Following Anthropic earlier this month, OpenAI also secretly submitted an initial public offering (IPO) application.
Just this morning, OpenAI officially announced that it had recently secretly submitted an S-1 registration statement.
However, OpenAI has not yet decided on a specific IPO date and may need some time, as some things may be easier to accomplish as a private company.
This involves a series of complex trade-offs. This submission gives OpenAI the option to go public earlier if necessary.
According to Bloomberg citing insiders, OpenAI is working with Goldman Sachs and Morgan Stanley, with plans to go public as early as this fall.
OpenAI's latest round of financing was on March 31, raising $122 billion in committed capital and a post-investment valuation of $8520 billion. Investors include Amazon, Nvidia, SoftBank, Microsoft, and others. This round of financing set a new record for the largest funding in OpenAI's history.
The comment section started joking, saying OpenAI is finally going "open." Yes, going public is also a kind of openness.
Alongside the official IPO announcement, Altman also co-authored a lengthy article with OpenAI's Chief Scientist Jakub Pachocki, outlining OpenAI's current and future plans.
The full text is as follows:
Every few generations, a new technology changes everything.
Imagine the scene in the 1920s, when electricity entered a rural American town. Before the wiring was laid, daily life was constrained by many practical conditions: people had to carry water themselves, wash clothes by hand, and store food with ice; After sunset, many of the day's activities come to an end. Electricity did not change every household overnight, nor did the benefits it bring reach everyone equally. But as electricity became widespread, ordinary people's lives began to change. With lights at night, the days are extended. Electric water pumps, household appliances, and refrigerators have eased many of the heaviest daily tasks. Radio brings news, music, and human connections from hundreds of miles away into homes and community spaces.
Electricity initially has practical value, but its deeper impact comes from new possibilities opened up after more people use it. As time went on, more and more new things emerged, and machines and computers greatly accelerated progress in medicine, engineering, and many other fields. By the end of the 20th century, human average life expectancy had increased by about 23 years, and average income after adjusting for inflation had risen by about 50%. These advances largely stem from improvements in healthcare, hygiene, and living standards, many of which have been driven or accelerated by electrification and related technological advances.
Similar things are happening again with AI. AI will soon be able to do many extraordinary things. But the focus is not on the technology itself, but on what people can do with it. It can help someone read medical bills, learn a new skill, start a small business, care for elderly parents, understand a legal or financial decision, turn an idea into reality, or even drive a scientific discovery.
The sense of wonder brought by lighting up at night may soon fade. But what people decide to do under the lights does not disappear. Technology has always been a reliable way to create long-term prosperity, so we believe AI should be accessible to everyone, and to the fullest extent possible, whenever and wherever they need it.
Such a future will not come automatically. Transformative technologies may further centralize power or make it more decentralized; It can make life easier for a few people and also expand opportunities for more people. Our starting point is clear: AI should serve people, help them pursue their goals, empower them, and distribute the benefits of this technology as broadly as possible.
Our first commitment is to create AI that serves all humanity. This means we want to empower everyone broadly, rather than letting power concentrate in the hands of a few companies, governments, or individuals. We believe that a safer future is one where power is distributed more broadly. This way, more people around the world can participate in building a more resilient ecosystem.
We remain optimistic about AI because we believe it can expand human capabilities and create more prosperity. But we also clearly see the risks involved. Powerful systems must remain secure, aligned with human intentions, and always under human control. OpenAI's mission is to ensure that AGI benefits all humanity. This means that the systems we want to build should help people do more of what they choose to do, rather than replacing human judgment of what matters.
Automating everything completely is not the future we want. Such a future would cause people to lose satisfaction and bring danger. AI should help people pursue their own goals, rather than operating independently of them. As AI systems become more powerful, human roles will become more important: setting direction, making trade-offs, using judgment, and bringing values, aesthetics, care, and responsibility into our work.
In the long run, a key human role will be to decide what is worth doing.
We believe that AI participation in AI research will become a key factor determining the pace of progress in the coming years. This is important because alignment itself is a challenging research problem. To achieve rapid and in-depth progress, our researchers need AI systems to help them test ideas, identify errors, explore alternatives, and iterate alongside them.
But the faster technological advances, the more important human judgment and public coordination become. The future should be shaped by people, institutions, and society together, not by companies building the most powerful systems.
As frontier AI continues to advance, we expect coordination at both national and global levels to become increasingly important. We have always believed that ultimately, there should be an international organization to help coordinate leading AI efforts to reduce catastrophic risks. Cooperation and common safety standards are important components on the road forward. This is especially critical when the incentives of business and national competition are hard to escape. One of the goals of such an organization should be to enable the world to take coordinated action, including slowing the development of cutting-edge AI when necessary, so that social resilience, security, and alignment can keep pace with technological progress.
Currently, OpenAI has three main goals:
First, build automated AI researchers. That is, an AI system that can accelerate and increasingly automate the research process itself, while remaining guided, accountable, and always connected to people. We internally believe that by March 2028, a significant portion of OpenAI's research work may be conducted by AI systems together with our researchers. To make sufficient progress in alignment, we believe AI needs to iterate alongside us. This will help us navigate the transition to a post-AGI world and jointly decide the path forward.
Second, accelerate economic development. Driving economic progress by accelerating scientific progress, productivity improvements, and economic growth, while striving to ensure that benefits are widely shared. Everyone should have the opportunity and fair share in the prosperity created by AI.
Third, to provide a personal AGI for every person on Earth. Allowing everyone to benefit from one of humanity's most transformative technologies, according to their own choices.
To achieve these goals, OpenAI is entering its third phase.
OpenAI's first phase focused on research focused on AGI. The second phase begins when our research starts relating to the real world, and we become a product company: deploying systems, learning from human usage, and continuing toward safe, aligned, and mission-aligned AGI.
Now, we are entering the third phase. The economy is beginning to be reshaped around AI. The core issue now is how to make advanced AI sufficient, affordable, secure, useful, and easy enough to use, benefiting everyone and every organization. Cutting-edge capabilities are just part of it. The bigger task is to turn these capabilities into tools that people can truly use and develop better because of them.
Most importantly, we believe that broad distribution of power will help lead to a better future. Human history shows that concentration of power creates vulnerability, while widespread sharing of power makes society more resilient, more adaptable to change, and freer.
That's exactly why access is so important. This is precisely why security, privacy, affordability, open ecosystems, and public oversight are equally important.
A good AI future cannot be one where a few institutions control most capabilities and most revenue. It should be a future where many people, many companies, many communities, and many countries can participate in building, benefit, and empower themselves. We believe this transformation should belong to everyone.
If we do this well, AI can become the foundation for driving higher productivity, stronger creativity, more scientific progress, and broader economic opportunities, ultimately fulfilling our mission: to ensure AGI benefits all humanity.
