Google's expanding Tensor Processing Unit (TPU) strategy is emerging as a serious challenge to Nvidia's long-running dominance in AI accelerators, particularly after a report from The Information revealed that Meta is in talks to begin using Google TPUs in its data centers in 2027 under a potential multibillion-dollar agreement.

Google's release of its Gemini 3 large language model (LLM) in November—trained primarily on the company's in-house TPU chips and performing at or above the level of OpenAI's ChatGPT—has become a catalyst for a broader strategic push. According to overseas reports, Google is now using its newest advances in AI models to pitch major clients, including Meta, on deploying TPU-based systems inside Google-operated data centers.