Microsoft appears to be getting more aggressive about its own AI products, and that means taking shots at the competition. According to TechCrunch, an internal strategy meeting held earlier this week saw Microsoft executives coach their sales team on how to position the company’s in-house AI models as better, faster, and cheaper than those from OpenAI, Google, and Anthropic.
The meeting was framed as a fiscal year 2027 kickoff session. Executive Vice President Jay Parikh reportedly told the room: “Everyone else is selling parts, we’re selling the full end-to-end system. That’s the story that we all need to get out there and tell in FY27.” Another exec, Jacob Andreou, reportedly went further, presenting a direct comparison between Microsoft’s Copilot and Anthropic’s Claude, arguing that Claude was “slower and less accurate, and lacked the proper security integrations” when used inside Microsoft’s office apps.
Microsoft and Anthropic have not commented publicly. TechCrunch said it reached out to both companies for a response.
A company training its sales team to pitch against rivals is nothing new. What makes this story worth watching is who Microsoft is now targeting. For years, OpenAI and Anthropic were not competitors, they were suppliers. Microsoft built much of its AI product lineup on top of their models, particularly OpenAI’s GPT series. The relationship with OpenAI was especially close, structured around a deal that gave Microsoft exclusive API access in exchange for billions in capital and computing resources.
That deal changed in April, when the two companies revised their partnership and dropped the exclusivity clause. OpenAI can now sell directly to Microsoft’s competitors. That shift appears to have accelerated Microsoft’s push to build and promote its own models rather than rely on outside providers.
This is not the first sign of that shift. Earlier this month, reports surfaced that Microsoft had quietly swapped out OpenAI and Anthropic models inside flagship apps like Word and Excel, replacing them with its own in-house alternatives. The reason cited was cost. Running your own models is cheaper than paying per token to a third-party provider, and at Microsoft’s scale, that difference adds up fast.
The financial pressure is real. Microsoft has been spending heavily on AI infrastructure, and investors have grown skeptical about when those investments will start paying off. The company’s stock performance over the past year has reflected that uncertainty. Coaching sales teams to talk up the quality and value of Microsoft’s own AI products is partly a commercial move, but it is also a message to the market: the spending is working, and the products can compete.
The broader picture here is a significant shift in how the AI industry is organizing itself. A year ago, the story was about which foundation model was most powerful. Now it is about who controls the full stack:
- Microsoft wants to own the model, the app, the infrastructure, and the enterprise relationship
- Google is doing the same thing with Gemini across Workspace and Cloud
- Amazon is pushing its own Titan and Nova models through AWS while also reselling Anthropic
- Anthropic and OpenAI are now competing directly for enterprise customers, including ones that used to reach them through Microsoft
For enterprise buyers, this creates real choices that did not exist 18 months ago. Microsoft’s pitch, based on what was reportedly said in that meeting, is that going with a single integrated vendor is safer, more secure, and ultimately cheaper than mixing and matching models from different providers. Whether that argument lands will depend a lot on whether Microsoft’s own models can actually match the performance benchmarks that OpenAI and Anthropic have set. That is a question the market will answer over the next year.
