Mustafa Suleyman: Scaling AI for the Agentic Revolution

Share this article
Share this article
Prioritise Us on Google
Mustafa Suleyman. Credit: Microsoft AI
Focusing on core model development, Microsoft’s AI leader plans to transform how software acts autonomously across apps to meet complex enterprise needs

Mustafa Suleyman’s role has shifted to allow him to fully concentrate on Microsoft’s “superintelligence mission” and the development of core AI models. 

Previously, he was the CEO of Consumer within Microsoft AI, having worked closely with Satya Nadella for some time on the long-term strategy for this next era of AI.

Youtube Placeholder

His new role as EVP of Frontier Models and Strategy for Microsoft AI elevates his commitment to “advancing the frontiers of model science”.

Following a new mandate, Mustafa plans to focus on creating superintelligence that delivers a transformative, positive impact for millions of people. 

This requires building frontier models at scale, pushing the boundaries of what’s possible by leveraging Microsoft’s “ambitious, long-term frontier scale compute roadmap” to deliver state-of-the-art models

“Since I began work on AI in 2010, the amount of training data that goes into frontier AI models has grown by one trillion times,” he writes on LinkedIn.

Microsoft AI’s Mustafa Suleyman visiting UBTECH Robotics. Credit: X/@mustafasuleyman

“It’s taken us from models that can sometimes identify cars in photos to models that can act autonomously for up to 12 hours at a time, using different apps, taking actions.

“It’s an exponential explosion. Now consider that we’re looking at something like another 1000x in effective compute by the end of 2028.”

A renewed goal

Mustafa is tasked with ensuring these models have a real product impact, specifically in terms of achieving research breakthroughs and meeting complex enterprise needs. 

His work is deemed “foundational to everything we build above it”, according to Microsoft, and critical to the company’s success over the next decade.

For the enterprise tech landscape, this transition underscores Microsoft’s commitment to owning the base layer of the AI stack. 

By doubling down on core model development, Mustafa’s team will build “enterprise tuned lineages” that improve all Microsoft products, while simultaneously achieving the necessary cost of goods sold efficiencies to serve AI workloads at the immense scale required in the coming years. 

This dedication to advancing the model layer, as well as having “human control, agency and economic opportunity”, is critical for the “agentic revolution” and redefining Microsoft’s position as a provider of foundational, high-performance technology for global organisations.

“I hope all of us start reimagining what might be possible, and how fast,” he says. “What do we want that world to look like? It'll be here sooner than we think.”

Company portals

Executives