In a surprising move, OpenAI, the company behind ChatGPT, has partnered with Google Cloud to meet its increasing need for compute power. The agreement, finalised in May 2025, comes at a time when OpenAI’s popularity is straining its infrastructure, especially after launching graphics-heavy features like Ghibli-style image generation.
ChatGPT’s soaring popularity overwhelms servers
OpenAI’s tools, including ChatGPT and image generation services, have seen a rapid rise in usage. In June 2025, the company hit a USD 10 billion annualised revenue run rate. But this growth has brought challenges. On June 10, ChatGPT suffered its third major global outage this year, impacting users worldwide. CEO Sam Altman even joked that their “GPUs are melting” under the pressure.
Diversifying beyond Microsoft Azure
While OpenAI has traditionally relied on Microsoft’s Azure cloud services, it is now diversifying partnerships. The Google deal follows similar collaborations with Oracle, CoreWeave, and SoftBank as part of the massive Stargate AI infrastructure project. OpenAI is also working with G42 in Abu Dhabi to build one of the largest AI data centres globally, Stargate UAE.
Google opens TPUs to rivals amid AI race
Google, despite being a competitor, is offering its tensor processing units (TPUs) to OpenAI. These chips, previously used only in-house, are now being rented out to companies like Apple, Anthropic, and now OpenAI. The move bolsters Google Cloud's position as a neutral compute provider even as it competes in AI services.
Custom chips coming: OpenAI’s next big bet
To gain more control, OpenAI is also designing its own AI chips, likely to be manufactured by TSMC. These custom chips are expected to roll out by 2026, reducing OpenAI's reliance on Nvidia GPUs. However, the cost is steep, with development expected to exceed USD 500 million per chip variant.
Compute is the new oil in AI wars
As AI’s appetite for compute continues to grow, companies like Google, Microsoft, and OpenAI are racing to invest. Alphabet plans to spend USD 75 billion on AI-related infrastructure in 2025 alone. But even Google CFO Anat Ashkenazi admitted that the company is already struggling to meet cloud demand, something OpenAI’s partnership may complicate further.
