OpenAI For-Profit Restructuring Passes As Microsoft Trains In-House AI Models

We live in a world where technology touches nearly everything we do—from how we communicate to how we work, create, and even think. Once considered a futuristic concept, artificial intelligence (AI) is now at the center of this transformation. While it inspires excitement and sparks curiosity, it also raises questions about what’s coming next.

So, let’s take a closer look at one of the biggest developments unfolding right now!

Microsoft and OpenAI are reshaping their partnership as the rapid growth of AI pushes both companies in new directions. On Thursday, the two confirmed they had signed a non-binding agreement that would allow OpenAI to transition into a for-profit business.

The move is designed to help OpenAI raise new funding under a standard corporate structure, though final commercial terms are still being negotiated.

Microsoft first invested $1 billion in OpenAI in 2019 and added $10 billion more in 2023. For years, it held exclusive rights to sell OpenAI’s tools through Azure. That exclusivity has since loosened, with OpenAI striking its own cloud partnerships, including a $300 billion agreement with Oracle and another with Google.

As part of the restructuring, OpenAI’s nonprofit arm could receive more than $100 billion, around 20% of the company’s $500 billion target valuation. Approval is still required in the states of California and Delaware, as OpenAI aims to complete the process by the year end to secure the funds.



Here’s a quick look at the investments and deals that have shaped this evolving partnership.

OpenAI is moving forward with plans to restructure, maintaining its nonprofit foundation while also giving it a financial stake in a Public Benefit Corporation (PBC).

In its statement, the nonprofit’s stake in the PBC will exceed $100 billion, placing it among the wealthiest philanthropic organizations worldwide.

The recapitalization is designed to secure the funds needed for OpenAI’s long-term mission while ensuring the nonprofit’s resources grow alongside the business. “Our core mission remains ensuring AGI benefits all of humanity,” the group noted, adding that safety decisions will remain guided by this principle.

In addition, OpenAI launched applications for a $50 million grant program aimed at promoting AI literacy, encouraging community-driven innovation, and creating economic opportunities.

As the exclusive partnership with OpenAI unravels, Microsoft is seeking to preserve its dominance in the AI race by investing heavily in in-house AI models to reduce reliance on the AI startup.

The company has begun developing its own frontier models, marking a new phase in its relationship with OpenAI. Last month, Microsoft introduced its first in-house models, with AI chief Mustafa Suleyman noting they are making “significant investments” in computing power to support the next wave of model development.

“We should have the capacity to build world class frontier models in house of all sizes, but we should be very pragmatic and use other models where we need to,” Suleyman said.

“We’re also going to be making significant investments in our own cluster, so today MAI-1-preview was only trained on 15,000 H100s, a tiny cluster in the grand scheme of things,” he added.

Indicating about the model development on a scale comparable to Meta, Google, and xAI, using clusters “six to ten times larger in size” than those used for the MAI-1-preview, Suleyman stated, “Much more to do, but it’s good to take the first steps.”

Alongside these developments, Microsoft CEO Satya Nadella underlined the importance of integrating multiple models into its products lineup, emphasizing “model-forward products” and supporting “multiple models,” pointing to GitHub Copilot as an example.

Microsoft also plans to integrate Anthropic’s AI models into Microsoft 365 Copilot, as they reportedly outperformed OpenAI’s models in Excel and PowerPoint.

Do you think moving to a for-profit structure puts OpenAI’s broader mission at risk? How will Microsoft’s training of in-house AI models affect its relationship with OpenAI?

Let us know your thoughts in the comments section below!