On November 25 (local time), NVIDIA's stock price plummeted by as much as 6.3%, hitting a new nadir since September 17. This decline drove its total market capitalization down to US$4.15 trillion, a staggering drop of nearly US$1 trillion from its all-time high of US$5.15 trillion recorded on October 29. Market analysts attribute the immediate catalyst for this steep decline to Meta's announcement of its intention to lease Google's in-house-developed TPU chips starting in 2026, a move that could potentially involve billions of dollars in orders.

However, the underlying pressures stem from deeper technological and competitive forces. Google's Gemini large language model, trained on TPUv5 chips, showcases markedly superior efficiency compared to NVIDIA's GPUs. Despite NVIDIA's assertion of its platform's supremacy, the market has begun to question the durability of its CUDA ecosystem's competitive moat.

Moreover, 'The Big Short' investor Michael Burry has weighed in, labeling NVIDIA as a bellwether for the impending burst of the AI bubble. He cautions that the market is susceptible to risks of oversupply and the looming threat of depreciation traps.