Policy review + reasoning about the shift from chip development to inference— the winners in AI infrastructure are about to change seats. Those betting on training cards could get trapped.

News from the crypto community portal, as reported by OneMillion_AI: the recent token consumption index has fallen nearly 20% from its May peak, while the growth gap between AI investment and sales has reached 46%. Bear-market commentators point out that, according to Allianz Research data, the degree of imbalance has surpassed the 32% seen when the telecom bubble burst in 2001. Although the average token price has crashed by 90% since 2023, total spending has still grown by nearly 1 times, meaning the index retreat is only a structural digestion following price cuts designed to stimulate consumption. The strengthening of policy and regulation has translated into hidden compliance costs for corporate users. Washington has imposed even more stringent policy review on the distribution of frontier models and cross-border access, placing a heavy policy burden on leading platforms. Subtle changes are also emerging in the hardware chip sector, with the main force in market procurement beginning to shift from training chips to inference-optimized hardware, and the winner structure is being reshuffled.