Microsoft and OpenAI may both be moving towards greater independence. According to a Wednesday report by The Wall Street Journal, OpenAI has reportedly signed an agreement with Oracle for computational access, a deal potentially worth $300 billion. This appears to be separate from an earlier agreement in July to develop 4.5 gigawatts of additional Stargate data center capacity for a project announced earlier this year. Oracle declined to comment. Meanwhile, Microsoft, a long-time investor in OpenAI, has not denied or confirmed reports from The Information on Tuesday suggesting that it plans to integrate Anthropic's technology into certain features of Office 365 applications, moving beyond its exclusive reliance on OpenAI for tools like Word, Excel, Outlook, and PowerPoint. If either development proves accurate, it would signal a significant shift in the evolving AI landscape. The Oracle-OpenAI report followed Oracle's quarterly earnings call, during which the vendor revealed it had signed major cloud contracts with companies such as OpenAI, xAI, Meta, Nvidia, and AMD. OpenAI did not respond to a request for comment prior to publication. “These indicators align with the booking and delivery of some of these large-scale compute contracts,” said Gartner analyst Chirag Dekate. He added that the deal complements the Stargate initiative, focusing on leveraging and securing GPU resources. OpenAI has been active in recent months, launching its Norway-based Stargate AI data center project in July and reportedly entering into a cloud resource agreement with Google. “They appear to be laying the groundwork for extreme-scale inference execution,” Dekate noted, adding that most of OpenAI's current inference scaling supports consumer products like ChatGPT. The Oracle deal raises questions about the future of OpenAI's relationship with Microsoft. “Microsoft still retains preferential rights,” Dekate said. “This seems to suggest that OpenAI is growing more rapidly and therefore needs faster access to extreme-scale capacity and a way to fund it.” Microsoft and Anthropic Microsoft responded to reports of its potential adoption of Anthropic AI technology by reaffirming its ongoing collaboration with OpenAI. A Microsoft spokesperson stated, “OpenAI will continue to be our partner for cutting-edge models, and we remain committed to our long-term partnership.” While Microsoft has invested over $13 billion in OpenAI over the past seven years, it continues to diversify its product portfolio by offering models from other providers such as xAI’s Grok and Anthropic’s Claude via GitHub Copilot. Last month, the cloud giant also launched its first in-house models, MAI-Voice 1 and MAI-1-preview. Microsoft Azure AI Foundry offers more than 11,000 AI models from various providers. “With the recent announcement of the MAI-1 preview, Microsoft has signaled a shift in its relationship with OpenAI,” said David Menninger, analyst at ISG Software Research. “Like other hyperscalers, Microsoft will both compete with OpenAI — introducing rival models from other teams — while also collaborating internally across different groups.” He noted that the two companies are evolving in divergent directions, likely due to their differing strategic objectives. “OpenAI has been chasing AGI (artificial general intelligence) for quite some time,” said Bradley Shimmin, analyst at Futurum Group, adding that this influences its go-to-market strategy. “That’s not the top priority for enterprise buyers or customers.” On the other hand, Microsoft’s reported interest in acquiring Anthropic AI technology illustrates its growing role as a developer of frontier models focused on responsible AI. Beyond its responsible AI efforts, Anthropic has also become a leading provider of agent tools and code generation, which aligns well with Microsoft’s broader strategy. “At its core, Microsoft is a company that builds tools for builders,” Shimmin said. Menninger added that Microsoft likely sees a business opportunity. “Anthropic’s customers consume a significant amount of cloud resources,” he said. “Microsoft wants a share of that business.” By partnering with Anthropic, Microsoft also strengthens its position in the competitive race with Google, which has accelerated its AI development since the release of Gemini 2.5. “If Microsoft wants to outpace its competitors — particularly Google in AI delivery — I don’t think they can catch up on their own,” Shimmin said. Partnerships are clearly playing a growing role in Microsoft’s strategy. On Monday, infrastructure provider Nebius announced a deal under which Microsoft will lease GPU-based computing capacity from Nebius over a five-year period under a $20 billion cloud agreement.