Now that this round of the US stock AI rally has come this far, you can’t just blindly chase hot stocks.
The market’s disagreements are pretty significant right now:
On one side, big tech companies are continuously increasing capital expenditures, and the orders for computing power are real.
On the other side, valuations could lead to a deep pullback at any moment under valuation pressure. So my own approach isn’t an all-in bet. I prioritize holding the core positions with solid underlying logic, and I keep some cash reserved to wait for pullback opportunities.
On the hardware side, I’ve taken a small position in $MU , mainly betting on the HBM memory track. As cloud server capacity expands, HBM demand is hard and non-negotiable. The expectation for product price increases is still there, but this stock is extremely volatile. My position size is kept very low, so it’s only suitable for long-term planning—not for frequent short-term trading back and forth.
On the software side, I allocated a bit into Microsoft $MSFT . On the one hand, Azure cloud continues to capture demand for model computing power. On the other hand, Copilot has started rolling out as a paid offering for enterprise customers. Compared with pure-play chip manufacturers, Microsoft has both the compute base layer and AI software products, giving it a more balanced business structure. That should help smooth out some of the impact from industry-cycle swings.
I didn’t add any additional AI stocks. In my view, the long-term direction for AI is fine, but the biggest risk at this stage comes from valuation. If market risk appetite declines, pullbacks in high-growth stocks can be quite frightening. Rather than digging for every kind of theme everywhere, it’s more prudent to stay with companies whose revenue realization you can already see, and then gradually add to positions when the market falls.
#AI股持续上涨还有哪些投资机会
The market’s disagreements are pretty significant right now:
On one side, big tech companies are continuously increasing capital expenditures, and the orders for computing power are real.
On the other side, valuations could lead to a deep pullback at any moment under valuation pressure. So my own approach isn’t an all-in bet. I prioritize holding the core positions with solid underlying logic, and I keep some cash reserved to wait for pullback opportunities.
On the hardware side, I’ve taken a small position in $MU , mainly betting on the HBM memory track. As cloud server capacity expands, HBM demand is hard and non-negotiable. The expectation for product price increases is still there, but this stock is extremely volatile. My position size is kept very low, so it’s only suitable for long-term planning—not for frequent short-term trading back and forth.
On the software side, I allocated a bit into Microsoft $MSFT . On the one hand, Azure cloud continues to capture demand for model computing power. On the other hand, Copilot has started rolling out as a paid offering for enterprise customers. Compared with pure-play chip manufacturers, Microsoft has both the compute base layer and AI software products, giving it a more balanced business structure. That should help smooth out some of the impact from industry-cycle swings.
I didn’t add any additional AI stocks. In my view, the long-term direction for AI is fine, but the biggest risk at this stage comes from valuation. If market risk appetite declines, pullbacks in high-growth stocks can be quite frightening. Rather than digging for every kind of theme everywhere, it’s more prudent to stay with companies whose revenue realization you can already see, and then gradually add to positions when the market falls.
#AI股持续上涨还有哪些投资机会