Image source: Generated by Unbounded AI

As the contribution of open source AI models to the world becomes stronger and stronger, OpenAI has begun to be criticized by the outside world for not being Open enough, and is only concerned with harvesting users and making money.

The vision set at the beginning of OpenAI is grand. The purpose of conducting AI research is to promote and develop friendly artificial intelligence, and thus benefit all mankind as a whole, and therefore has received widespread support and support.

In order to regain the industry's influence and appeal, OpenAI today announced the launch of NextGenAI, and promised to provide $50 million in research funding, computing resources and API access, using artificial intelligence to promote research and education, and while making money, it began to show its "Open" side.

At present, competition in the field of AI big model is very fierce. OpenAI's latest GPT-4.5 has not set off a wave of impact in the market as expected. Moreover, the extremely expensive pricing and unsatisfactory performance have caused a lot of complaints. Whether it can continue to firmly sit in the top position in the industry has become a suspense.

In-depth layout of "industry, education and research"

What is NextGenAI for? According to reports, this is the first alliance of 15 leading research institutions, committed to using artificial intelligence to accelerate research breakthroughs and transform education.

OpenAI said that NextGenAI will join forces with institutions in the United States and abroad to promote progress faster than any AI organization. This plan is not only to promote the next generation of discoveries, but also to shape the future of artificial intelligence. Committed to provide $50 million in research grants, computing funding and API access to support students, educators and researchers to advance the knowledge frontier.

In addition to OpenAI, NextGenAI's first founding partners include California Institute of Technology, California State University Systems, Duke University, University of Georgia, Harvard University, Howard University, MIT, University of Michigan, University of Mississippi, Ohio State University, Oxford University, Paris School of Political Science, Texas A&M University, Boston Children's Hospital, Boston Public Library, etc.

Each partner will be able to use OpenAI technology to solve high-impact scientific research challenges, such as:

Ohio State University is using AI to accelerate development in areas such as digital health, advanced therapy and agriculture;

Researchers at Harvard and Boston Children's Hospital are using OpenAI tools and NextGenAI funds to reduce the time it takes for patients to find the right diagnosis, especially for rare orphan diseases;

Texas A&M University is using NextGenAI resources to promote its Generative Artificial Intelligence Literacy Program;

MITStudents and teachers will be able to use OpenAI's API and computing funds to train and fine-tune their own research AI models and develop new applications and more.

OpenAI Chief Operating Officer Brad Lightcap said NextGenAI will further strengthen the important partnership between academia and industry to ensure that the benefits of artificial intelligence extend to laboratories, libraries, hospitals and classrooms around the world.

This is another market brand for OpenAI outside its technical layout. As early as May last year, OpenAI launched the ChatGPT Edu version, tailored for universities, supported by the GPT‑4o model, aiming to responsibly deploy AI to students, faculty, researchers and campus operations, and bring AI to campus on a large scale into a accessible option.

In January this year, OpenAI signed an agreement with the U.S. National Laboratory to use the latest inference model to enhance scientific research. With approximately 15,000 scientists, OpenAI and Microsoft are working with Microsoft to deploy O1 and other O series advanced models on the NVIDIA supercomputer Venado at Los Alamos National Laboratory (LANL), becoming a shared resource for national laboratory researchers. Venado is designed to promote scientific breakthroughs in materials science, renewable energy, astrophysics and other fields. The in-depth binding with a wide range of scientific research and educational institutions has become a one-step market operation for OpenAI's leading competitors, and it also provides a reference for the market application of my country's AI model.

Regain the "ambition" of robot models

It is worth noting that in addition to the dialogue-based AI model, OpenAI may be going to accelerate the expansion of the exploration of the basic robot models in the next step, and is currently recruiting troops.

According to information shared by OpenAI researcher Xingyu Lin on social media, OpenAI’s robot learning team is opening up for recruitment. It is currently a small but talented team dedicated to building a robot stack from zero to one and exploring the basic problems in the robot’s basic model.

It is understood that the robot team's mission is to expand the functions of the basic model to support universal robots in dynamic real-life environments, which are not limited to action generation, motion planning, world modeling, and real-time communication through voice and emotion.

In addition, the team will also work on the entire robotics stack, including cutting-edge hardware, sensors, end effectors and models, to explore the possibilities of various humanoid robots, with the goal of developing models that can provide state-of-the-art intelligent and seamless physical skills under the actual constraints of the robotic platform. In terms of treatment, engineers will be given an annual salary of $290,000 to $530,000 + equity reward.

In RobotIn the model field, OpenAI previously maintained a cooperation with the American celebrity robot company Figure AI, but two weeks ago, Figure AI suddenly announced that it would stop its marriage with OpenAI and released its own machine learning model Helix on February 20.

Helix is ​​a universal "vision-language-action" (VLA) model specially designed for humanoid robots that unify perception, language understanding and learning controls to overcome several long-term challenges in the field of robotics and unlock multiple robots in the demonstrations for doing housework.

As of the moment, OpenAI has not given up its ambition to further advance the field of basic robot models. It is unknown whether it will launch a better model than Helix. However, with the influence of OpenAI in the developer group and the AI ​​industry, if it succeeds, it will soon set off a new wave of technology in the hottest humanoid robot track at the moment.

The evolving business model

What kind of AI institution will OpenAI eventually develop into is not yet finalized, but it is certain that it has successfully escaped the shackles of the "non-profit" charity label at its beginning of its establishment and has received support from the judicial department.

A U.S. federal judge today rejected Elon Musk's request to prevent OpenAI from transforming into a for-profit entity, according to Bloomberg. However, the document says other aspects of the lawsuit against Musk's joint funding for the establishment of OpenAI can continue.

In early 2025, OpenAI built a new business architecture, which created a public welfare company to oversee business operations and eliminated the "limited profit" model to make it more like a fast-growing startup. In X's post, Musk described OpenAI's transformation as a "complete scam" more than once, and claimed that "OpenAI is evil."

Musk repeatedly stated in his statement: "It is time for OpenAI to return to open source and focus on security." But Ultraman rejected the proposal and believed that the move was just an unfair competition method that hindered the development of OpenAI.

At present, due to the huge initial investment cost, it is not realistic to open source OpenAI, but OpenAI will partially open its technology through methods such as the NextGenAI Alliance to empower all walks of life, especially in the fields of scientific research and education, which is worth learning from by domestic AI model companies.