Hua Er Jie Jian Wen·2025-11-26 08:51Core Insights - The AI chip market is experiencing turbulence as rumors suggest that Meta, a key customer of Nvidia, is considering adopting Google's custom AI chips, Tensor Processing Units (TPUs), potentially starting next year [1] - This news led to a significant market reaction, with Nvidia's stock dropping by 6% and AMD's by 10%, while Alphabet's stock rose by 4% [1] - Both Google and Nvidia responded to the market concerns, emphasizing their commitment to supporting both TPUs and Nvidia GPUs [1] Group 1: Market Reactions - Nvidia's stock ultimately closed down approximately 2.6%, while Alphabet's stock closed up 1.6%, reaching a new historical high [1] - The market's reaction indicates heightened competition and uncertainty surrounding Nvidia's dominance in the AI chip sector [1][4] Group 2: Analyst Perspectives - Citigroup's report acknowledges the growing competition from custom AI accelerators like Google's TPUs and Amazon's Trainium, but maintains that Nvidia's market share will remain high, predicting a decline from 90% in 2025 to 81% by 2028 [2] - Analysts note that major companies like Microsoft and Meta still heavily rely on Nvidia's platform due to delays in their own custom chip projects [2] Group 3: Nvidia's Defensive Strategy - Nvidia has taken unusual defensive communication measures, publicly asserting its technological superiority over TPUs and emphasizing its ability to run all AI models [5] - The company also distributed a detailed memo to Wall Street analysts addressing various criticisms, which some analysts interpreted as a sign of insecurity [5][7] Group 4: Competitive Landscape - Google's AI model, Gemini 3, has been trained entirely on its TPUs, enhancing the credibility of TPUs as a viable alternative to Nvidia's products [6] - The competitive landscape is shifting, with major players like Google demonstrating significant capabilities in AI chip development [6]