Microsoft threatens OpenAI and Amazon with legal action amid the cloud wars

For several years, Microsoft and OpenAI have been working hand in hand in the field of artificial intelligence.

But this balance could be disrupted, because behind the scenes, tensions are rising around OpenAI's cloud strategy, which is now being courted by other tech giants. This situation reveals deep disagreements about the future of the partnership…

An agreement with AWS that threatens Azure exclusivity

At the heart of the conflict is the recent closer relationship between OpenAI and Amazon Web Services (AWS). The two companies have concluded a massive agreement, estimated at several $38 billion, aimed in particular at distributing Frontier, OpenAI's new AI agent platform.

This partnership directly challenges a central clause of the agreement between OpenAI and Microsoft, because it stipulates that access to AI models must be exclusively through Azure. A strategic condition for Microsoft, which has invested heavily in OpenAI and integrated its technologies into its own services. The point of contention lies in the technical interpretation of this exclusivity, and Microsoft considers that any use of the models, including in an AWS environment, violates the agreement. For its part, as reported by Reuters, Frontier, a symbol of a deliberate multi-cloud strategy, is at the heart of all the tensions between the two companies. Presented as a solution for businesses, it allows for the creation and deployment of autonomous agents capable of retaining memory and context over time.

To achieve this, OpenAI is reportedly working with AWS on a stateful runtime environment, notably via AWS Bedrock. This approach would allow AI agents to interact persistently with internal company data, a major evolution compared to traditional, so-called "stateless" APIs.

It is precisely this distinction that fuels the disagreement, because OpenAI considers that these new environments do not fall under the exclusivity granted to Azure. Microsoft, on the other hand, believes that these systems still rely, in practice, on its technologies and should therefore remain under its control.

A legal threat that reveals a rift…

Faced with this situation, Microsoft is raising its voice, and according to the Financial Times, the company is not ruling out taking legal action against OpenAI and AWS if it believes its contractual rights are being violated.

Meanwhile, OpenAI is displaying a certain serenity in the face of these threats, relying on a more flexible interpretation of its agreement with Microsoft. Furthermore, discussions are reportedly still underway between the various parties to avoid legal escalation. Beyond the legal aspect, OpenAI now seems to want to break free from its dependence on Azure by adopting a multi-cloud approach. This direction could redefine the balance of power in the AI ecosystem, where alliances are formed and dissolved as industrial stakes intensify.