OpenAI and Microsoft are collaborating with Oracle to obtain more computing power to support the operation of ChatGPT. The partnership announced this week demonstrates the joint efforts of these three companies, enabling OpenAI to leverage Microsoft Azure AI platform on Oracle's infrastructure. OpenAI's CEO, Sam Altman, openly acknowledges the need for more infrastructure to support their services. He is even discussing raising billions of dollars for their AI chip business. In the press release regarding the collaboration with Oracle, Altman mentioned that Oracle's chips will "help OpenAI continue to scale." Currently, OpenAI relies entirely on Microsoft to meet its computing needs. Microsoft has invested $13 billion, acquiring a 49% stake in OpenAI's profitable subsidiary and obtaining exclusive commercial licensing rights to its technology. However, as demonstrated by this collaboration with Oracle, OpenAI requires more computing power than what Microsoft alone can provide to keep up with demand and prevent future failures of ChatGPT. Both Microsoft and OpenAI approach this collaboration with Oracle cautiously. On Wednesday, OpenAI released a follow-up statement emphasizing that their "strategic cloud relationship with Microsoft remains unchanged" and the new partnership "will enable OpenAI to use Azure AI platform for inference and other needs on Oracle Cloud Infrastructure (OCI)." (Inference refers to the behavior of running AI models in production environments, such as ChatGPT.) OpenAI also explicitly states that the pre-training of their cutting-edge models "will continue on the supercomputer built in collaboration with Microsoft." Another subtlety of OpenAI's collaboration with Oracle is that Oracle also provides infrastructure for xAI, which is Elon Musk's OpenAI's competitor.