Despite investing $13.75 billion in OpenAI since 2019, Microsoft is reportedly exploring alternatives to the ChatGPT maker’s AI model, particularly for its 365 Copilot – an AI tool designed to assist with creating documents, presentations, spreadsheets, notes, emails, and more. The move is believed to be part of the tech giant’s effort to diversify its AI dependency and cut costs, according to a report by Reuters, citing people familiar with the development.
Launched in 2023, Microsoft 365 Copilot – which also got a ‘voice’ this year, named Copilot Voice – aggressively highlighted the use of OpenAI’s GPT-4 model in its early days. But now the scenario seems to be changing, possibly due to three main reasons: first, to avoid monopoly or exclusive dependency; second, to reduce costs; and third, to improve speed. Yes, the report suggests that the ‘speed improvement’ is also a factor that led Microsoft down this path.
Going forward, we could see Microsoft achieve its goals by integrating internal and third-party artificial intelligence models into 365 Copilot, which has over 28 million active users across Windows, apps, and websites as of November 2024. This is notable, as just a few weeks ago, Microsoft officially unveiled the latest iteration in its Phi series of generative AI models – Phi-4. With approximately 14 billion parameters, Phi-4 is claimed to be a compact yet powerful model, noted for its excellence in specialized tasks such as mathematical reasoning.
To this date, Microsoft has invested over $20 billion into the artificial intelligence segment, with a major chunk of it going to OpenAI alone. Priced at $30 per user per month, the AI-powered 365 Copilot is expected to see significant growth in both its user base and revenue. According to a forecast by Macquarie, Microsoft could generate approximately $7.3 billion in incremental annual recurring revenue (ARR) from 365 Copilot by the end of FY’25 (June), based on an estimated initial uptake of 5%, or about 20 million users. By FY’26, the portfolio of Copilot products is expected to generate $9.1 billion in revenue.
Speaking about OpenAI—which triggered the AI race with its chatbot ChatGPT—it now appears to be expanding further in the AI sector, reportedly planning to develop a full-fledged humanoid robot. The company has also recently announced its new generation of reasoning models – ‘o3’, which are claimed to be approaching ‘AGI (Artificial General Intelligence).
The development also comes just days after reports emerged suggesting that Google urged the U.S. trade regulator, the FTC, to end Microsoft’s exclusive agreement to host OpenAI’s AI technologies on its cloud, including its chatbot, ChatGPT.
