As the financial demands of artificial intelligence continue to escalate, many companies are actively seeking ways to reduce expenses. The latest example comes from Microsoft, which has reportedly started implementing a cost-cutting strategy by scaling back its reliance on software from OpenAI and Anthropic, opting instead to utilize its own internally developed models. According to a report from Bloomberg on Tuesday, Microsoft has begun using its proprietary MAI models to handle a certain portion of user prompts in two of its most popular applications—Excel and Word. In the past, the company had prominently highlighted that much of Office 365 was powered by models from both OpenAI and Anthropic.

While Microsoft still depends on those third-party models to some extent, it has increasingly focused on building its own AI agents. At its annual Build conference last month, the company introduced seven new MAI models, including an agentic coder and a text-to-image generator. This shift is part of a broader trend across the tech industry. After a brief period of aggressive "tokenmaxxing" earlier this year, recent months have been filled with news stories about major technology companies becoming far more frugal.

Other large corporations—such as Amazon, Uber, Meta, and Accenture—have also reportedly taken steps to curb their AI-related spending. The enormous cost of both providing and purchasing AI services has become a contentious issue within the industry. The sticker shock has grown so severe in some parts of Silicon Valley that certain companies are now exploring Chinese models as more affordable alternatives for agentic solutions, despite lingering concerns about potential security risks.