OpenAI Chief Executive Officer Sam Altman said on Aug. 12 that the company will roughly double its compute fleet over the next five months to meet surging demand sparked by the launch of GPT-5. The expanded capacity will first ensure that paying ChatGPT subscribers receive more total usage than they did before the latest model’s debut. Once subscriber needs are met, OpenAI will allocate remaining resources to its application-programming-interface customers and is diverting some compute previously reserved for research toward product operations. The move is intended to ease recent service throttling and could translate into additional orders for high-performance chip suppliers such as Nvidia.