In a move shaking up the artificial intelligence industry, OpenAI has finalized an agreement to use Google’s cloud infrastructure, marking a significant pivot in its cloud strategy and signaling a rare collaboration between two of AI’s fiercest competitors. Sources told Reuters that the deal was completed in May following months of discussions, as OpenAI seeks to diversify its computing capabilities beyond longtime partner Microsoft.
This alliance grants OpenAI access to Google Cloud’s infrastructure, including its advanced tensor processing units (TPUs), expanding OpenAI’s ability to train and deploy large-scale AI models such as ChatGPT. The deal also underscores the immense computing needs driving today’s AI race, with cloud capacity emerging as a critical resource for the world’s top tech firms.
OpenAI’s partnership with Google follows its earlier moves to secure compute from SoftBank, Oracle, and CoreWeave as part of its massive Stargate infrastructure initiative. The company is also developing its own in-house AI chips, aiming to reduce long-term dependency on external suppliers. According to sources, OpenAI had been limited by exclusivity clauses in its arrangement with Microsoft Azure until earlier this year, but with those restrictions lifted, it is now engaging with other major players in cloud computing.
The agreement is also a major win for Google Cloud, which brought in $43 billion in revenue in 2024 and is aggressively expanding its AI client base. The addition of OpenAI, even as it competes directly with Google’s DeepMind and Gemini projects, strengthens Google’s positioning as a neutral infrastructure provider and bolsters its case against top cloud rivals Microsoft and Amazon. Google’s in-house TPU chips, once reserved for internal use, are now being leveraged to serve both corporate giants like Apple and AI startups such as Anthropic and Safe Superintelligence.
However, the partnership comes with internal tension. Selling compute to a rival like OpenAI means allocating limited chip resources that could otherwise be used to advance Google’s own consumer-facing AI tools. CFO Anat Ashkenazi recently acknowledged that Google Cloud struggled to meet demand in the last quarter, further complicating CEO Sundar Pichai’s task of balancing internal needs with external growth.
Despite ChatGPT’s continued dominance over Google’s Gemini chatbot in terms of monthly active users, Pichai has publicly dismissed concerns that OpenAI threatens Google’s broader business. Still, analysts caution that this collaboration could reshape AI’s competitive landscape as firms blur the lines between rivals and strategic partners, all in pursuit of greater computing power.
