Microsoft and NVIDIA Disagree About the Arrival of AGI

NVIDIA CEO Jensen Huang recently made waves when he claimed humanity has already achieved artificial general intelligence (AGI). In an interview on the Lex Fridman podcast, Huang was asked for a timeline on when an AI could run a tech company and generate millions of dollars. 

His response was bold and immediate: “I think it’s now.” While that makes for a great headline, the legal team at Microsoft might be sweating over that specific definition.

The billion-dollar legal loophole is hiding in their contract

For most people, AGI means an AI model that can adapt, think, and solve complex problems across different subjects just like a real human. However, the legal definition between Microsoft and OpenAI is much more specific, and millions of dollars hinge on it.

When Microsoft poured billions into OpenAI, they baked a crucial escape clause into their partnership. If OpenAI officially achieves AGI, they no longer have to give Microsoft exclusive access to their future, most powerful models.

According to their specific contract, as explained by Windows Central, OpenAI can only trigger this clause if they create a model capable of generating $100 billion in profit. Right now, even the best AI models are losing money, not making it. Notably, the NVIDIA CEO also addressed concerns about the DLSS 5 enhancements recently.

The AI chip maker and research company are setting the stage for a messy breakup over AGI

The problem is that NVIDIA and OpenAI seem eager to shift the goalposts on what actually counts as AGI. If OpenAI’s board decides they have met the philosophical benchmark for AGI, even without hitting the massive profit requirement, it could spark a massive court battle.

Microsoft is already quietly preparing for a future where they rely less on OpenAI, restructuring its own internal teams to build new tools.

If the two tech giants cannot agree on when the future has officially arrived, a very public and very messy breakup feels inevitable.