Microsoft and OpenAI have formed a high-profile partnership that has garnered significant attention in the tech industry. However, insiders familiar with the matter reveal that the relationship has been accompanied by conflicts and confusion.
Microsoft made a substantial investment in OpenAI, a smaller but pioneering company in the field of generative artificial intelligence. To avoid antitrust scrutiny, Microsoft acquired only a 49% stake in OpenAI, creating a unique arrangement that can be likened to an open relationship rather than an exclusive one. While Microsoft hosts OpenAI’s technology on its cloud and incorporates it into its popular products, OpenAI has been providing its technology to Microsoft’s competitors. This situation has led to limitations on potential search-engine customers for OpenAI, effectively reducing its market reach.
Despite the positive alignments and mutual admiration between Microsoft CEO Satya Nadella and OpenAI CEO Sam Altman, there have been complaints from within Microsoft regarding reduced investment in the company’s in-house AI projects. Additionally, most Microsoft employees do not have access to the inner workings of OpenAI’s technology, creating further frustrations. Both Microsoft and OpenAI sales teams sometimes approach the same customers, causing overlap and confusion.
The partnership between Microsoft and OpenAI has deviated from the conventional norms in the tech industry, where larger companies usually acquire smaller startups and their technologies. By maintaining a certain distance, Microsoft enjoys influence over OpenAI while still allowing the latter to explore other partnerships. However, this unconventional arrangement may face challenges as both sides compete to generate revenue from similar software and services.
Both Microsoft and OpenAI have publicly praised the benefits of their partnership, with Microsoft’s CFO Amy Hood describing it as mutually advantageous. However, the situation becomes more complex as both companies sell access to OpenAI’s technology, either through Microsoft’s Azure cloud services or directly from OpenAI.
There have been instances where potential customers have been approached by both OpenAI and Microsoft’s Azure team for access to products like ChatGPT, leading to confusion among companies. OpenAI has also established partnerships with other firms, including Salesforce, which offers a product called Einstein GPT that competes with OpenAI-powered features in Microsoft’s software. Moreover, OpenAI has engaged with various search engines to discuss licensing its products, coinciding with Microsoft’s integration of OpenAI technology into a new version of Bing. However, Microsoft has implemented policy changes that make it costlier for search engines to develop their own chatbots using OpenAI’s technology.
Some researchers at Microsoft have expressed dissatisfaction with the limited access to OpenAI’s technology, as most employees are required to treat OpenAI’s models like any other external vendor without access to its inner workings. Furthermore, tensions arose during the rollout of ChatGPT and Microsoft’s AI-infused Bing. Microsoft executives were concerned about ChatGPT’s launch potentially overshadowing the new Bing, while OpenAI advised Microsoft to proceed with caution and allow more training for the AI model to avoid issues. Despite the warnings, Microsoft released the Bing chatbot, which encountered problems and prompted the imposition of restrictions on its usage.
While ChatGPT has already gained significant popularity with 200 million monthly users and a high number of daily search sessions, the new Bing has struggled to achieve similar success. Satya Nadella recently announced plans to integrate Microsoft’s Bing search engine into the ChatGPT experience, signaling further collaboration between Microsoft and OpenAI.