Index  ›  ai  ›  Times of India

After thousands of layoffs, Microsoft’s cost cutting drive hits OpenAI and Anthropic as company ‘dumps’ …

Times of India Published Jul 8, 2026 Reviewed Jul 8, 2026 ✓ Reviewed by citations.press editors
Citation-ready fact
Microsoft AI CEO Mustafa Suleyman stated at the June Build developer conference that Microsoft’s goal is to reduce and ultimately eliminate its payments to Anthropic.
View source ↗
Citation-ready fact
Microsoft is replacing tens of thousands of AI prompts in Excel and Outlook each week with its internally built MAI models, according to a Bloomberg report citing a person familiar with the matter.
about 10000 prompts · AI prompts in Excel and Outlook
View source ↗
Citation-ready fact
Microsoft’s MAI models are now integrated into GitHub Copilot for software developers, according to a Bloomberg report.
View source ↗
Citation-ready fact
Microsoft claims one of its new in-house AI models, introduced at the June Build developer conference, can match the coding capabilities of Anthropic’s Opus 4.6 model at a significantly lower operational cost.
View source ↗

Microsoft’s push to cut costs following thousands of layoffs has reached its premier artificial intelligence (AI) partnerships, a report has said, claiming that the tech giant is beginning to “dump” models from OpenAI and Anthropic inside its most popular software products, replacing them with its own internally built AI.Citing a person familiar with the matter, Bloomberg reported that tens of thousands of AI prompts within Microsoft Excel and Outlook are now being quietly handled each week by Microsoft’s home-grown “MAI” models.

Previously, these spreadsheet and applications relied on technology developed by OpenAI and Anthropic – the makers of ChatGPT and Claude, respectively.While internally handled tasks represent a relatively small portion of Microsoft’s total AI footprint, the shift proves the company is making major strides in building highly competitive artificial intelligence (AI) at a fraction of the cost.Microsoft AI CEO Mustafa Suleyman’s June ‘warning’The move follows a clear warning issued in June by Microsoft’s AI division chief, Mustafa Suleyman, who stated that the company was actively trying to reduce its financial dependence on Anthropic.Every time a user asks an AI assistant to write an email or analyse data, it consumes massive quantities of “tokens” – the basic unit of measurement for AI computing power.

Currently, Microsoft’s workplace assistant, Copilot, consumes these tokens at an extraordinary rate. While Microsoft currently enjoys steep discounts on this computing power due to its multi-billion dollar partnership with OpenAI, executives know that special pricing agreement won’t last forever, the report said.

Suleyman’s team is working to ensure Microsoft isn’t left at the mercy of outside AI labs' pricing models.Eliminating the outside costsAt its annual Build developer conference in June, Microsoft debuted seven new in-house AI models. The company claims one of its new models can match the coding capabilities of Anthropic’s highly popular, prior-generation Opus 4.6 model – but at a significantly lower operational cost.“We pay a lot of money to Anthropic — so our goal is to reduce and ultimately eliminate that cost,” Suleyman stated at the conference.The rollout of Microsoft’s MAI models is already expanding beyond basic office work, the report noted.

The home-grown tech is now available within GitHub Copilot for software developers, and Suleyman confirmed that a Microsoft-built voice transcription model will be integrated into Microsoft Teams videoconferencing and other core products in the coming months.Get the latest technology news and updates.

Download the TOI App.

This article was originally published by Times of India ↗. citations.press indexes the source-backed facts above and links to the original. Something wrong? Corrections policy · Report an error