As the costs of AI continue to rise, companies are looking for ways to reduce them. The most recent example is Microsoft, which has reportedly begun implementing a cost-saving strategy by relying less on software from OpenAI and Anthropic and instead implementing its own internal models.

In fact, when it comes to two of its most used programs (Excel and Word), Microsoft has started using its home-made MAI models to respond to a certain percentage of user requests, Bloomberg reported on Tuesday. In the past, the company had announced the fact that much of Office 365 is powered by both OpenAI and Anthropic models.

While Microsoft still relies on those third-party models, it has also increasingly sought to champion its own AI agents. Last month at its annual Build conference, the company announced the launch of seven new MAI models, including an agent encoder and a text-to-image generator.

When contacted by TechCrunch for comment, Microsoft said it had nothing further to share.

Microsoft’s apparent cuts are part of a broader trend. After a brief blitz of “tokenmaxxing” earlier this year, recent months have seen a news cycle flooded with stories about tech companies acting significantly more thrifty. Other large companies, such as Amazon, Uber, Meta and Accenture, have also taken measures to curb spending.

The immense cost of providing and purchasing AI services has become a controversial part of the industry. The impact of the labels has become so severe in some parts of Silicon Valley that some companies are reportedly looking to Chinese models for more affordable agent solutions, despite some concerns about potential security issues.