Google aims to match NVIDIA's AI chip output with massive TPU v9 deployment by 2028

Market analysis from August indicates Google plans to deploy between 12 and 15 million TPU v9 AI chips by 2028, a production volume comparable to NVIDIA's estimated output of approximately 12.4 million GPU units in the same timeframe. The new processors employ a quad-core architecture, substantially increasing demands for advanced manufacturing processes and chip packaging capacity. This requirement is pushing Google to seek production partnerships beyond Taiwan Semiconductor Manufacturing Company, potentially benefiting Intel and Samsung's foundry services. The initiative underscores intensifying competition for AI processing capacity, with Google pursuing supply chain independence and greater strategic control over its computational infrastructure.