Microsoft Forms MAI Superintelligence Team Under Mustafa Suleyman to Build ‘Humanist Superintelligence,’ Moves Beyond Relying On OpenAI

Highlights

• Microsoft has formed a dedicated division led by Mustafa Suleyman to build “Humanist Superintelligence”.
• Microsoft’s updated agreement with OpenAI lifts previous restrictions on AGI development and allows independent superintelligence research while retaining a 27% stake in OpenAI.
• Suleyman emphasises a human-centered approach, rejecting the AI arms race narrative with a roadmap extending to 2032.

Caption – (Photo by BoliviaInteligente on Unsplash)

Microsoft has officially entered the race to develop superintelligent AI. The tech giant has announced the creation of a new division called the MAI Superintelligence Team, led by Mustafa Suleyman, head of Microsoft’s AI division. The company says it no longer plans to wait for approval or restrictions to pursue this next phase of artificial intelligence development.

Microsoft’s MAI Superintelligence

According to reports, Microsoft’s new mission is to go far beyond achieving artificial general intelligence (AGI), systems that can match human cognitive ability. Instead, the company will develop superintelligence or AI that can outperform humans in every domain including healthcare, clean energy, and materials science.

Caption – Microsoft AI CEO Mustafa Suleyman. (Image credit – Stephen Brashear, Getty Images)

In a blog post, Suleyman explained that while AGI represents AI that can do anything a human can, superintelligence goes “far beyond that performance.” He described Microsoft’s approach as building “Humanist Superintelligence,” a system designed to serve humanity directly. Examples include advanced personal AI assistants, improved medical outcomes, and sustainable energy innovations.

Microsoft-OpenAI New Contract Lifts AI Restrictions

This new direction comes after Microsoft removed a key limitation in its partnership agreement with OpenAI. Previously, Microsoft was restricted from developing its own AGI and from training AI models beyond a set computing power threshold, measured in FLOPS (floating-point operations per second). FLOPS represent the number of mathematical calculations an AI model performs per second and serve as a rough measure of the total computing power used during training.

“For a company of our scale, that’s a big limitation,” Suleyman told Fortune.

The updated contract now lifts these restrictions allowing Microsoft to develop its own AGI and superintelligent systems. It also grants Microsoft a 27% stake in OpenAI’s public-benefit company, signaling a shift from full collaboration to what now appears to be strategic competition between the two tech giants.

Microsoft’s previous agreement with OpenAI not only capped model sizes but also prohibited it from building beyond a certain computational limit. The arrangement that effectively prevented independent AGI research. With the revised deal, those boundaries have been removed and opened the door for Microsoft to pursue more advanced models.

Microsoft’s Superintelligence Push – What We Know So Far

Caption – Microsoft Forms MAI Superintelligence Team Under Mustafa Suleyman. (Photo by Steve Johnson on Unsplash)

Suleyman, 41, joined Microsoft in March 2024 after co-founding Inflection AI, whose technology and intellectual property Microsoft acquired. He previously co-founded Google DeepMind and has since reorganized internal Microsoft AI teams, integrating key staff into the new group and launching what he called an internal “AI self-sufficiency effort” to make Microsoft less dependent on outside partners.

He has also recruited top researchers from Google, DeepMind, Meta, OpenAI, and Anthropic to strengthen the company’s AI division.

Microsoft is simultaneously working to extend its OpenAI partnership through 2030 to maintain early access to OpenAI’s most advanced models and intellectual property, even as it builds independence through its own superintelligence program.

Karén Simonyan, who joined Microsoft alongside Suleyman from Inflection AI, will serve as the chief scientist of the new Humanist Superintelligence team.

Suleyman has also told employees that Microsoft plans to invest heavily in its AI infrastructure. While its current AI models remain smaller than those of OpenAI or Google, the company’s long-term roadmap extends to 2032. It is the year its licensing rights to OpenAI’s models are set to expire.

Trend Toward “Superintelligence” Branding

Microsoft’s new initiative aligns with similar efforts across the AI industry. It basically joins a growing list of tech giants rebranding their AI programs around superintelligence.

Former OpenAI chief scientist Ilya Sutskever recently launched Safe Superintelligence, a startup focused on building and controlling superintelligent AI. Meanwhile, Anthropic continues research into the safe development of advanced AI systems.

In June 2025, Meta renamed its AI division Meta Superintelligence Labs, while OpenAI CEO Sam Altman has claimed that his company is already moving beyond AGI toward superintelligence.

However, Microsoft has taken a distinct philosophical approach. Suleyman emphasised that his team rejects the idea of an AI arms race, writing, “We reject narratives about a race to AGI, and instead see it as part of a wider and deeply human endeavor to improve our lives and future prospects. We also reject binaries of boom and doom; we’re in this for the long haul to deliver tangible, specific, safe benefits for billions of people. We feel a deep responsibility to get this right.”

Despite its focus on ethics and safety, Suleyman said Microsoft will still innovate rapidly. He acknowledged that building a “humanist” superintelligence requires caution with capabilities that aren’t ready for widespread deployment, but insisted that this won’t slow the company’s pace of progress.

Interestingly, a noteworthy promise in Suleyman’s essay is that “everyone who wants one will have a perfect and cheap AI companion helping you learn, act, be productive and feel supported.”

As of now, the concept of superintelligence remains theoretical with no existing system close to achieving it. Having said that, Microsoft’s new team represents one of the most significant corporate pushes yet toward that goal. This evident race could very well shape the next decade of AI innovation.

FAQs

Q1. What is Microsoft’s MAI Superintelligence Team?

Answer. Microsoft has launched the MAI Superintelligence Team under AI chief Mustafa Suleyman to build “Humanist Superintelligence,” AI that surpasses human capabilities across domains like healthcare, energy, and science.

Q2. How has Microsoft’s partnership with OpenAI changed?

Answer. Microsoft’s updated contract with OpenAI lifts previous restrictions on AGI development and computing power, allowing Microsoft to build its own superintelligent systems while retaining a 27% stake in OpenAI’s public-benefit company.

Q3. What makes Microsoft’s superintelligence approach different?

Answer. Led by Suleyman, Microsoft emphasizes a human-centered philosophy, rejecting the AI arms race and focusing on safe, long-term benefits for billions through ethical innovation.

Also Read –

https://www.mymobileindia.com/microsoft-releases-windows-11-25h2-update-with-major-security-copilot-experiences-accessibility-upgrades-and-more/

https://www.mymobileindia.com/microsoft-officially-ends-free-windows-10-support-heres-what-users-need-to-know-about-esu-windows-11-upgrade-and-office-2019/