OpenAI begins using Google’s AI chips in unlikely alliance between rivals

In a surprising twist in the AI arms race, OpenAI has started renting Google’s artificial intelligence chips to power ChatGPT and its other tools, according to a source familiar with the matter.

The collaboration, initially reported exclusively by Reuters earlier this month, is seen as a strategic step by OpenAI to diversify its computing capacity and potentially lower the substantial costs associated with AI inference – the process where trained AI models make predictions and decisions. OpenAI is one of the largest purchasers of Nvidia's GPUs for both model training and inference computing.

The move marks a significant shift for OpenAI, which has until now been one of the largest buyers of Nvidia’s GPUs, the high-performance chips critical to training and running large AI models. It also reflects a departure from the company’s near-exclusive reliance on Microsoft’s Azure data centers, despite Microsoft being OpenAI’s primary investor and infrastructure partner.

For Google, this deal represents a major win for its cloud business and its in-house TPU technology. Historically, Google's TPUs were primarily reserved for internal use. However, the tech giant has been expanding their external availability, successfully attracting other major customers, including Apple and AI startups like Anthropic and Safe Superintelligence – both founded by former OpenAI leaders.

Despite this surprising partnership, the competition between the two AI powerhouses remains. A Google Cloud employee reportedly indicated that Google is not renting its most powerful TPUs to OpenAI, likely reserving its most advanced hardware for its own AI development, including its Gemini models.

Implications for Microsoft and Nvidia

OpenAI’s decision could have ripple effects across the AI ecosystem. It raises questions about the durability of its partnership with Microsoft, whose cloud services have been central to OpenAI's growth. It also underscores a growing trend of major AI labs looking to hedge against reliance on Nvidia, whose dominance in the AI chip market has led to high costs and limited availability.