The AGI escape hatch is gone: Microsoft and OpenAI rewrite their deal

Microsoft and OpenAI’s relationship has been weird for a while now. It started as a straightforward investment and cloud deal, then turned into whatever you call that phase where one side keeps pouring billions in while the other tries to date other cloud providers on the side. This morning, the two companies finally acknowledged the awkwardness and rewrote the terms.

The headline change: the AGI clause is dead.

For years, that clause was the nuclear option in their contract. It said that once OpenAI achieved artificial general intelligence — a system smart enough to do most economically valuable work — Microsoft’s exclusive license to OpenAI’s tech would evaporate. The reasoning was that AGI shouldn’t be locked behind a single corporate partner. It was noble in theory, but in practice it created a weird incentive: Microsoft had every reason to downplay any AGI claims, because admitting OpenAI had built AGI would cost them their exclusive access.

Now that clause is gone. No more philosophical debates about whether GPT-6 or whatever qualifies as AGI. The deal is simpler and more commercial.

Here’s what stays: Microsoft remains OpenAI’s “primary cloud partner.” OpenAI products still ship first on Azure, unless Microsoft can’t or won’t support the needed capabilities. That last bit is new and interesting — it gives OpenAI an escape hatch if Microsoft drags its feet on some future infrastructure requirement.

Here’s what changes: OpenAI can now sell its products through any cloud provider. That’s a big deal. Previously, if you wanted to run OpenAI models in production, you were basically forced onto Azure. Now OpenAI can court enterprises on AWS, Google Cloud, or whoever else shows up with a checkbook. This is OpenAI’s play to become a proper platform company rather than just a feature living inside Microsoft’s stack.

I’ve seen this pattern before with other tech partnerships. The exclusive cloud deal, the strategic investment, the joint press conferences — it always looks great on stage, but eventually the startup outgrows the arrangement. OpenAI is now big enough and cash-rich enough that it doesn’t need to be Microsoft’s pet project.

What nobody’s saying out loud is that this also lets Microsoft off the hook a bit. The AGI clause was a ticking bomb in their financial reporting. If OpenAI ever declared AGI, Microsoft would have to explain to shareholders why their exclusive access just vanished. Now that uncertainty is gone. Microsoft can keep selling Azure credits to OpenAI without worrying about some existential trigger clause.

The enterprise customers I’ve talked to are split on this. Some like the flexibility of choosing their cloud provider. Others are nervous that OpenAI will spread itself too thin and lose focus. I lean toward the former — competition usually makes these services better and cheaper.

One thing that bothers me: the original AGI clause was one of the few pieces of forward-thinking governance in the AI industry. It acknowledged that AGI shouldn’t be a proprietary product. Now it’s just another cloud reseller agreement with extra steps. I get why both companies did it — the clause was creating more friction than value — but it feels like a missed opportunity to set a precedent.

For now, the practical impact is straightforward. If you’re an enterprise, you’ll soon have more options for where to run OpenAI’s models. If you’re an investor, you can stop worrying about AGI definitions messing up contract terms. And if you’re Sam Altman, you just got a lot more freedom to play the field.

The AGI escape hatch is gone. Whether that’s progress or a retreat depends on how much you trust market forces to handle something as weird as artificial general intelligence.