OpenAI’s $10 billion raise was initially meant to fuel its chatbot development, but the company burned through most of that money in just 18 months. To avoid another cash crisis, OpenAI plans to borrow an additional $4 billion and expects to spend over $5.4 billion a year by next year.
The main reason for these massive expenses is the enormous amounts of data required to train AI systems like ChatGPT. Feeding these systems with vast amounts of data makes them more powerful, similar to how a student improves their skills by reading more books. However, this comes at a high cost, including specialized computer chips called GPUs that can cost up to $30,000 each.
The shortage of computing power, data centers, and electricity needed for AI development has pushed costs higher. To address this, companies like OpenAI, Google, and Microsoft are expanding their global pool of data centers and securing the necessary electricity. Despite offering free chats, these companies charge consumers around $20 a month to use their most powerful technologies.
OpenAI’s chatbot technology has improved steadily since its initial version, and it recently unveiled a new “reasoning” feature that uses reinforcement learning to solve math problems and computer programming issues. This innovation requires even more computing power, which OpenAI expects to grow sevenfold by 2029 as it pursues artificial general intelligence.
The high costs of AI development have sparked concerns about sustainability, with some experts warning that chasing science fiction can be costly. As companies push the boundaries of AI, they must balance their ambitions with financial realities to make these technologies accessible and affordable for everyone.
Source: https://www.nytimes.com/2024/12/17/technology/openai-chatgpt-funding.html