OpenAI is transforming from a research lab into a vertically integrated "AI empire."
On October 8, OpenAI CEO Sam Altman revealed in a latest conversation with Ben Horowitz, co-founder of the well-known venture capital firm a16z, that OpenAI has decided to make a "very aggressive infrastructure bet" on a scale that requires the participation of the entire industry.
He explained that the decision was based on strong confidence in the model's capabilities in the next one to two years, as they foresaw that the upcoming model would create huge economic value and the current expansion rate could no longer meet future demand.
This strategy directly explains OpenAI's recent series of partnerships with tech giants like Nvidia, Oracle, and AMD. Altman predicts that more such partnerships will be announced in the coming months, demonstrating its ambition to leverage the entire industry chain, from electronics to model distribution.
This may also mean that the AI competition is shifting from algorithms to an all-round struggle involving computing power, capital and energy.
Altman also directly linked the future of AI to the future of energy, noting that AI's exponential growth will require cheaper and more abundant energy. He predicted that the long-term solution will be a combination of solar energy and energy storage with advanced nuclear power, and asserted that the cost of nuclear energy will be the key variable in determining its rapid adoption and, in turn, its ability to support AI development.
When talking about the company's vision, Altman said that OpenAI is not just a research lab, but also a complex that integrates consumer AI subscription services, ultra-large-scale infrastructure operations and cutting-edge AI research, dedicated to building general artificial intelligence (AGI) and making it beneficial to mankind. 
Highlights from Sam Altman's interview:
Aggressive Infrastructure Bet: Altman revealed that the company is making a "very aggressive infrastructure bet," the scale of which requires industry-wide collaboration. This substantial investment is based on strong confidence in the capabilities of the model over the next one to two years, rather than the current model, and he foresaw more industry collaboration in the coming months.
Energy Future Blueprint: Altman noted that AI and energy have become "integrated," and that AI's exponential growth will rely on cheaper and more abundant energy. He predicted that the future energy landscape will be dominated by "solar + storage" and nuclear power. He believed that once nuclear power demonstrates "overwhelming economic advantages," its development will be extremely rapid, and called past restrictions on nuclear power "an extremely foolish decision."
Sora's strategic role: Sora is not only a tool for video generation, but also a strategic tool for building a "world model" to advance AGI and help society adapt to the development of AI.
"AI scientists" are coming: Altman predicts that AI models will be able to make major scientific discoveries within the next two years, which he sees as a true sign of AI changing the world. He revealed that GPT-5 is beginning to demonstrate the ability to make small, novel scientific discoveries.
Strategic Shift Toward Vertical Integration: Altman admitted that he was wrong about vertical integration in the past and now believes it is a necessary path for OpenAI to achieve its mission, comparing it to the success of Apple's iPhone.
New copyright models: He foresees that AI training may be considered fair use in the future, but using specific IP to generate content will give rise to new business models. Some copyright holders are even more concerned about their IP being underutilized by AI rather than overused.
Commercialization and Trust: In terms of commercialization, Altman is open but cautious about advertising, emphasizing that user trust in ChatGPT must not be undermined. He believes that recommending paid products instead of the best products will destroy this trust.
A "very aggressive gamble": infrastructure expansion based on future demand
Amidst ongoing questions about whether AI is a bubble, Altman's statement has undoubtedly fueled market enthusiasm. He stated bluntly, "We've decided it's time to make a very aggressive infrastructure bet."
This decision wasn't driven by demand for current products like ChatGPT, but rather by a firm belief in the future. Altman revealed, "We're being so aggressive not because of the models we have now... We can see [future model capabilities] a year or two in advance." He believes the economic value created by upcoming models will far exceed expectations, necessitating early planning.
The scale of this gamble is so enormous that it requires the support of an entire industry. “In order to make a bet of this size, we kind of need the entire industry… to support it,” Altman said. This encompasses “from electronics to model distribution and everything in between.”
This also explains why OpenAI is actively establishing partnerships with companies such as AMD, Oracle, and Nvidia. He further hinted that this is just the beginning, "You should see more action from us in the coming months."
The energy lifeline of AI: The future of nuclear and solar energy
Altman admitted that the two areas of his career that have been of greatest concern—AI and energy—have now merged. He believes that historically, the most significant factor in improving human quality of life has been cheaper and more abundant energy. The enormous computing power demands of AI are bringing energy issues to the forefront.
Altman laid out a clear roadmap for addressing future energy needs. He predicted that in the short term, additional baseload energy in the United States would primarily come from natural gas. However, in the long term, he believed that "the two dominant energy sources will be solar plus storage, and nuclear." The nuclear energy he was referring to encompasses the entire advanced nuclear technology stack, including small modular reactors (SMRs) and nuclear fusion.
Altman makes a key economic point about nuclear energy development. He argues that the speed at which nuclear energy becomes widespread depends entirely on its cost. “If it’s economically overwhelmingly advantageous relative to everything else, then I would expect it to happen very quickly.”
He added that intense political pressure would push regulators to act quickly. Conversely, if costs were comparable to other energy sources, anti-nuclear sentiment could make the process extremely slow. He bluntly called the decision to ban nuclear power in many regions in the past "a profoundly foolish one."
Sora's strategic value: more than just video production
Regarding the recently released Vincent film model, Sora, Altman also elaborated on its multiple roles within the company's strategy. While Sora may appear unrelated to artificial general intelligence (AGI) on the surface, he believes that building robust "world models" is far more crucial to achieving AGI than people realize.
Sora is also a crucial tool for OpenAI to guide the "co-evolution" of society and technology. Altman said that just as ChatGPT helped the world take large language models seriously, Sora can give society a forewarning of the impact and opportunities that powerful video models will bring. "It's crucial to educate the world about the direction of video technology development," he said.
Of course, Sora also presents new commercial challenges. Altman observed that users use it not only for professional creation but also for creating and sharing fun memes. The high production costs and the high frequency of recreational use conflicted, meaning OpenAI needed to explore a completely different business model for Sora than ChatGPT.
The catalyst for AGI: "AI scientists" are coming
Of all the potential applications of AI, Altman says he's "most excited" about "AI scientists." He believes the world will truly change when AI can independently make scientific discoveries.
“We’ve seen small examples of this happening for the first time with GPT-5,” he revealed. He observed that the model has been able to make some novel mathematical discoveries or achieve small advances in physics and biology research, and believes that this trend will go further.
He predicts that within the next two years, models will be able to take on larger chunks of scientific work and make important discoveries that will have a "significant impact" on the world.
From investor to operator: a shift in perspective on vertical integration
In the interview, Altman also reflected on his transition from investor to company operator and its impact on OpenAI's strategy. He admitted that he had always been against vertical integration, but "now I think I was wrong."
He attributes this shift in perspective to his practical experience running a company. As an investor, he favored the theoretical notion of market efficiency, where each company focused on just one thing. But as CEO, he realized that achieving his company's mission required him to personally do more than he anticipated. The phenomenal success of Apple's iPhone is a prime example of extreme vertical integration; Altman calls it "the most incredible product ever created in the tech industry."
This shift in thinking from "suggestion" to "execution" explains why OpenAI has gradually expanded from a pure research laboratory to building its own large-scale infrastructure, striving to control the full stack capabilities from underlying computing power to upper-level applications, laying the foundation for its grand AI empire.
