Microsoft shifts thousands of Office prompts to in-house AI models to slash soaring costs

Microsoft joins AI cost-cutting trend by relying more on its own models

As AI costs continue to rise, companies are looking for ways to cut back. The most recent example is Microsoft, which has reportedly begun to deploy a cost-savings strategy by relying less on software from OpenAI and Anthropic and instead deploying its own in-house models. Indeed, when it comes to two of its most widely used programs -- Excel and Word -- Microsoft has begun to use its homemade MAI models to respond to a certain percentage of user prompts, Bloomberg reported Tuesday. In the past, the company had advertised the fact that large parts of Office 365 are powered by models from both OpenAI and Anthropic. While Microsoft still relies on those third-party models, it has also increasingly sought to stand up its own AI agents. Last month, at its annual Build conference, the company announced the launch of seven new MAI models, including an agentic coder and a text-to-image generator. When reached for comment by TechCrunch, Microsoft said that it had nothing further to share. Microsoft's apparent cutbacks are part of a broader trend. After a brief blitz of "tokenmaxxing" earlier this year, the last few months have seen a news cycle awash in stories about tech companies acting significantly more thrifty. Other large companies -- like Amazon, Uber, Meta, and Accenture -- have also reportedly made moves to curb spending. The immense cost of providing and buying AI services has become a controversial part of the industry. The sticker shock has gotten so bad in some parts of Silicon Valley that some companies are reportedly looking to Chinese models for more affordable agentic solutions -- despite some concerns over potential security issues.

Microsoft 365 Apps Favoring Company's AI Models Over ChatGPT, Claude

To cut costs, Microsoft is pushing to have Microsoft 365 AI prompts answered by its internal MAI models rather than by Anthropic and OpenAI models, according to Bloomberg. Although Microsoft has historically championed the use of Claude and ChatGPT in its Microsoft 365 Office applications like Excel and Outlook, it has begun funneling at least some of the AI prompts from those tools to its own AI models. It's a small share of overall prompts, and Microsoft previously said it's working on using its own models to help reduce monthly token costs, but this is the first clear sign of how it might do that. Microsoft's CEO of AI, Mustafa Suleyman, said during the June Build conference that the plan is to "reduce and ultimately eliminate" the amount of money it gives to Anthropic. Microsoft's MAI models are already available in Copilot for Business and Enterprise users. MAI-Code-1-Flash is Microsoft's coding assistant tool, designed to compete with Claude Code and OpenAI's Codex. Microsoft has also announced plans to release its own Teams-based transcription model. This is just the latest pullback by external AI providers, including Microsoft, in 2026. Earlier this year, it began reducing its workers' access to external tools such as Anthropic's Claude Code after costs skyrocketed. It later moved GitHub Copilot users to token-based billing and pulled its employees' access to Claude Fable over data-retention concerns. It also joins a growing trend of even major tech companies reducing their AI spend as token costs spiral. Uber reportedly spent its entire annual AI budget in the first three months of this year, and an undisclosed company spent half a billion dollars on AI tokens in a single month due to uncapped engineer limits. AI tokenmaxxing leaderboards have quickly fallen out of fashion. Microsoft pushing everyone toward its own models makes costs easier to manage, even if the return on investment is still almost impossible to quantify.

Microsoft swaps in its own AI over OpenAI in some apps

Microsoft has begun swapping OpenAI and Anthropic models out for its own in-house MAI models in some app features where cost or data residency favours them, per Bloomberg. The shift is incremental, OpenAI and Anthropic still handle most Copilot traffic, but it marks Microsoft steadily reducing a dependence it spent years and billions building, enabled by the 2025 renegotiation that freed it to build competing models. Microsoft has begun replacing OpenAI and Anthropic models with its own AI in some product features, Bloomberg reports. The shift routes selected tasks to Microsoft's in-house MAI models where cost or data residency favours them. The change is incremental rather than a clean break. OpenAI and Anthropic still handle most production traffic inside Copilot, with MAI slotting in where its economics stack up. Microsoft unveiled seven MAI models at its Build conference, including its first reasoning model and image, voice, and transcription systems. It has been testing MAI-Transcribe-1 across Teams and Copilot, and rolling MAI-Image-2 into Bing and PowerPoint. The pitch is efficiency, as Microsoft can run these models on its own Azure infrastructure and skip paying third parties. It says one MAI model, tuned for consulting firm McKinsey, beat OpenAI's GPT-5.5 on cost efficiency by a factor of ten. None of this signals a full divorce. Microsoft launched the MAI line as a direct challenge to OpenAI while keeping its partners' models available for the work they do best. The leash comes off For years, Microsoft's original OpenAI contract barred it from independently pursuing frontier AI. That changed when the two firms renegotiated and ended Microsoft's exclusivity, freeing it to build competing models while keeping a licence to OpenAI's technology through 2032. The deal cut both ways, letting OpenAI sell through rivals so that AWS could offer its models too. Microsoft's answer has been a three-way hedge, holding a large OpenAI stake, embedding Anthropic's Claude in Copilot, and now shipping its own MAI models. The strategic logic is one Satya Nadella has hinted at directly, having reportedly feared Microsoft becoming "the next IBM" if it leaned too hard on a single partner. Owning the model, not just renting it, is how it avoids that fate. A distribution game, not a benchmark race Microsoft does not need MAI to top every leaderboard, since its reach into hundreds of millions of Office and Teams seats does the heavy lifting. Shifting even a slice of that traffic to first-party models moves real money. That reach comes with a nagging problem, as Microsoft is pushing deeper into AI with the paying-customer question still unresolved. Cheaper in-house models help the margins whether or not users pay up. The bill for outside AI is not trivial, and every feature Microsoft brings in-house trims it. The partnership that once defined the company's AI story is being quietly rebalanced, one swapped-in model at a time.

Claude and ChatGPT Are Getting Too Expensive, Even for Microsoft

Using the most advanced AI models at a corporate scale isn't cheap, even for a company as massive as Microsoft. The cost of AI tokens, the unit used to measure how much computing work a model is doing, is getting so expensive that Microsoft is reportedly turning to its own models to save some cash. Bloomberg reports, citing an unnamed source, that tens of thousands of AI prompts each week in Microsoft's Excel and Outlook software are now being completed using Microsoft's own MAI models. The spreadsheet and email programs previously relied more heavily on models from OpenAI and Anthropic to complete some tasks. That is still only a small fraction of Microsoft's overall AI usage, Bloomberg notes. For example, Copilot, the company's workplace AI assistant, requires massive amounts of AI tokens. Microsoft declined to comment on the report. The news comes a little over a month after Microsoft announced seven new in-house models, including MAI-Thinking-1, Microsoft AI's first reasoning model. At the time, the company said MAI-Thinking-1 was built for "high efficiency and performance, but importantly, at a low-token cost." Microsoft describes the model as a mid-sized, 35 billion active-parameter model with a 256K context window. According to a blind test, it matched the coding abilities of Anthropic's popular Claude Opus 4.6. Microsoft also rolled out new image, transcription, voice, and coding models. The new models arrive as cheaper and more efficient AI models are getting more attention across the industry. China's DeepSeek made headlines earlier this year when it realeased new budget-friendly models. For instance, DeepSeek charges $0.435 per million input tokens and $0.87 per million output tokens for its V4-Pro model. While, Anthropic charges $10 per million input tokens and $50 per million output tokens for its most advanced Fable 5 model. "Anthropic is extremely expensive and I think many people are urgently looking for alternatives," Microsoft AI CEO Mustafa Suleyman told Bloomberg last month. "We pay a lot of money to Anthropic -- so our goal is to reduce and ultimately eliminate that cost." OpenAI's pricing is cheaper than Anthropic's, at $5 per million input tokens and $30 per million output tokens for API use of GPT-5.5. Microsoft also enjoys a discount through its partnership with OpenAI. But the clock is ticking on the current deal, which ends in 2032. Still, it looks like Microsoft is trying to get ahead of those pricey costs. Suleyman told Bloomberg that "many, many people in our organization are spending millions of dollars" on AI tokens.

Microsoft Quietly Shifts Thousands of Office Prompts to In-House AI | PYMNTS.com

While there have been previous reports that Microsoft plans to transition to its own AI models, the fact that the company's MAI is now handling tens of thousands of prompts in those applications each week was not previously known, according to the report. Microsoft did not immediately reply to PYMNTS' request for comment. According to the Bloomberg report, the company is shifting to its own models to reduce its AI costs and to avoid being reliant on other AI labs. The report cited previous Bloomberg reports in which the CEO of Microsoft AI, Mustafa Suleyman, said that the company aims to reduce its spending on Anthropic models by using its own and that within months, a Microsoft model will start performing transcriptions in the company's videoconferencing app, Teams, and other products. It was reported in March 2025 that Microsoft was testing AI models from other companies that could replace OpenAI's models in Copilot and that the tech giant was also training artificial intelligence models that could compete directly with OpenAI's. At that time, Microsoft had completed training of a family of models that performed nearly as well as OpenAI's leading models on commonly accepted benchmarks, per that report. A Microsoft spokesperson told PYMNTS at the time: "As we've said, we are using a mix of models, which includes models from OpenAI as part of our partnership, as well as Microsoft AI and open-source models." It was reported in January that Microsoft was increasing its spending on Anthropic, that it had become one of Anthropic's top customers and that it was on track to spend around $500 million per year to use Anthropic's AI models in Microsoft products. PYMNTS reported in April that during the most recent quarter, Microsoft delivered double-digit growth across its core segments, primarily fueled by massive demand for AI and cloud services. At the same time, the company has faced investor concerns over elevated capital spending.

Microsoft edges higher as in-house AI push takes aim at copilot costs By Investing.com

Investing.com -- Microsoft (NASDAQ:MSFT) shares ticked up 1.75% on Tuesday, outperforming a sluggish Nasdaq following reports that the tech giant is beginning to ditch pricey third-party AI models in favor of its own homegrown technology. According to Bloomberg, Microsoft has quietly deployed its newly minted, internally developed MAI models directly into Excel and Outlook. The under-the-hood swap is already handling tens of thousands of AI prompts a week across the spreadsheet and email apps, an anonymous source familiar with the matter revealed. While Microsoft's flagship workplace apps previously leaned heavily on outside pioneers like OpenAI and Anthropic, this marks the first time the true scale of its proprietary MAI rollout has come to light. The tactical pivot signals a critical evolution in Microsoft's AI playbook: aggressively slashing the massive infrastructure costs of artificial intelligence while trying to keep its core software suite running at full steam.