According to Fortune, Microsoft is forming a new MAI Superintelligence Team led by AI chief Mustafa Suleyman, who joined in March 2024. The company was previously barred from pursuing its own AGI research due to its landmark deal with OpenAI, which even capped how large of a model Microsoft could train using FLOPS measurements. Suleyman called this “a big limitation” for a company of Microsoft’s scale. The new team will work toward “Humanist Superintelligence” defined as incredibly advanced AI that always serves people and humanity. Microsoft is extending its OpenAI partnership through 2030 while pursuing this new independent research path. Karén Simonyan will serve as chief scientist of the new team, joining from the same Inflection deal that brought Suleyman to Microsoft.
Microsoft’s Double Game
Here’s what’s really interesting about Microsoft’s move. They’re playing both sides – building their own superintelligence capability while maintaining that crucial OpenAI partnership through 2030. Suleyman calls it a “best-of-both environment,” but it’s really about hedging their bets. If OpenAI delivers superintelligence first, Microsoft gets early access. If they beat OpenAI to it, they control their own destiny. Either way, they’re covered.
And let’s be real – the timing here is everything. Microsoft waited until they had their “self-sufficiency effort” in place before making this move. They’ve been quietly building their GPU stash (though Suleyman won’t say how big), poaching researchers from Google DeepMind, Meta, OpenAI and Anthropic, and establishing what he calls a “culture focused on developing the absolute frontier.” This wasn’t a sudden decision – it’s been in the works for at least a year.
The Superintelligence Marketing Game
Now, let’s talk about this “superintelligence” branding. Microsoft is just the latest company to jump on this bandwagon. Meta rebranded as Meta Superintelligence Labs in June 2025. OpenAI’s Sam Altman says they’re looking beyond AGI to superintelligence. Ilya Sutskever founded Safe Superintelligence. Anthropic has teams working on controlling hypothetical superintelligence.
But here’s the thing – nobody actually has superintelligence yet. Scientists debate whether it’s even achievable with current methods. So why all the announcements? Basically, it’s become the new marketing buzzword for “we’re doing really advanced AI stuff.” Microsoft’s twist is calling it “humanist” superintelligence – a deliberate contrast to what they see as the more purely technological focus of rivals.
The Regulation Dance
Suleyman’s comments on regulation are particularly revealing. When asked about AI “czar” David Sacks pushing for no-holds-barred acceleration, Suleyman agreed… but with a huge caveat. “David’s totally right, we should accelerate,” he said, then immediately added “within the constraints of making sure it doesn’t harm us.”
That’s the tightrope every major AI company is walking right now. They want to move fast and beat competitors, but they also need to appear responsible enough to avoid heavy regulation. Suleyman’s “humanist” framing is part of that balancing act – it sounds safer and more ethical than just chasing raw capability. Whether that translates to actual safety practices remains to be seen, but it’s smart positioning in today’s regulatory environment.
What This Means for AI Competition
So where does this leave the AI race? Microsoft is now fully independent in research while maintaining their OpenAI advantage. They’re building their own team with top talent from across the industry. And they’re framing their mission in terms that might appeal to regulators and the public.
The real question is whether any of these companies can actually deliver on the superintelligence promise. We’re talking about systems potentially wiser than all of humanity combined – that’s a pretty high bar. For now, it’s mostly marketing and positioning. But Microsoft’s move signals they’re serious about competing at the very highest levels of AI development, with or without their partners. The next few years will show whether they can turn that ambition into reality.
