OpenAI and Google are reportedly set to sign a deal whereby Google will provide cloud services to the ChatGPT creator.
This is according to Reuters, citing sources. Until January, Microsoft Azure was the exclusive data center provider for OpenAI. This exclusivity had previously blocked talks with Google. Microsoft and OpenAI are now negotiating new terms for their billion-dollar partnership, including Microsoft’s future shareholding. The agreement with Google was finalized in May after months of negotiations.
Computing capacity is becoming scarcer
OpenAI has been facing a rising demand for computing power since the launch of ChatGPT in late 2022. The company needs computing power to train language models and for inference—processing information so that users can use these models. OpenAI’s monthly revenue recently reached an annualized level of $10 billion.
Google Cloud will provide additional capacity to OpenAI’s existing infrastructure. This will help train and run AI models.
The collaboration brings challenges. Google’s DeepMind unit competes directly with OpenAI in the race to develop the best AI models. By selling computing power to competitors, Google is reducing its supply of chips while giving capacity to rivals.
For Google, the deal will be a win. The company is expanding its own TPU (tensor processing unit) chips, which have historically been used only internally. Apple, Anthropic, and Safe Superintelligence have already become customers of this technology. Google Cloud, with $43 billion in revenue in 2024 (12 percent of Alphabet’s total revenue), is positioning itself as a neutral provider of computing resources.
Tip: ChatGPT has been experiencing technical difficulties for hours