MICROSOFT AND OPENAI : It's complicated

Microsoft has always been a platform and tools company. It builds foundational layers—operating systems, developer frameworks, cloud infrastructure—and lets others build atop them. In the era of hyperscale, under Satya Nadella’s leadership, Microsoft evolved into a behemoth, expanding fast with bold bets: acquiring LinkedIn and GitHub, scaling Azure, and doubling down on cloud. I witnessed that evolution firsthand, as one of the early engineers on Microsoft 365’s cloud journey. It was a time when the company moved faster than ever before. It knew that to lead, it had to go beyond what was comfortable.

That same instinct is what led Microsoft to OpenAI.

In 2019, the partnership began quietly—with Microsoft investing $1 billion in OpenAI’s vision of Artificial General Intelligence (AGI) that would benefit all of humanity. It was part strategic, part philosophical. Microsoft saw in OpenAI a shot at the future. OpenAI saw in

Microsoft the infrastructure and scale to get there.

Since then, both companies have changed—dramatically.

The AI Bet That Changed Everything

By 2023, Microsoft had poured over $13 billion into OpenAI, securing exclusive rights to offer OpenAI’s models via Azure, integrating GPT into Microsoft 365 Copilot, and embedding it across Bing, Teams, and Dynamics. The move wasn’t just about product features—it was a statement. Microsoft was no longer following the AI wave. It was the wave.

But in 2024, the dynamic began to shift. OpenAI raised a staggering $6.6 billion at a $157 billion valuation, bringing in new backers like Thrive Capital, Nvidia, SoftBank—and yes, Microsoft too. But Microsoft was no longer the lone patron. OpenAI was becoming a powerhouse in its own right.

At the same time, OpenAI announced Stargate, a $500 billion infrastructure initiative with SoftBank and Oracle. The aim? To build massive AI datacenters across the U.S. Microsoft approved the move and retains a right of first refusal for new compute capacity—but the message was clear: OpenAI was preparing for a world where it didn’t rely solely on Azure.

That shift marked a new chapter. Infrastructure exclusivity is no longer on the table. What remains is a delicate alliance—rooted in shared success, but shaped by evolving needs.

Navigating Co-opetition: Microsoft’s Playbook

If there’s one thing Microsoft understands better than most, it’s co-opetition—the art of collaborating with companies that also compete with you. For decades, it’s worked alongside rivals while building platforms that outlast any single product.

Its partnership with OpenAI follows that tradition.

Even as Microsoft integrates OpenAI’s models into its flagship products, it’s also building its own AI future. In 2024, it hired Mustafa Suleyman, co-founder of DeepMind and Inflection AI, to lead a new division: Microsoft AI. His team is now training an internal family of large language models, dubbed MAI, designed to match or exceed current frontier models.

Meanwhile, Microsoft continues to open up Azure to the broader AI ecosystem. It offers Meta’s Llama models, Anthropic’s Claude, Mistral, and even integrations with xAI. The strategy is clear: provide enterprises with choice while retaining leadership in infrastructure.

Microsoft is also investing heavily in hardware. Its custom AI chips—like the Maia 100 accelerator—are now powering its own workloads and reducing dependency on Nvidia’s supply chain. With $80 billion committed to global datacenter expansion, Microsoft is positioning Azure as the world’s supercomputer for AI.

This dual approach—deep partnership with OpenAI and simultaneous development of internal capabilities—ensures Microsoft remains adaptable, no matter how the landscape evolves.

Strategic Distance: OpenAI’s Rise and Microsoft’s Balancing Act

As OpenAI expanded, it also changed. The company that once focused solely on long-term research is now a commercial juggernaut. That transformation hasn’t come without friction.

In 2024, several high-profile leaders departed—Mira Murati, John Schulman, and even Ilya Sutskever—each exit exposing internal tensions between OpenAI’s founding ideals and the commercial momentum it now carries. Can a mission-driven lab stay true to its vision while chasing a $157 billion valuation and scaling for enterprise adoption? That’s the open question.

At the same time, OpenAI was broadening its alliances and ambition. It raised $6.6 billion in a late-2024 funding round led by Thrive Capital, bringing in Nvidia, SoftBank, and Microsoft. But more importantly, it launched Stargate—a $500 billion infrastructure initiative backed by SoftBank and Oracle, aimed at building AI supercomputing capacity across the U.S.

This marked a clear shift. While Microsoft approved the move—retaining a right of first refusal on infrastructure capacity—it also signaled that OpenAI was preparing for a world beyond Azure. The company now expects 75% of its future compute capacity to come from Stargate by 2030, diversifying beyond Microsoft without severing ties.

Still, the relationship remains deeply interwoven. Microsoft retains exclusive commercialization rights for OpenAI’s core models, and OpenAI services continue to run on Azure for now. The partnership is locked in through at least 2030, including tight distribution alignment for services like Azure OpenAI and ChatGPT Enterprise.

This is where Microsoft’s strength shows. As a platform company, it’s used to ecosystems filled with overlap and tension. Its strategy has always been to build the foundation and invite everyone to build on it—even competitors. That same strategy is playing out in AI: hosting OpenAI, Mistral, Meta’s Llama, and others on Azure; while simultaneously developing its own internal MAI models and custom chips like Maia 100 to hedge its future.

This is "co-opetition" at the scale of supercomputers and AGI. Both companies are growing in parallel—sometimes collaborating, sometimes preparing to compete.

The question that lingers for me is this: Should Microsoft author its own frontier model—not to replace OpenAI, but to reduce its risk surface? Or is it more valuable to remain the richest, most flexible platform for AI innovation?

Maybe both can be true.

What Comes Next

The Microsoft–OpenAI story is no longer one of startup and sponsor. It’s now a complex partnership between two giants—each with leverage, ambition, and a vision for the future.

OpenAI wants to lead us into the AGI era. Microsoft wants to be the platform that powers that journey—while building its own capabilities along the way. Whether that path stays unified or splinters into competition remains to be seen.

But one thing is certain: what began as a bold $1 billion experiment has become one of the defining alliances of the tech industry.

And like all great alliances, it’s not about control. It’s about timing, trust, and knowing when to build together—and when to build your own. Well, maybe it is a little bit about control.

The AI future is still unfolding. But the center of gravity? It’s right here, between these two.