Microsoft plans to stop sharing revenue with OpenAI, according to Bloomberg. A major shift in one of tech's most watched partnerships. But the details are murky. Bloomberg's paywall means we're working from limited information, and commenters are already flagging contradictions in the reporting.
The original 2019 deal used a "capped profit" structure, not a standard equity investment. Microsoft put in $1 billion and would collect profits until hitting a cap estimated around $20 billion, roughly 20x the initial investment. Some earlier reporting said profit sharing would continue until OpenAI achieved AGI, though what counts as AGI was never defined. Microsoft has since poured in another $10 billion (2023) and OpenAI's valuation has exploded. Something had to give.
There are a few ways to read this. Maybe Microsoft hit its cap and the original terms are playing out as designed, or OpenAI got enough clout to force a renegotiation. The original terms might have been reported wrong from the start either way. The Hacker News discussion raises another point worth considering: the story you read probably depends on which side leaked. Take coverage with that in mind.
What's clear is the power dynamic has moved. OpenAI isn't the scrappy lab Microsoft first backed. With real revenue and a massive valuation, it can push back on unfavorable terms now. That said, as openings to wield more influence appear, the risk of misjudged power moves increases. The new arrangement reflects where the power actually sits.
This mirrors Microsoft's recent decision to label Copilot as 'entertainment only' in its terms of use, signaling a broader shift away from blanket liability. The new arrangement reflects where the power actually sits.