Openai says it’s under non-profit control and rejects a full commercial shift amid the mask lawsuit

Openai is not a normal company and will never happen.

Our mission is to ensure that artificial general information (AGI) benefits all humanity.

When we first started Openai, we had no detailed sense of how we were going to achieve our mission. We began staring at each other around the kitchen table and wondered what research we should do. At the time, we weren’t pondering our business model, our products. The direct benefits of AI being used for medical advice, learning, productivity, etc., or the hundreds of billions of dollars of computing needs to train models and serve users.

I really didn’t know how AGI would be constructed or used. Many people can imagine an oracle that can tell scientists and presidents what to do, which can be incredibly dangerous, but perhaps those few people could trust them.

Many people around Openai in the early days thought that AI should be in the hands of a few trustworthy people who can “handle it.”

Now we are looking at how AGI can directly empower everyone as the most capable tool in human history. If we can do this, we believe that people will build something incredible for each other and continue to advance society and quality of life. Of course, it’s not all used for good, but we trust humanity and believe that good is a few orders of magnitude better than bad.

We are committed to this democratic AI path. We want to put incredible tools in everyone’s hands. We are amazed and delighted at what they are creating with our tools and how much they want to use them. We want to open source a very capable model. We want to give users great freedom in how we can use our tools within broad boundaries, even if users are able to make decisions about ChatGPT actions, rather than always sharing the same moral framework.

We believe this is the best way. AGI should allow all humanity to benefit one another. I find that some people have very different opinions.

We want to build our brains for the world and make them available for what people want (freedom should not hinder the freedom of others, for example).

People use ChatGpt to increase productivity as well as scientists, coders and more(Opens in a new window). People use ChatGpt to solve the serious healthcare challenges they face and learn more than ever. People use ChatGpt to get advice on how to handle difficult situations. We are extremely proud to provide a service that does so much for so many people. It is one of the most direct fulfillment of our mission we can imagine.

But they want to use it more. Now we cannot supply AI as much as the world wants, and we need to post usage restrictions in our systems and do it slowly. As the system becomes more capable, they will want to use it more for even more great things.

When I set up my lab almost ten years ago, I never thought this would become a world-wide state. But now that we see this picture we are excited.

It’s time for us to evolve our structure. There are three things I want to achieve.

We want to operate and acquire resources so that our services are widely available to all humanity who currently need hundreds of millions of dollars and could ultimately require hundreds of millions of dollars. We believe this is our best way to fulfill our mission and generate great benefits for each other with these new tools. We hope that nonprofits are the largest and most effective nonprofits in history, and are focused on using AI to enable people with the highest leverage results. I want to provide useful AGIs. This includes contributing to safety and alignment shape. We are proud of our track record of transparency into model behavior using systems we launched, alignment research we did, processes such as red teaming, and innovations such as Model Spec⁠.(Opens in a new window). As AI accelerates, our commitment to safety becomes stronger. We want to make sure that democratic AI wins over authoritarian AI.

We heard the voices from our civic leaders and discussed with the California Attorney General and the Delaware office before the nonprofit decided to maintain control. We look forward to moving forward with the details of this plan in our ongoing conversations with them, Microsoft, and the newly appointed nonprofit committee members.

Openai was founded as a nonprofit organization and is today a nonprofit that oversees and manages for-profit organizations, and will now remain a nonprofit organization that oversees and manages for-profit organizations. That remains the same.

For-profit LLC under nonprofit organizations will move to Public Benefits Companies (PBCs) on the same mission. PBC has become the standard for commercial structures for other AGI labs such as Anthropic and X.AI, and for many purpose-driven companies such as Patagonia. I think it makes sense for us too.

Instead of our current complex cap commercial structure, it makes sense when there may be one dominant AGI effort, but in the world of many great AGI companies it doesn’t seem to be – we’re moving into a normal capital structure where everyone has stock. This is not a sale, it is a structure change.

Nonprofits will continue to manage PBCs, become major shareholders of PBCs with amounts supported by independent financial advisors, and AI can benefit many different communities in line with their mission to provide nonprofit resources and support the program. And as PBC grows, more can be done as nonprofit resources grow. We look forward to getting immediate recommendations from our nonprofit committee on how we can help ensure that AI benefits everyone, not just a small amount. Their ideas will focus on how our nonprofit work can support a more democratic AI future and have a real impact on areas such as health, education, public services, and scientific discovery.

This has led to the continued progress in speed and safe ways, putting a large AI in everyone’s hands. Creating AGIs is our bricks on the path of human progress. I can’t wait to see the next brick I’ll add.

Sam Altman
May 2025 “”