineering appears to have the same AI adoption footprint as John Deere, the tractor company," he wrote on X.

The post sparked plenty of discussion and solicited a blunt response from Google DeepMind CEO Demis Hassabis: "Maybe tell your buddy to do some actual work and to stop spreading absolute nonsense. This post is completely false and just pure clickbait."

Yegge posted again on Monday, saying he'd heard from some Google employees who corroborated his claims. Yegge also said that some DeepMind engineers use Claude, while engineers across the rest of the company aren't allowed.

"When the question of equalizing access came up internally, the proposed response was to remove Claude for everyone — which DeepMind objected to so strongly that several engineers reportedly threatened to leave," he wrote.

Have something to share? Contact this reporter via email at hlangley@businessinsider.com or Signal at 628-228-1836. Use a personal email address and a non-work device; here's our guide to sharing information securely.