Microsoft’s significant $13 billion investment in OpenAI, the creators of ChatGPT, marks one of the most notable partnerships in the tech world. Despite this hefty financial commitment, the relationship might be encountering some challenges. Earlier this year, Microsoft briefly claimed the title of the world’s most valuable company, surpassing both Apple and NVIDIA, thanks to its early adoption of AI and strategic investment in OpenAI. This partnership has granted Microsoft early access to the latest AI advancements, which, in turn, are integrated into many of its products and services.
However, a recent Reuters report indicates that Microsoft might be planning to use new AI models for its Microsoft 365 Copilot service, bypassing those from OpenAI. This potential shift is attributed to concerns about the cost and speed of OpenAI’s GPT-4 model, which some believe does not meet the needs of Microsoft’s enterprise customers. Microsoft’s focus appears to be on reducing costs for features like GitHub Copilot, with the intention of lowering prices for consumers.
These revelations come amidst reports that hint at tensions in the Microsoft-OpenAI partnership, primarily due to disagreements over their exclusivity agreement and the high computing costs required for OpenAI’s advancements. There are speculations from OpenAI insiders that Microsoft’s struggle to meet computing demands might hinder its goal of reaching the AGI benchmark, particularly as other AI labs continue to make rapid strides.
Copilot 365, a critical part of Microsoft’s tools like PowerPoint and Word, illustrates some of these challenges. This tool is designed to streamline productivity by quickly accessing and summarizing data, meetings, and emails. Yet, despite having early access to OpenAI’s technologies, a recent report highlights Microsoft’s hurdles with effectively implementing Copilot’s advanced AI capabilities. An executive from Microsoft even labeled some Copilot tools as “gimmicky,” and revealed a dependency on third-party support to integrate Copilot across its products. Some users have complained about the service’s performance issues, claiming it doesn’t meet expectations 75% of the time, which raises questions about its $30 per user monthly fee.
In response, OpenAI is reportedly looking to remove a clause that would end its partnership with Microsoft upon reaching the AGI milestone. Sam Altman, CEO of OpenAI, speculates that achieving AGI might happen sooner than expected, potentially without significant societal upheaval. Moreover, sources within OpenAI suggest that the company might have already reached AGI with the release of OpenAI o1.
Considering these factors, Microsoft may be exploring alternative paths for its AI developments. Splitting its investment and reducing dependence on OpenAI could be a strategic move, especially following rumors of potential bankruptcy for OpenAI, with projected losses of up to $5 billion within a year. Microsoft CEO Satya Nadella has insinuated that parting ways with OpenAI once AGI is achieved might be the logical next step.
