Microsoft and OpenAI end exclusive partnership: OpenAI will be available on AWS Bedrock, revenue sharing agreement terminated

Core conclusion

On April 27, 2026, Microsoft and OpenAI officially ended their seven-year exclusive partnership and revenue sharing agreement. This decision means that OpenAI will no longer exclusively use Microsoft Azure cloud services for model training and inference, and Microsoft will no longer receive a share of OpenAI’s revenue. What’s even more noteworthy is that OpenAI’s models are about to land on the AWS Bedrock platform—this marks that the AI ​​cloud service market has entered a new multi-cloud competitive landscape.

Key Points

• Time of incident: April 27, 2026 (Exclusive report by Bloomberg) -Affected: All developers and enterprise customers using Azure OpenAI services
• Core changes: OpenAI gains freedom to use non-Microsoft cloud infrastructure, and Microsoft no longer enjoys OpenAI revenue share

Background and trigger events

The cooperation between Microsoft and OpenAI began in 2019. Microsoft has invested a total of approximately US$13 billion in exchange for exclusive commercial licensing and revenue sharing rights for OpenAI technology. In exchange, OpenAI must run all of its training and inference workloads on the Microsoft Azure cloud.

However, as OpenAI’s commercialization accelerates—ChatGPT’s monthly active users have exceeded 400 million, and its API services cover more than 90% of the world’s top 500 companies—this exclusive arrangement has gradually become a constraint on the continued development of both parties.

After the report was released, the OpenAI official blog simultaneously published the article "Next Phase of Microsoft Partnership" (the next phase of Microsoft partnership), confirming this shift. Amazon CEO Andy Jassy later responded on X: "OpenAI's announcement this morning was very interesting. We will make OpenAI models available directly on Bedrock in the coming weeks."

Key Impact (by Dimension)

Adaptation suggestions

In response to this change, content producers and AI tool developers should take the following actions immediately:

• Check cloud dependencies in existing workflows: If your automation pipeline relies heavily on Azure OpenAI, you can now plan a multi-cloud backup solution
• Follow the launch date of Bedrock OpenAI: The OpenAI model on AWS Bedrock is expected to be launched in the next few weeks. It is recommended to activate the Bedrock preview in advance.
• Compare API pricing: Multi-platform competition often brings room for price reductions. It is recommended to continuously monitor the API pricing adjustments of each platform.
• Updated internal link strategy: This news related tools include OpenAI, Claude, n8n, AWS Bedrock, and the trigger tool floating card can be mentioned naturally in the text

Task List

• Add multi-cloud API switching logic to the automated pipeline
• Compare the pricing differences between Azure OpenAI and AWS Bedrock OpenAI
• Follow the follow-up updates on whether OpenAI adopts Google TPU (Gen 8)

Reactions from all parties and future prospects

Judging from community discussions, developers generally believe that this change will have a profound impact on the industry. The top comment on HN noted: "The real winner may be Google - almost all cutting-edge AI labs are using TPUs, and the only one that is useless is OpenAI because of the exclusive agreement with Microsoft. In the context of Google's Gen 8 TPU just released this month, OpenAI is likely to consider using TPUs."

Amazon CEO Andy Jassy's response confirmed AWS's rapid action: "We are pleased to provide OpenAI models directly on Bedrock, as well as the upcoming Stateful Runtime Environment. Developers will have more choices."

It is worth noting that this change coincides with the upcoming legal battle between OpenAI and Elon Musk - Musk is suing OpenAI for deviating from its original intention of being a non-profit. The two things are superimposed, and the future direction of OpenAI will become one of the most watched narratives in the AI ​​industry in 2026.

Further reading

Google DeepMind paper: LLM will not have consciousness

Also this week, Google DeepMind senior scientist Alexander Lerchner published a paper stating that "phenomenal consciousness" is a physical state and "is not a software artifact that can be created accidentally or deliberately." DeepMind CEO Demis Hassabis' previous statements about the potential of AGI sparked discussion. After 404 Media reported that DeepMind has removed its brand logo from the paper and moved the "personal opinion" statement to the top, clearly distancing itself from the relationship.

This means that the discussion on the growth of AI capabilities is shifting from "can it be done" to "should it be done" - for content practitioners, the more powerful the tool, the higher the requirements for those who use the tool.

Taylor Swift lays out trademark to protect AI rights

Taylor Swift's company has filed multiple sound trademark applications, including catchphrases such as "Hey, it's Taylor Swift" and a trademark for a specific performance photo. IP attorney Josh Gerben noted that celebrity voice trademarks "have not yet been tested" in court. This incident shows that when AI synthesis capabilities exceed legal protection, creators have begun to use legal means to build a moat - an important signal for content producers.

Tool entry

The following tools are involved in the text, and the platform side will automatically match the maintained tools library to trigger the floating card:

• OpenAI, ChatGPT, Claude, DeepSeek, Gemini, n8n, LangGraph, Hermes Agent

Internal link guidance

• Want to build your own AI automated workflow? See: Claude Code automated writing practice: build an AI content production pipeline in 30 minutes
• AI Agent introductory practice: AI Agent Tools 2026 Complete Tutorial: 5 Tools to Build an Automated Pipeline in 30 Minutes
• Someone helped us make money using AI Agent: Claude Code 48 hours to start a business: one person + US$29 monthly fee, monthly income in 3 months $9,000