TECH NEWS – Nvidia could lose a major client as this tech company may soon switch to a rival’s chips…

According to a source cited by The Information, Google has managed to secure a portion of OpenAI’s AI computing workload from Nvidia. OpenAI has turned to Google’s Tensor Processing Units (TPUs) to power its AI products. Until now, it was widely believed that OpenAI primarily relied on Microsoft, Oracle, and others to access Nvidia GPUs for training and running its AI software. The report indicates that OpenAI is now using Google’s TPUs to cut operating costs—and Google has not leased its latest generation of TPUs to AI rivals.

Google launched its seventh-generation TPUs in April, which it claimed were the first chips optimized specifically for AI inference. These TPUs made headlines last year when Apple revealed it had used Google’s hardware to train its Apple Intelligence platform. But training is only part of the AI equation—inference is the process that actually delivers output based on user prompts. OpenAI is now using these TPUs to run ChatGPT and other AI products, with the switch reportedly occurring recently.

This strategic pivot stems from OpenAI’s need to reduce costs amid rising demand, especially after the launch of its image generation platform earlier this year. Google reportedly plans to offer its TPU chips to cloud infrastructure providers. With Nvidia’s high prices and GPU shortages, there’s room for competitors offering more affordable solutions to gain market share.

If Google succeeds in convincing these providers to adopt its chips, it could significantly cut into Nvidia’s dominance—especially given the ongoing supply constraints. Google also powers its own AI models using its proprietary hardware. The company is considered one of the most versatile AI players in the industry, spanning both hardware and software. In addition to TPUs, Google offers the Gemini AI model, which has already been integrated into Gmail and Google Search.

So far, OpenAI has relied heavily on Nvidia GPUs due to its partnerships with Microsoft and Oracle. Oracle has positioned itself as one of the largest holders of Nvidia GPU inventory in the industry. But if OpenAI shifts more of its operations to Google TPUs, it could potentially break Nvidia’s near-monopoly in the high-end AI sector.

Source: WCCFTech, The Information