Mustafa Suleyman, CEO of Microsoft AI, has outlined a bold new direction for the company as it moves to build artificial general intelligence (AGI) independently, marking a significant shift from its long-standing partnership with OpenAI. In a recent interview with Business Insider, Suleyman revealed the launch of a new team called Microsoft AI Superintelligence, tasked with developing frontier-grade AI models in-house and advancing the company’s long-term goals in superintelligence. Suleyman emphasized that Microsoft now has the freedom to pursue AGI on its own or in collaboration with third parties, following a major renegotiation of its partnership with OpenAI. Previously, a contractual agreement limited Microsoft from developing its own AGI until 2030, a condition designed to protect OpenAI’s leadership in frontier AI while allowing Microsoft to provide critical cloud and financial support. That restriction has now been lifted, enabling Microsoft to build its own large-scale models and invest heavily in foundational research. The new team will focus on solving key technical challenges, such as enabling AI models to transfer knowledge across tasks and continuously learn without forgetting previous information. These capabilities are seen as essential for creating systems that can operate across diverse real-world domains, from healthcare and energy to transportation and economic efficiency. To support this ambition, Microsoft is making substantial investments in compute infrastructure. This includes expanding its partnership with Nvidia, accelerating its own AI chip development, and building a dedicated AI supercluster. Suleyman described this as a top priority, stating that Microsoft aims to create the most performant AI infrastructure in the world. Despite this push for independence, Suleyman stressed that Microsoft remains open to using a range of AI models—whether from OpenAI, Anthropic, open-source projects, or its own MAI series. “There’s no reason for us to be religious about that,” he said, underscoring a pragmatic approach focused on delivering effective, reliable AI products. Suleyman also highlighted Microsoft’s commitment to responsible AI. The company has brought in Trevor Callaghan, a former legal leader at DeepMind and Google, as vice president of responsible AI, signaling a strong focus on safety and alignment with human values. Suleyman framed the company’s mission not as a race for technological supremacy, but as a responsible effort to build AI that enhances human potential and remains under human control. “We’re not chasing a mountain for its own sake,” he said. “The goal is to ensure these systems are aligned with human interests and don’t outpace our ability to guide them.” With its vast data, global distribution network, and deep integration across products, Microsoft believes it is uniquely positioned to deploy AI agents at scale and transform workflows across industries. Suleyman described the future as one where AI becomes “really, really magical” in how it supports users and organizations. As Microsoft enters a competitive landscape already populated by superintelligence teams at Meta, Google, Anthropic, and xAI, it aims to stand out not through hype, but through responsibility, infrastructure strength, and a commitment to building AI that serves humanity.