Report: Mustafa Suleyman Says Microsoft Now Free to Chase AGI

Microsoft’s AI goals have started to entering a new era. According to AI chief Mustafa Suleyman, the company is now free to independently pursue artificial general intelligence (AGI) after years of contractual limits tied to its partnership with OpenAI.

In a new interview with Business Insider, Suleyman said that the is building a “world-class, frontier-grade research capability in-house” under a newly formed unit called Microsoft AI Superintelligence. The team will reportedly advance the kind of AI that could one day outperform humans across multiple tasks.

Until recently, Microsoft’s earlier agreement with OpenAI barred it from racing towards AGI until 2030. This limited the company to remain focused on smaller models and post-training work. That limitation is now gone. Following a renegotiated partnership, Microsoft can reach AGI on its own or with other partners.

Suleyman said the company now plans to train its own large-scale models, develop custom AI chips, and invest heavily in compute infrastructure with support from NVIDIA and its in-house silicon efforts. “Microsoft needs to be self-sufficient in AI,” he told the news outlet.

As rivals like Google, Meta, Anthropic, and xAI chase their own “superintelligence” goals, Suleyman said Microsoft wants to be the “adult in the room.” The focus, he stressed, is on humanist intent to design systems that remain safe and aligned with human interests.

He recently added Trevor Callaghan, formerly of DeepMind, as VP of responsible AI to oversee this mission. “There’s a risk these systems get extremely smart and run away from us,” Suleyman warned. “We have to design them so they don’t,” he added.

With new autonomy and vast resources, Microsoft’s AI future now looks more independent than ever.