Microsoft is making a major push to develop its own artificial intelligence (AI) reasoning models, potentially positioning itself as a direct competitor to OpenAI, despite being one of its biggest financial backers. According to a report from The Information, the tech giant is not only working on in-house AI models but is also considering selling them to external developers, marking a shift in its AI strategy. This move comes as Microsoft looks to reduce its dependence on OpenAI, whose ChatGPT models have powered key Microsoft products, including Copilot and other AI-driven offerings.
Microsoft has already started testing AI models from xAI (the AI startup founded by Elon Musk), Meta, and DeepSeek as potential alternatives to OpenAI’s technology. The company is exploring whether these models could serve as replacements for OpenAI’s advanced AI models in its Copilot product, which integrates AI-driven assistance into Microsoft 365 applications. This development reflects Microsoft’s desire to hedge its bets in the AI space and avoid relying solely on OpenAI, especially given the high costs associated with using OpenAI’s cutting-edge models.
In December, Reuters exclusively reported that Microsoft had been working on integrating both internal and third-party AI models into Microsoft 365 Copilot. The move was aimed at diversifying its AI model sources while also reducing operational expenses tied to OpenAI’s technology. When Microsoft first launched 365 Copilot in 2023, one of its biggest selling points was that it leveraged OpenAI’s GPT-4 model, widely considered one of the most powerful AI models available at the time. However, with the AI landscape rapidly evolving, Microsoft now appears intent on building its own proprietary AI models that could eventually compete with OpenAI’s offerings.
Leading Microsoft’s AI efforts is Mustafa Suleyman, co-founder of DeepMind and former CEO of Inflection AI, who joined Microsoft in early 2024. Under his leadership, Microsoft’s AI division has successfully trained a new family of AI models internally referred to as MAI. These models reportedly perform almost as well as OpenAI’s and Anthropic’s top-tier AI models on widely accepted industry benchmarks. The company is also working on reasoning models that incorporate chain-of-thought techniques, a method designed to improve AI's ability to generate answers with step-by-step logical reasoning. This approach is expected to make Microsoft's AI models more capable of handling complex problem-solving tasks, directly competing with OpenAI’s latest innovations.
According to The Information, Microsoft’s AI team is already experimenting with swapping out OpenAI’s models in Copilot for its new MAI models. The MAI models are described as being significantly larger and more advanced than Microsoft’s previous Phi models, which were smaller-scale AI models developed for specific applications. Additionally, Microsoft is considering making the MAI models available as an application programming interface (API) later this year, which would allow external developers to integrate Microsoft’s AI technology into their own applications and services. If successful, this move could further establish Microsoft as a major independent player in the AI space, providing businesses and developers with a viable alternative to OpenAI, Google DeepMind, and Anthropic.
Microsoft’s growing independence in AI development signals a broader shift in the tech industry, where companies are increasingly looking to build and control their own AI models rather than relying on external providers. This trend is driven by multiple factors, including cost considerations, data privacy concerns, and the desire for greater flexibility in customizing AI solutions. For Microsoft, developing its own AI models could lead to significant cost savings, as licensing and running OpenAI’s GPT-4 and future models are believed to be expensive. Additionally, having proprietary AI models would give Microsoft more control over its AI roadmap, reducing potential risks associated with being tied to a single AI provider.
Despite its strategic pivot, Microsoft remains deeply invested in OpenAI, both financially and technologically. The company has invested billions of dollars in OpenAI and continues to provide critical cloud computing resources via its Azure platform. However, as competition in the AI space heats up, Microsoft’s decision to build its own AI models suggests that it is preparing for a future where it may need to rely less on OpenAI and more on its in-house capabilities.
Neither Microsoft nor OpenAI has responded to requests for comment on this development. However, if Microsoft’s MAI models prove to be competitive, it could reshape the AI industry by introducing more competition and giving businesses more choices when selecting AI solutions. As Microsoft continues to push forward with its AI ambitions, all eyes will be on how these new models perform in real-world applications and whether they can truly rival OpenAI’s cutting-edge AI technology.
