OpenAI says it has no plan to use Google's in-house chip

OpenAI has clarified it has no active plans to deploy Google's in-house Tensor Processing Units (TPUs) at scale, despite recent reports suggesting a significant shift in its compute strategy. While the AI lab is conducting early tests with Google's chips, it continues to primarily utilize Nvidia's GPUs and AMD's AI chips for its growing compute demands, alongside developing its own custom silicon. This statement underscores OpenAI's diversified hardware strategy and clarifies the competitive landscape among AI chip providers, even as Google expands external TPU availability to other prominent AI firms.

Analysis

OpenAI's clarification that it has no active plans to deploy Google's Tensor Processing Units (TPUs) at scale significantly tempers recent market speculation about a major shift in its hardware strategy. While confirming early testing of Google's chips, the statement reinforces the incumbent positions of Nvidia (NVDA) and AMD (AMD) as the primary suppliers for OpenAI's current, high-demand compute needs. This development represents a setback for Google's (GOOGL) ambition to have its in-house chips adopted by the definitive AI market leader, even as it successfully secures other high-profile customers like Apple and Anthropic. The news also highlights OpenAI's sophisticated, multi-pronged approach to securing computing power, which involves diversifying across established chipmakers, exploring new hardware, utilizing specialized cloud providers like CoreWeave (CRWV) for the majority of its GPU capacity, and pursuing vertical integration through its own chip development, which is reportedly on track for a 'tape-out' milestone this year. This strategy is designed to mitigate supply-chain risk and manage long-term costs in the intensely competitive AI landscape.

Key Decisions for Investors

• This news reinforces the market-leading positions of Nvidia and AMD, as OpenAI's clarification confirms its continued reliance on their hardware and downplays the immediate competitive threat from Google's TPUs for this key account.
• Investors in Google should moderate expectations for its TPU business; the failure to secure a large-scale commitment from a marquee customer like OpenAI underscores the significant challenge of displacing entrenched competitors in the high-end AI training market.
• The key long-term development to monitor is OpenAI's progress toward its in-house chip 'tape-out' milestone, as a successful outcome would signal a strategic move that could eventually reduce its reliance on all external suppliers.