OpenAI discreetly collaborates with Google to maintain online presence

OpenAI partners with Google Cloud to overcome GPU shortages and meet demand.

: OpenAI has partnered with Google Cloud to address its infrastructure shortages, specifically the lack of enough GPUs to handle increasing user demand. This collaboration, disclosed quietly via a sub-processor list on OpenAI's website, allows OpenAI access to Google's extensive data center infrastructure. The partnership is significant since OpenAI is backed by Microsoft and has been a competitor to Google in AI ventures. With Google's support, users might see improved performance and stability in OpenAI's popular tools like ChatGPT.

OpenAI has quietly entered into a partnership with Google Cloud to expand its computing infrastructure, marking a significant shift in its long-standing reliance on Microsoft Azure. This move was disclosed in mid-2025 as OpenAI began listing Google Cloud as a formal partner on its official documentation pages. The decision reflects OpenAI's strategy to diversify its cloud resources amidst rapidly increasing demand for artificial intelligence services, particularly with the scaling of products like ChatGPT.

Reports revealed that OpenAI is now engaging multiple cloud providers, including Google, Oracle, and CoreWeave, in a broader effort to mitigate dependency risks and ensure the resilience of its operations. Google Cloud's infrastructure, including Tensor Processing Units (TPUs), provides OpenAI with additional performance capabilities, especially important as AI models become larger and more resource-intensive. This partnership is notable because Google also operates its own competing AI products like Gemini, making the arrangement unusually cooperative among rivals.

While neither company has aggressively publicized the collaboration, industry insiders note that it benefits both sides: OpenAI gains compute redundancy and flexibility, while Google secures a prestigious AI customer that can showcase the power of its cloud offerings. The deal reportedly includes the use of advanced TPUs for inference and possibly training workloads, which had previously been used primarily by Google internally.

OpenAI’s recent growth has been exponential, with revenues surpassing $10 billion in annual run rate by mid-2025. This has necessitated an unprecedented expansion in infrastructure to meet user demand and support ongoing research. The decision to work with Google underscores OpenAI’s pragmatic approach, placing operational needs above platform loyalty or competitive positioning.

The development signals a broader industry trend where the sheer scale of AI infrastructure needs is compelling tech companies to work together, even when their products compete directly. In doing so, OpenAI and Google are shaping a new model of strategic cooperation in the AI era, where access to compute power is as crucial as the AI itself.

Sources: Reuters, Cloud Wars, TechCrunch

CloudAIdata center