Rival or Partner? Microsoft Develops Its Own LLMs to Compete with OpenAI 

Shutterstock

Microsoft, one of OpenAI’s most powerful backers, is reportedly building its own large language models (LLMs) to compete directly with the company it has invested billions in. Internally, these LLMS are being referred to as MAI, a name that could potentially stand for “Microsoft Artificial Intelligence.” 

The initiative, first reported by The Information, indicates a change in Microsoft’s AI ambitions. The tech giant may be trying to reduce its dependence on external partners, such as OpenAI, one of the leading AI companies in the world and the creator of ChatGPT.  

Microsoft has invested nearly $14 billion in OpenAI since 2019. The two companies have also been planning on a $100 billion ‘Stargate’ data center to support OpenAI’s advanced AI models. 

Mircosoft is experimenting with how these LLMs could improve its existing AI products, most of which are based on Microsoft Copilot. In addition, Microsoft is considering ways to integrate MAI into other products, including Microsoft Teams and Azure cloud services. 

The LLMs under training are primary reasoning models, which use chain-of-thought techniques known to offer enhanced problem-solving abilities by breaking down complex tasks into smaller and logical steps. Microsoft is reported to have plans to release the new LMLs as an API for external developers by the end of the year. The APIs will allow outside developers to weave these models into their own apps.

Along with the aim to reduce its dependence on OpenAI, Microsoft also wants to reduce its costs of using OpenAI’s models. Running the LLMs requires significant computing resources and potential licensing fees for external models. With in-house LLMs, Microsoft can optimize the AI infrastructure and create models tailored specifically for their enterprise applications. 

The move to create its own LLMs in-house is not groundbreaking news; however, for a major investor in OpenAI to do so has raised eyebrows. Some see this as a conflict of interest, as Mictosoft could be competing with the company that it supported both financially and strategically. 

Microsoft's decision to develop its own LLMs could strain its relationship with OpenAI and other AI developers. However, reports suggest the tension between the two companies has been building over the last couple of years after OpenAI released a competing enterprise product. 

There have also been reports that OpenAI employees have not been getting along with Microsoft workers. In addition, Microsoft has been exerting influence over the company, especially after OpenAI's CEO, Sam Altman, was fired and then reinstated under pressure from Microsoft. Some viewed this as a change in Microsoft’s role from an investor and supporter to having excessive control over the company. 

The relationship was further complicated when Mustafa Suleyman was hired by Microsoft. Suleyman, co-founder of DeepMind and CEO of Inflection AI, is an influential figure in the AI world but has been surrounded by controversies. He has faced allegations of workplace bullying and has been criticized for his role in data privacy disputes. 

When Microsoft brought Suleyman on board to lead its AI division, it not only signaled an aggressive push to dominate the AI space but also raised concerns due to his controversial past.

While the conflict with OpenAI is getting attention, Microsoft is also testing other models, signaling a broader strategy of AI diversification. The company has reportedly tested algorithms from Anthropic, Meta, and DeepSeek. This move could also be a significant step toward the development of Artificial General Intelligence (AGI), a goal that has rapidly become a priority for tech companies. 

Whether this decision ultimately enhances or complicates Microsoft’s AI strategy is yet to be seen. However, what is evident is that there is fierce competition in the AI industry, and old alliances are unlikely to stand in the way of their AI ambitions.