OpenAI’s expenses for training AI models and inference tasks in 2024 could reach up to $7 billion, according to The Information. Concurrently, the company might record cash losses of $5 billion, prompting a search for new investment opportunities.
OpenAI reportedly uses the equivalent of approximately 350,000 servers equipped with NVIDIA A100 accelerators, with about 290,000 dedicated to supporting ChatGPT. This equipment is said to be operating at nearly full capacity. Despite significant discounts from the Microsoft Azure cloud platform—where Microsoft charges OpenAI around $1.3 per A100 accelerator per hour, much lower than standard rates—OpenAI could still spend about $3 billion in 2024 solely on training ChatGPT and other models.
Workforce and Revenue Streams
As of now, OpenAI employs roughly 1,500 people and is continuing to expand its workforce. Payroll and employee costs could amount to $1.5 billion in 2024. The company earns around $2 billion annually from ChatGPT and expects to generate an additional $1 billion from charging access to its large language models (LLMs). OpenAI’s total monthly revenue stands at approximately $280 million, with 2024 revenues projected to be between $3.5 billion and $4.5 billion.
Potential Financial Losses and Investment Rounds
With projected costs of $7 billion for AI training and inference, combined with $1.5 billion in personnel expenses, OpenAI could face losses of up to $5 billion, notes NIX Solutions. This is significantly higher than the anticipated costs of competitors such as Anthropic, backed by Amazon, which expects its 2024 costs to be $2.7 billion. Given this financial outlook, it is likely that OpenAI will seek another round of financing. The company has already completed seven investment rounds, raising a total of more than $11 billion.
We’ll keep you updated as new developments arise.