Microsoft (MSFT+0.24%) is celebrating its anniversary with an event on Friday at its headquarters in Redmond, Washington.
Alongside an employee-only anniversary celebration of the tech giant’s 50th birthday, Microsoft is also preparing to announce next steps for its consumer AI initiatives and updates regarding the company’s AI assistant Microsoft Copilot.
The event comes on the heels of recent rumors that Microsoft is in the works to produce in-house artificial intelligence models, including reasoning models like OpenAI’s o1 which the company incorporated into its Copilot products earlier this year, in an effort to wean off its reliance on OpenAI.
Both Bloomberg and The Information reported last month that a number of these AI models, which Microsoft has dubbed MAI, produced results that proved competitive with rival products from OpenAI and Anthropic on accepted benchmarks. The models were also tested to see if they could power Copilot tasks currently executed using OpenAI models.
Microsoft is OpenAI’s core investor, and used to be its exclusive compute provider until January 2025 when the companies announced a new agreement that allows OpenAI to partner with competing cloud providers. Microsoft, however, still has right of first refusal.
Microsoft is also testing out models from Meta (META-1.69%), DeepSeek, and Elon Musk’s xAI as potential replacements for OpenAI, according to The Information report.
The company also announced brand new deep reasoning capabilities and AI agent workflows for Microsoft Copilot last week.
Copilot Studio will now enable two deep reasoning agents for Microsoft 365 Copilot users later this month. It’s part of a new initiative called “Frontier” meant to give 365 Copilot customers early access to innovations while still in the development phase.
Researcher agent, which utilizes OpenAI’s deep research AI model, will help users with complex, multi-step research, while Analyst, built on OpenAI’s o3-mini reasoning model, will assist researchers with making sense of raw data.