On September 25, according to the Wall Street Journal, the day before, a major blog post was released revealing that OpenAI is implementing a grand global computing infrastructure expansion plan, with an investment estimated to reach up to $1 trillion, approximately RMB 7.11 trillion. As part of the plan, the computing center OpenAI has built in Abilene, Texas, USA, has officially launched. The scale of this project is comparable to Central Park in New York, demonstrating the company's ambitious aspirations in the field of AI computing power.
OpenAI has shifted from being primarily "model-driven" to "computing power-driven," aiming to address the massive computing power required for training and reasoning of next-generation artificial intelligence models. The company plans to build large-scale computing center clusters in multiple regions in the US and overseas to meet the growing demand for computing power in the future.
The day before this trillion-dollar investment plan was announced, OpenAI just announced a $100 billion partnership with chip giant NVIDIA. This collaboration not only guarantees high-performance computing hardware for OpenAI but also effectively alleviates concerns in the market about its financial sustainability.
In addition, OpenAI has also reached new partnerships with Oracle and SoftBank this week, planning to build five artificial intelligence data centers in the United States. These include three data centers in collaboration with Oracle and two data centers in collaboration with SoftBank, as well as expanding an Oracle data center located in Abilene.
The computing center in Abilene is the core of this series of plans, located about 180 miles west of Dallas, covering 1,100 acres, and the initial construction has been completed. The computing capacity of this center is expected to be more than 13 times the original scale, becoming the largest artificial intelligence supercomputing center in the world. With workers working hard in temperatures of 100 degrees Fahrenheit, the construction of this base is progressing rapidly.
The Wall Street Journal pointed out that this wave of large-scale computing center construction marks that the AI industry is shifting from competition in model innovation to an arms race in infrastructure. OpenAI's initiatives aim not only to build an insurmountable computing power barrier but may also reshape the global distribution of AI computing power, providing strong support for the development of next-generation large models including the GPT series.
