OpenAI to Acquire Google’s TPU Chips to Reduce Nvidia Dependence

OpenAI is reportedly planning to purchase Google’s Tensor Processing Unit (TPU) chips, a move that would lessen its reliance on Nvidia, the primary supplier of AI hardware. The Information reports that OpenAI began renting TPUs in June 2025 and has now hired an engineer to optimize its frontier models for them. The company is also developing its own Broadcom‑built chip, Jalapeño, slated for deployment by the end of 2026. Alphabet’s CFO expects most revenue from TPU system sales to come in 2027. No price, volume, or delivery dates have been disclosed, and neither company has confirmed the plan. The acquisition would allow OpenAI to own the hardware rather than pay hourly usage fees to Google. This development signals a significant shift in the AI hardware ecosystem, as OpenAI seeks to diversify its supply chain beyond Nvidia’s dominance. The move could influence pricing, availability, and competitive dynamics among AI hardware providers, impacting the broader AI and data‑center industry.