Microsoft is adjusting its AI strategy as the cost of advanced models continues to climb. The company is now using more of its own MAI models in selected parts of Word and Excel, reducing its dependence on OpenAI and Anthropic for some user requests.
According to recent reports, Microsoft has started routing a portion of prompts in its flagship productivity tools to in-house systems designed to handle everyday tasks more efficiently. The move reflects a broader effort to balance performance, flexibility, and operating costs across its software ecosystem.
At its Build conference last month, Microsoft introduced seven new MAI models, including an agent-focused coding model and a text-to-image generator. These additions signal a clearer push toward building a more self-reliant AI stack inside the company's core products.
The shift also mirrors a wider industry pattern. Several major technology firms are now looking for more efficient ways to deploy AI, with cost control becoming a central part of product planning and infrastructure decisions.
As AI tools become more deeply embedded in everyday software, companies that can combine strong model performance with lower operating costs may shape the next phase of digital productivity.
