Goldman Sachs believes the bullish logic for AI hasn't disappeared, but the market is shifting from a highly correlated "basket of AI trades" to repricing on individual themes.
During the July correction, sectors like Memory, AI Semiconductors, Optical Communications, Data Centers, and Neocloud were largely sold off simultaneously, almost a general liquidation. However, since August, the rebounds have shown significant divergence.
From their lows, Optical Communications rebounded by approximately 32%, Neocloud by about 20%, and AI Data Centers by about 17%, while Memory only rebounded by about 12% and AI Power by about 6%. Goldman Sachs believes this indicates that funds are beginning to re-differentiate the profit cycles, valuations, and fundamentals of different AI segments: a new "Inference Economy" theme has emerged for software, while Memory is shifting its focus from price increases and upward profit revisions to price stability, long-term agreements, and capital returns. The AI trading is not over, but the era of obtaining a uniform valuation premium solely based on the AI label is passing.
