OpenAI's data center ambitions collide with reality

OpenAI is facing challenges in financing and building massive data centers to meet its ambitions, despite its hunger for computing capacity. The company's partnership with SoftBank has hit roadblocks, and no specific deals have been inked yet. However, OpenAI has announced an expansion of its work with Oracle to build a further 4.5 gigawatts of capacity. OpenAI's CEO Sam Altman acknowledged that the company's thirst for computing is straining the supply chain and will require creativity. The company is partnering with various companies and governments to get more processing power, including Microsoft, Oracle, and the United Arab Emirates. OpenAI expects to have the equivalent of 2 million Nvidia A100 processors available by the end of August and double that capacity by the end of the year. The company is not alone in its need for massive data centers, as global data center capacity is expected to grow by 23% per year through 2030. Other companies like Meta, Google, Microsoft, and xAI are also driving this growth. Elon Musk's xAI company has 230,000 GPUs up and running for training models, and Anthropic is acknowledging the need for flexibility in choosing partners to meet its data center needs. Ultimately, all companies chasing AI with superhuman capabilities will need more money, electricity, and processing power to achieve their goals.