Tech giants Meta and Microsoft are taking steps to significantly reduce the use of Anthropic’s Claude AI models among their employees, citing the rapidly rising costs of third-party AI and a desire to shift teams to in-house development.
Tech giants Meta and Microsoft are taking steps to significantly reduce the use of Anthropic’s Claude AI models among their employees, citing the rapidly rising costs of third-party AI and a desire to shift teams to in-house development.
According to The Information , Microsoft's internal spending on Claude this year could reach at least $1 billion. In this regard, the company's management urged employees to limit the use of a third-party tool and switch to its own Copilot system and other internal products.
Meta is taking similar steps. The company has reportedly already halved the number of employees using its Claude Code tool from 60,000 to 30,000. At the same time, the popularity of its internal alternative, MetaCode, has crossed the 30,000 user mark. Meta previously planned to spend billions of dollars on internal AI, but was forced to implement strict limits on tokens, spending controls, and monitoring.
For both corporations, switching to their own software is a logical economic solution that allows for significant cost reductions, even if their own models are currently slightly inferior in terms of individual characteristics.
Losing or significantly reducing contracts with its two largest clients could deal a serious blow to Anthropic's financial performance. According to Reuters, the startup is preparing for an initial public offering (IPO) with a potential valuation of more than $2 trillion.
The situation for Anthropic is also complicated by recent reputational problems: the US Department of Defense has recognized the company as a supply chain risk in the context of national security.
“The Company may experience material revenue losses or business disruptions as a result of these events,” Anthropic noted in its IPO prospectus.
This trend reflects a broader trend in the market: American organizations are increasingly looking for cheaper alternatives to expensive flagship models, optimizing their budgets for artificial intelligence.
