- OpenAI and Microsoft end exclusivity in the cloud and in the sale of models, keeping Azure as a priority but not the only partner.
- Microsoft stops paying royalties to OpenAI, while the startup will continue to pay a limited share of revenue until 2030.
- The new pact opens the door to in-depth agreements with Amazon, Google and other cloud providers, strengthening a multi-cloud AI ecosystem.
- The alliance remains, but with more flexibility, which could facilitate a future IPO for OpenAI and reshape the AI competition.
The breaking of the exclusivity agreement between OpenAI and Microsoft marks a Before and after in the race for artificial intelligence in the cloudWhat for years was a nearly closed alliance, with Azure as OpenAI's sole major technological pillar, has transformed into a much more flexible agreement, open to rivals like Amazon Web Services (AWS) and Google Cloud. It's not a complete separation, but it is a huge strategic shift that changes the balance of power in the sector.
After several tensions, veiled threats of legal action, and OpenAI's moves toward other providers, both companies have opted to redefine their relationship to simplify the contractual part and gain room to maneuverMicrosoft no longer holds exclusive control over OpenAI's business models; OpenAI can now deploy its products on any cloud platform, and revenue streams between the two parties are being reconfigured. All of this is happening while maintaining multi-million dollar investments, long-term intellectual property licenses, and a continued central role for Azure.
What was the original agreement between OpenAI and Microsoft and why has it fallen short?
For years, the agreement between the two firms was based on a Very clear exclusivity: OpenAI relied solely on Azure, and Microsoft was the only one that could sell its modelsMicrosoft invested in OpenAI in 2019 with an initial investment of $1.000 billion and, following the success of ChatGPT, increased its commitment to approximately $13.000 billion. This investment enabled OpenAI to develop models such as GPT-3, GPT-4, and successors, while Microsoft rapidly integrated this technology into its ecosystem, particularly in Copilot and its cloud services.
The economic design of the agreement was peculiar: Microsoft obtained up to 49% of OpenAI's economic value, but without formal ordinary shares.In other words, he participated in the financial returns without controlling the company like a traditional shareholder. In return, Sam Altman's startup committed to hosting ChatGPT and the rest of its models exclusively on the Azure cloud, and both parties shared a percentage of the revenue generated by the AI products each marketed.
That scheme worked well at first, when the demand for computing was high but manageable and Microsoft saw Azure sales skyrocket thanks to the exclusivity of OpenAI models.However, as generative AI took off, infrastructure needs grew dramatically. Training and serving increasingly large models required more data centers, more specialized chips, and ultimately, more providers.
OpenAI then began to sound out other partners, such as Google or Amazon, trying Diversify access to computing power to avoid dependence on a single providerThe problem was that these moves clashed head-on with Microsoft's exclusivity clauses regarding cloud usage and the commercialization of its models. Every attempt to open the door to a new partner was accompanied by friction, renegotiations, and, according to several reports, even warnings of potential lawsuits.
The new agreement: end of exclusivity but continuity of the alliance
Finally, both parties have opted for a compromise: break the exclusivity, but maintain close collaboration and economic incentivesMicrosoft remains OpenAI's primary cloud partner, but it is no longer the only one. The new contract wording specifies that OpenAI products will launch first on Azure, unless Microsoft is unable or unwilling to provide the necessary capabilities. From then on, OpenAI is free to operate with any provider it chooses.
In parallel, Microsoft has relinquished the exclusive right it held to resell and market OpenAI's AI models to third parties from your cloudThis exclusivity was one of the key drivers of Azure's growth in the early stages of the AI boom, because it forced any company interested in these models to rely, in one way or another, on Redmond's infrastructure. With the new agreement, OpenAI can sign similar deals with other cloud giants, and its models can appear on rival platforms.
Despite these changes, the joint statement insists that the alliance remains key: both companies emphasize that the The new structure provides greater predictability and at the same time more flexibility to explore joint opportunities.For example, there is talk of expanding the capacity of shared data centers by gigawatts, developing next-generation chips collaboratively, or applying AI to strengthen global cybersecurity.
The tensions don't disappear magically, but the framework is less rigid. The amendment seeks to prevent any new project with third parties from turning into a contractual battle.The idea is for Microsoft to continue benefiting from OpenAI's growth, but without blocking the startup's expansion into other cloud ecosystems, something essential if it wants to continue scaling rapidly.
Economic changes: who pays whom and for how long
One of the most delicate elements of the new pact is the reconfiguration of how revenue generated by AI products is distributedPreviously, Microsoft and OpenAI paid each other royalties on sales derived from each other's technology. With the new model, Microsoft will no longer pay OpenAI royalties for the sale of products built on its models through Azure.
OpenAI, on the other hand, will continue paying Microsoft a portion of its revenue until 2030This is presumably around 20%, as previously estimated, although now with a maximum cumulative limit whose amount has not been made public. In other words, the startup still has to compensate Microsoft for several years, but with a cap that reduces long-term financial uncertainty.
This is combined with the existing investment structure: Microsoft retains around a 27% stake in OpenAI, resulting from the company's reorganization as a for-profit companyAlthough it is not a controlling stake in the classic sense, it does guarantee Microsoft a privileged position in the future distribution of the company's value.
In addition, there is a key intellectual property component. Microsoft maintains a license on OpenAI IP related to AI models and products until 2032The key difference is that this license is no longer exclusive, meaning OpenAI can offer the same technology to other companies simultaneously. Previously, Microsoft had a temporary lock on access to the most advanced versions, allowing it to get ahead of competitors in integrating new capabilities.
Another relevant point has to do with the so-called artificial general intelligence (IAG or AGI)Until now, a clause existed that limited Microsoft's access to OpenAI's cutting-edge technology in the hypothetical case that the board of directors deemed it had reached a level of intelligence comparable to human intelligence. Leaked information suggests that the recent review has softened this aspect, eliminating the obligation for Microsoft to make specific decisions should that technological frontier be reached, although the details of that section remain confidential.
The entry of Amazon, Google and other giants: towards an open ecosystem
The major practical consequence of the end of exclusivity is that OpenAI It can now close full agreements with other cloud providers such as Amazon or GoogleIn fact, the rapprochement with Amazon had been brewing for some time and was one of the main sources of friction with Microsoft before the renegotiation of the agreement.
In February, a massive alliance between OpenAI and Amazon was announced, which includes a potential investment of up to $50.000 billion by the e-commerce giantIn parallel, a previous cloud services agreement with AWS, already worth around $38.000 billion, was expanded to approximately $138.000 billion over eight years. In this context, Amazon became a key partner in both financing and infrastructure.
One of the most striking elements of that collaboration is that Amazon will exclusively distribute the OpenAI Frontier AI agent platform on its cloud platform.The idea is for AWS to become the primary home for certain advanced automation and artificial intelligence products from Altman's company, thereby strengthening the appeal of its ecosystem compared to other providers.
Until the renegotiation with Microsoft, these moves clashed with existing exclusivity clauses. In fact, it was even reported that Microsoft He seriously considered taking legal action against Amazon and OpenAI over that $50.000 billion cloud deal, interpreting it as a potential violation of their exclusive intellectual property rights to OpenAI. The new agreement eliminates this head-on clash and normalizes OpenAI being shared among several providers.
The door also remains open to a Google Cloud's increased role in the deployment of OpenAI modelsAlthough the relationship between OpenAI and Google is both competitive and collaborative, the elimination of cloud exclusivity allows companies already operating on Google Cloud to directly integrate OpenAI products without having to go through Azure as an intermediary layer, simplifying architectures and reducing dependencies.
Market reactions, internal tensions and the future of the alliance
The announcement of the revised agreement has not gone unnoticed in the markets. After it was made public, Microsoft shares fell around 1-2% in pre-market trading.This reflects some investors' fears that the company could lose a key competitive advantage in the AI race. After all, for the past few years, preferential access to OpenAI models has been one of Azure's biggest selling points.
In that same context, Amazon's stock registered more moderate declines, less than 1%This is logical considering that the new scenario reinforces the cloud giant's position as an essential partner for OpenAI. In the medium term, market perception will depend on whether Microsoft can continue to capitalize on its relationship with OpenAI while expanding its own internal model development, and to what extent Amazon and other players manage to capture a share of the market.
On a strategic level, the general feeling is that The alliance between the two companies has become too large and complex to continue being based on rigid exclusivity agreements.Back in the recent past, when a non-binding memorandum of understanding was signed to redefine the collaboration, many analysts assumed that a major change was imminent. That preliminary document served as the basis for the negotiations that have led to the current agreement.
There is also a relevant corporate angle: with a more flexible agreement and fewer ties to Microsoft, OpenAI gains ground to consider a future IPOBy easing some restrictions on where it can operate its products and how revenue is distributed, the company can better structure its business model for public investors. However, from now on it will have to bear the cost of hosting on Azure and other providers like any other customer, instead of enjoying the advantageous deal it had initially.
For Microsoft, the risk is that Copilot and other AI services should stop relying so heavily and so soon on GPT.This opens the door to integrating alternative models from third parties or its own. The company cannot ignore OpenAI, but neither can it tie its hands if its rivals begin to deploy equally or more competitive models. It is not unreasonable to imagine a future in which Microsoft combines OpenAI models with others developed internally or even with technologies from other vendors, although integrating direct rival tools like Google's Gemini would still be politically complicated.
Underlying all this movement is a general trend in the sector: moving from closed alliances to more open and multi-cloud ecosystemsComputing needs, regulatory pressure, and companies' desire to avoid being locked into a single provider are all driving this trend. The new agreement between OpenAI and Microsoft fits perfectly into this paradigm shift and, incidentally, reshapes the competitive landscape between Azure, AWS, and Google Cloud.
With all of the above on the table, the feeling that remains is that The relationship between OpenAI and Microsoft is not broken, it is transformed.The exclusivity that enabled their rise in generative AI is diminishing, but in return, OpenAI gains much more freedom to work with other cloud giants and Microsoft gains much more freedom to rebalance how it monetizes collaboration. The future of commercial AI lies in increasingly flexible agreements, multiple clouds competing to host the most advanced models, and a handful of companies simultaneously acting as partners and rivals in one of the most strategic markets of the digital economy.