Microsoft is preparing to overhaul its Office 365 suite by adding artificial intelligence models from Anthropic — the latest sign of a growing rift with OpenAI, according to a report.
The software giant will pay to use Anthropic’s Claude models for some Office 365 Copilot features, according to two people involved in the project cited by The Information.
The decision follows internal testing that found Anthropic’s models outperformed OpenAI’s in generating PowerPoint decks and handling complex Excel functions.
The change blends Anthropic and OpenAI technology inside Microsoft’s most important business software, which serves more than 430 million paying users worldwide, according to The Information.
OpenAI’s flagship GPT-5 remains in use for several Copilot tasks, but Anthropic’s Claude Sonnet 4 will take over advanced work like spreadsheet automation and presentation design.
The move comes as Microsoft and OpenAI continue tense negotiations over OpenAI’s restructuring into a for-profit company ahead of an eventual public listing, The Information reported.
OpenAI’s plan to restructure into a public benefit corporation has triggered tough negotiations with Microsoft, its largest backer, over how much equity Microsoft will hold and what level of privileged access it will retain to future AI models.
The drawn-out talks, along with regulatory scrutiny in California, have delayed OpenAI’s IPO until at least 2026, complicating liquidity plans for investors and employees.
Investors including SoftBank are pressing for resolution, while OpenAI’s nonprofit parent insists it will maintain control to ensure its mission of developing AI for the benefit of humanity remains central.
OpenAI declined to comment when reached by The Information. Anthropic and Amazon Web Services, which hosts Anthropic’s models, also did not respond to The Information’s requests for comment.
A Microsoft spokesperson told The Information: “OpenAI will continue to be our partner on frontier models and we remain committed to our long-term partnership.”
Unlike its arrangement with OpenAI, where Microsoft’s deep financial investment grants free use of the startup’s models, Microsoft will pay AWS to access Anthropic’s technology.
Amazon is one of Anthropic’s biggest backers, making the deal notable as Microsoft relies on a rival cloud to power its productivity suite, according to The Information.
The Office 365 Copilot subscription will remain priced at $30 per user per month. Equity analysts estimate the add-on could already be generating more than $1 billion annually, assuming roughly 1% of Office 365 customers have signed up, The Information said.
Some enterprises have been slow to adopt the service, citing bugs in AI-generated presentations and spreadsheets. Microsoft insiders told The Information that Anthropic’s performance edge convinced executives to move forward.
CEO Satya Nadella recently elevated Charles Lamanna to oversee business applications, including Copilot. Lamanna has been leading the integration of Anthropic models into Office in recent weeks, The Information reported.
The shift isn’t limited to Office. Earlier this year, Microsoft announced that its GitHub Copilot would also tap Anthropic models for advanced features, moving away from exclusive reliance on OpenAI.
The company has additionally explored alternatives for its consumer-facing Copilot app, which competes with ChatGPT but has far fewer users, according to The Information.
The rivalry between OpenAI and Anthropic is intensifying. Both startups are developing workplace productivity tools that compete directly with Microsoft’s software.
OpenAI agents can generate Excel sheets and PowerPoint files, while Anthropic unveiled a similar tool this week, The Information noted.
Despite the competition, Microsoft is now depending on both companies to power its flagship products, a dual-track strategy that underscores both OpenAI’s continued importance and Anthropic’s rising influence in the enterprise AI market.
The Post has sought comment from Microsoft, Anthropic and OpenAI.
