Meta and Microsoft Ask Employees to Reduce Claude Use, Favor Own AI Tools
Meta and Microsoft have instructed employees to decrease their reliance on Anthropic's Claude AI model and utilize in-house developed tools to curb costs.
Technology giants Meta and Microsoft are reportedly directing their employees to reduce usage of Anthropic's Claude AI model, prioritizing internally developed artificial intelligence tools to cut expenses. This strategic shift aims to manage significant operational costs associated with third-party AI services.
Microsoft had previously projected annual expenses for its internal use of Anthropic AI to reach at least $1 billion. The company has since reduced this budget by over a third. This directive primarily targets internal daily operations and does not affect Microsoft's external cloud platform services, through which it continues to offer Claude to its clients.
At Meta, the number of employees using Claude Code internally has reportedly fallen to around 30,000 from approximately 60,000 earlier this year. This decrease is attributed to both Meta's workforce reductions and the company's strong promotion of its own AI tools, such as MetaCode and Muse Code. Muse Code, seen as a direct competitor to Claude Code, has garnered over 6,000 internal users since its August launch.
Despite the reduction in users, Meta's spending on Claude Code remains substantial, exceeding $105 million in the past 28 days. Both Meta and Microsoft are considered key clients for Anthropic, collectively contributing an estimated 25% of its revenue. Anthropic has noted in its filings that many of its large customers do not have long-term contracts, allowing for potential shifts in spending.