📌 AI Capital Is Rotating Toward Platforms, Not Cooling on Hardware
AI trading is making a relative rotation from chip stocks to large platform stocks. On the day, both SPY and QQQ rose only about 0.2%, Nvidia fell about 1.7%, and AMD dropped about 1%; Microsoft rose about 2.5%, Meta gained about 2%, and Google also moved higher. This indicates that the slight rise in the indices was not broad-based buying, as capital is more willing to pay for platform companies that can directly package AI capabilities into products and services.
But this cannot yet be interpreted as weakening demand for AI infrastructure. According to Micron, data-center memory demand remains so strong that the company cannot meet about half of customer demand, and buyers are not very price-sensitive either. Tight storage supply means order and pricing pressure across the hardware chain has not disappeared. Chip stocks’ weakness that day is more likely to reflect capital adjusting positions among different AI segments, rather than a sudden collapse in demand.
The real disagreement lies in the path to earnings realization. Platform companies can generate revenue through the combination of existing channels such as cloud services, software, and advertising with AI products, making them easier to view as a layer closer to monetization. However, platform stocks are not without costs. Taking Microsoft as an example, its current price is considered to already factor in Azure maintaining growth above 40%, as well as the pressure from AI infrastructure on gross margins gradually easing in the latter part of FY27.
Therefore, the core of this rotation is not that “hardware has lost demand,” but that capital is comparing who can turn large computing-power investments into revenue and profit more quickly. If cloud-business growth or margin improvement falls short of expectations, the relative strength of platform stocks will also need to be reassessed; if storage supply and demand remain tight, the explanation of cooling hardware demand will become harder to sustain.
▌ Sources