OpenAI is set to lose $5B this year

OpenAI's training and inference costs are projected to reach $7 billion in 2024, with the company potentially facing a $5 billion loss. As of March, OpenAI was expected to spend nearly $4 billion on Microsoft Azure servers for running inference workloads for ChatGPT. The company operates approximately 350,000 servers equipped with Nvidia A100 chips, with around 290,000 dedicated to ChatGPT, running at near full capacity. Training costs for ChatGPT and new models could add another $3 billion this year. OpenAI benefits from discounted rates from Microsoft, paying about $1.30 per A100 server per hour. The workforce, now around 1,500 employees, is estimated to cost $1.5 billion, significantly higher than the initial projection of $500 million for 2023. OpenAI generates about $2 billion annually from ChatGPT and anticipates nearly $1 billion from access to large language models (LLMs). Recent revenue figures indicate a monthly total of $283 million, suggesting potential full-year sales between $3.5 billion and $4.5 billion. This financial outlook indicates a substantial shortfall, prompting the need for additional funding within the next year to cover operational costs and losses.