OpenAI CEO Sam Altman described the situation earlier in a San Francisco dinner attended by only a dozen media outlets.The roadmap after GPT-5This includes plans for trillion-dollar data center construction, new AI hardware in collaboration with Jony Ive, and even a recent exploration of the brain-computer interface market. Sam Altman even bluntly stated, "The CEO of three years from now may be AI."
Sam Altman said frankly,GPT-5It's indeed inferior to GPT-4o in emotional expression, which led OpenAI to redesign the model selection interface. However, for OpenAI, publishing a single model is no longer the core of its development. The focus has been on disrupting search, hardware, and social experiences, and expanding AI applications on a global scale.
Trillion-dollar data centers and new financial models
Sam Altman revealed that OpenAI will invest trillions of dollars in data center construction to support its inference services and new product operations. He even suggested that the company might fund its computing capacity through the development of new financing tools. His logic is straightforward: "Spend $3000 billion on data center construction in exchange for $4000 billion in service revenue. Otherwise, we'll continue to disappoint our customers."
Sam Altman also stated that OpenAI's current inference services are already profitable, but the high cost of training still creates a financial burden. Excluding training costs, OpenAI would already be "an extremely profitable company."
AI hardware and brain-computer interface competition
In terms of hardware, Sam Altman emphasized the importance of working with Jony Ive, the former design director of Apple.New AI equipment, saying that this is a product "worth waiting for", and even joked that if anyone dared to put a phone case on it, he would "personally find out". In addition, Sam Altman also confirmedExploring opportunities for brain-computer interface development, hoping that in the future we will be able to "talk to ChatGPT directly through thoughts."
GPT-5 Controversy and the Path to Humanization
In response to criticism of GPT-5's "coldness," Sam Altman announced that the model's personality will be customizable, allowing users to choose whether ChatGPT's responses are "warm" or "ruthless." This not only addresses external criticism but also aligns with the company's previous emphasis on "products remaining neutral, rather than forcing a particular stance."
Currently, ChatGPT has over 7 million weekly users worldwide, making it the fifth-largest website in the world. Sam Altman even predicts it will soon surpass Instagram and Facebook. Sam Altman further believes that as conversations continue to grow, "one day, ChatGPT will produce more words per day than the entire human population combined."
The AI Bubble and Long-Term Ambition
Meanwhile, Sam Altman acknowledged the risk of an "AI bubble," but emphasized that, similar to the dot-com bubble, there's still real value behind it. He called AI "one of the most important developments of recent times," and added that OpenAI's goal isn't just to build models; it aims to reshape search, social behavior, and the hardware ecosystem through AI. He even stated that if Google were to sell its Chrome browser business, he would consider acquiring it.
It now appears that the recent controversy over GPT-5 is just a transition, and OpenAI is fully committed to building an AI empire with a larger scale and more complex logic.
