GuShiio.com Article source: Head Technology
Image source: Generated by AI
As the open source AI model contributes more and more strongly to the world, OpenAI has begun to be criticized by the outside world for not being Open enough and only focusing on harvesting users and making money.
The vision set by OpenAI at the beginning of its establishment is grand. The purpose of conducting AI research is to promote and develop friendly artificial intelligence, which will benefit all mankind as a whole. Therefore, it has received widespread support and support.
In order to regain the influence and appeal of the industry, OpenAI today announced the launch of NextGenAI, and promised to provide US$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.
Currently, the competition in the field of AI models is very fierce. OpenAI’s latest GPT-4.5 has not set off a wave of impact in the market as expected, and its extremely expensive pricing and unsatisfactory performance have caused a lot of complaints. Whether it can still continue to sit firmly on the top spot in the industry has become suspense.
In-depth layout of “industry-university-research”
What does NextGenAI do? 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 institution. The plan is not only to promote the next generation of discovery, but also to prepare for shaping the future of artificial intelligence. Commit $50 million in research grants, computing funds and API access to support students, educators and researchers advancing the knowledge frontier.
In addition to OpenAI, the first founding partners of NextGenAI include the California Institute of Technology, California State University System, Duke University, University of Georgia, Harvard University, Howard University, Massachusetts Institute of Technology, University of Michigan, University of Mississippi, Ohio State University, Oxford University, Political School of Paris, Texas A & M University, Boston Children’s Hospital, Boston Public Library, etc.
Each partner institution 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 therapies and agriculture;
Researchers at Harvard University and Boston Children’s Hospital are using OpenAI tools and NextGenAI funding 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;
MIT students and faculty will be able to use OpenAI’s APIs and computing funding to train and fine-tune their own scientific AI models and develop new applications, among other things.
Brad Lightcap, chief operating officer of OpenAI, 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 OpenAI’s other market card outside of its technical layout. As early as May last year, OpenAI launched the ChatGPT Edu version, specifically tailored for universities and supported by the GPT‑ 4o model. It aims to responsibly deploy AI to students, faculty, researchers and campus operations, bringing AI to campuses on a large scale and becoming an accessible option.
In January, OpenAI signed an agreement with the U.S. National Laboratory to use the latest inference models to enhance scientific research. The U.S. National Laboratory has approximately 15000 scientists. OpenAI has partnered 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 deep binding with a wide range of scientific research and educational institutions has become a step in the market operation of OpenAI’s leading rivals, and also provides a reference for the market application of my country’s AI models.
Restoring the “ambition” of the robot model
It is worth noting that in addition to the conversational AI model, OpenAI may have to accelerate the expansion and exploration of the basic robot model next, and is currently recruiting people.
According to information shared by OpenAI researcher Xingyu Lin on social media, OpenAI’s robot learning team is open for recruitment. Currently, this is a small but talented team committed to building a robot stack from zero to one and exploring the basic model of robots. Basic issues.
It is understood that the robot team’s mission is to extend the functions of the basic model to support universal robots in dynamic reality environments. These functions 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 robotic technology stack, including cutting-edge hardware, sensors, end effectors and models, exploring the possibilities of various humanoid robots, with the goal of developing a model that can provide state-of-the-art intelligence and seamless physical skills within the practical constraints of the robotic platform. In terms of benefits, engineers will be offered an annual salary + equity award of US$290,000 to US$530,000.
In the field of robot models, OpenAI had previously maintained cooperation with Figure AI, the American star robot company, 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 specifically built for humanoid robots that unifies perception, language understanding, and learning control to overcome many long-term challenges in the field of robotics and in demonstrations. Scenes where multiple robots collaborate to do housework are unlocked.
For now, 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 OpenAI’s influence among the developer community and the AI industry, if it succeeds, it will surely soon set off a new wave of technology on the hottest humanoid robot track today.
Evolving business models
What kind of AI organization OpenAI will eventually develop into has not yet been finalized, but what is certain is that it has successfully escaped the shackles of the “non-profit” public welfare label at the beginning of its establishment and has received support from the judicial department.
According to Bloomberg, a U.S. federal judge today rejected Elon Musk’s request to prevent OpenAI from transforming into a for-profit entity. However, the documents stated that other aspects of Musk’s lawsuit against him and Sam Ultraman and others for co-financing of OpenAI could continue.
In early 2025, OpenAI built a new business architecture. The company 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 posts, Musk more than once described OpenAI’s transformation as a “complete hoax” and claimed that “OpenAI is evil.”
Musk has repeatedly said in statements: “It is time for OpenAI to return to the power of open source and security.” But Ultraman rejected the proposal, arguing that the move was just an unfair competitive means to hinder the speed of OpenAI’s development.
At present, due to the huge initial investment costs, it is not realistic for OpenAI to engage in open source. However, OpenAI will partially open its own 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 domestic AI model companies.
