OpenAI and Microsoft built one of the most closely watched alliances in artificial intelligence. Now, a reported contract clause about artificial general intelligence is putting a sharper focus on who controls OpenAI's technology, and when that control could change.
The New York Times, citing 19 people familiar with the companies' relationship, described a partnership under strain. According to the report, the relationship has cooled because of financial pressure on OpenAI, limited computing power from Microsoft, and disagreements over ground rules.
A partnership under pressure
Microsoft is described in the source as OpenAI's investor, partner, and increasingly, rival. That mix helps explain why the relationship is unusually complex. A company can support another company, depend on it, compete with it, and still argue over the limits of the arrangement.
The reported tension is not presented as one single dispute. Instead, the source points to several overlapping pressures. OpenAI faces financial pressure. Microsoft is providing limited computing power. The two companies also disagree about the rules that should govern their work together.
Those details matter because artificial intelligence development depends on both technology and access. If one company builds systems and another provides critical infrastructure or investment, the contract between them becomes more than a business formality. It becomes part of the power structure around the technology itself.
The reported AGI clause
The most striking detail is a reported clause in OpenAI's contract with Microsoft. According to the source, the clause would cut off Microsoft's access to OpenAI's technology if OpenAI develops artificial general intelligence, or AGI.
The source defines AGI as an AI system capable of rivaling human thinking. That definition is central to the issue because it describes a category of technology that could carry far greater consequences than today's systems. In the reported contract structure, reaching that threshold would not simply mark a technical milestone. It could change Microsoft's rights.
The Times report says the clause is meant to ensure Microsoft can never misuse the technology. That framing puts the clause in the category of control and safety, rather than ordinary commercial leverage. It suggests the agreement was designed with a future scenario in mind: OpenAI reaches a level of capability so significant that access by a major partner must be restricted.
For readers trying to understand the stakes, the reported mechanism can be simplified into three linked points:
- OpenAI and Microsoft have a contract governing access to OpenAI technology.
- The contract reportedly includes a clause tied to the arrival of AGI.
- If OpenAI develops AGI, Microsoft's access to OpenAI technology could reportedly be cut off.
That is why the clause stands out. It creates a possible escape route from a powerful partnership, but only if OpenAI reaches a disputed and difficult-to-define technological threshold.
Who decides when AGI has arrived?
The reported clause becomes more complicated because OpenAI's board can reportedly decide when AGI has arrived. That gives the board an important role in interpreting the moment that would trigger the contract's consequences.
Sam Altman has already said that the arrival of AGI will be somewhat subjective. The source also notes that he described the shift toward AGI as likely to be blurrier and more gradual than many people expect.
That matters because a contract trigger usually works best when the triggering event is clear. A date can be checked. A payment can be verified. A delivery can be confirmed. But a judgment about whether an AI system rivals human thinking may be harder to pin down, especially if the transition is gradual rather than sudden.
If OpenAI's board has the authority to make that call, the contract is not only about technology. It is also about governance. The board's interpretation could determine when Microsoft's access changes, assuming the reported clause operates as described.
Why the clause changes the Microsoft question
The reported clause reframes the OpenAI-Microsoft relationship. Microsoft is not only an investor or partner in this account. It is also a company whose access to OpenAI's future technology could be limited under specific conditions.
That makes the relationship different from a simple funding story. The source describes a five-year romance that has cooled, but the deeper issue is control over advanced AI. If OpenAI develops AGI and its board says AGI has arrived, the reported clause could place a boundary between OpenAI's most important technology and Microsoft.
The same facts also show why tensions could grow. OpenAI needs resources. Microsoft provides support but is also increasingly a rival. The companies disagree about ground rules. In that setting, a clause that could cut off access becomes more than a technical legal detail.
It is also important to keep the limits of the reporting clear. TechCrunch said it asked OpenAI for comment. The source does not say that AGI has arrived. It does not say the clause has been triggered. It reports that such a clause exists and that the board can decide when the AGI threshold has been met.
The bigger signal
The reported OpenAI Microsoft AGI clause shows how artificial intelligence partnerships can contain built-in conflict. The same deal can support rapid development while also reserving the right to restrict access if the technology becomes powerful enough.
For OpenAI, the clause reportedly protects against misuse of technology. For Microsoft, it introduces uncertainty around long-term access. For everyone watching the AI industry, it highlights a broader question: who gets to decide when an AI system has crossed a line that changes the rules?
Based on the source, the answer may rest with OpenAI's board. That is what makes the reported clause so consequential. It connects technical progress, corporate governance, and partner access in one decision point.