OpenAI Reportedly Planning to Use Google Cloud to Meet Computational Needs

OpenAI is reportedly planning to use Google Cloud’s services to fuel its computational requirements. As per the report, the San Francisco-based artificial intelligence (AI) firm and the Mountain View-based tech giant have been discussing this partnership for several months. With this move, OpenAI could shift some of the server requirements for ChatGPT and Sora to Google. Currently, the financial details of this deal are not known, although it is likely that the Gemini-maker is getting its fair share of remuneration.

Google Cloud to Reportedly Carry OpenAI’s Compute Burden

According to a Reuters report, the ChatGPT maker has struck a deal to use Google Cloud’s servers to power ChatGPT’s computational needs. Citing three different unnamed sources, the publication claimed that OpenAI’s requirement for additional cloud servers has increased in recent times, which has led to them seeking this partnership.

Not a lot of details about the deal are known, although the two entities reportedly discussed the partnership for several months before finalising it in May. No financial terms of the deal were revealed.

This partnership between OpenAI and Google is intriguing, given that both of them are currently competing for the same share of the AI market. Both companies have been aggressively releasing large language models, expanding the use case of their chatbots, and finding newer ways to reach users (while Google is integrating Gemini into all of its products, OpenAI has been exploring partnerships with OEMs).

Interestingly, less than a week after Google released the upgraded Gemini 2.5 Pro model, OpenAI introduced its o3-pro model. Similarly, after the former announced reduced application programming interface (API) costs for its latest AI models compared to the latter, the ChatGPT maker announced an 80 percent price cut on its o3 model.

Despite the competition, however, OpenAI cannot ignore its dire need to increase compute capacity. After the release of the image generation feature powered by GPT-4o, Altman had posted that the company’s GPUs were “melting.” Separately, ChatGPT servers were down for several hours on Tuesday, resulting in a widespread outage.