In an interview with CNBC’s Steve Kovach, Suleyman explained that Microsoft isn’t in a race to be first in developing frontier AI models. The company’s strategy hinges on staying just behind the bleeding edge—a position Suleyman refers to as “off-frontier.” The rationale? It’s significantly cheaper and allows Microsoft to focus on more targeted, practical use cases.
“It’s cheaper to give a specific answer once you’ve waited for the first three or six months for the frontier to go first,” Suleyman said. “That’s actually our strategy—to really play a very tight second, given the capital-intensiveness of these models.”
Suleyman is no stranger to the world of AI innovation. He co-founded DeepMind, the renowned AI lab acquired by Google in 2014 for a reported $400–$650 million. He later led Inflection, another AI startup, before bringing his expertise to Microsoft last year.
Today, Microsoft is leveraging its extensive partnerships to fuel its AI ambitions. Rather than solely relying on internal innovation, the tech giant collaborates closely with external players—most notably, San Francisco-based OpenAI. Microsoft has infused OpenAI’s technology into its ecosystem, integrating AI models into products like Bing, Windows, and the Copilot assistant. These integrations offer users powerful language and image-generation capabilities.
At an event celebrating Microsoft’s 50th anniversary in Redmond, Washington, Suleyman revealed that Copilot will soon gain a “memory” feature—allowing it to remember key facts about frequent users. That function first debuted in OpenAI’s ChatGPT, which now boasts over 500 million weekly users. Microsoft later added a comparable feature called “Think Deeper” to its Copilot assistant, underscoring its strategy of rapid but deliberate follow-through.
Beyond major AI models, Microsoft is also investing in accessible solutions. It occasionally releases open-source small language models that can run directly on PCs—models that don’t rely on GPU-heavy server infrastructure like OpenAI’s more complex offerings.
Despite their close partnership, recent developments suggest a shifting dynamic between Microsoft and OpenAI. Microsoft has poured $13.75 billion into OpenAI since 2022, helping to catalyze the generative AI boom. But cracks have started to appear. In July 2024, Microsoft officially listed OpenAI as a competitor. Then in January 2025, OpenAI announced a collaboration with Microsoft’s cloud rival Oracle on the ambitious $500 billion Stargate project. Nevertheless, Microsoft stated that OpenAI had recently made a fresh “large Azure commitment,” signaling that the collaboration isn’t over yet.
Suleyman emphasized the long-term importance of Microsoft building its own AI capabilities in-house. “It’s absolutely mission-critical that long-term, we are able to do AI self-sufficiently at Microsoft,” he said. “At the same time, I think about these things over five- and 10-year periods. You know, until 2030 at least, we are deeply partnered with OpenAI, who have had an enormously successful relationship with us.”
Microsoft may not be charging ahead with the most cutting-edge AI models, but that’s not necessarily a weakness—it’s a strategy. “We have an incredibly strong AI team, huge amounts of compute, and it’s very important to us that maybe we don’t develop the absolute frontier, the best model in the world first,” Suleyman said. “That’s very, very expensive to do and unnecessary to cause that duplication.”
In short, Microsoft is proving that in the AI race, being second doesn’t mean being behind—it might just mean being smarter.