To be honest, every time the AI sector gets hyped up, I feel that instead of guessing who will be the next NVIDIA, it's better to just buy the "shovel sellers."

Applied Materials follows this logic—no matter whether TSMC, Samsung, or Intel wins, the equipment to make chips has to be bought from $Applied Materials(AMAT.US). The more intense the AI arms race becomes, the fuller their order book gets.

The stock price has risen from last year's low of $128 to nearly $400 now, tripling in value. Many people think it's too expensive and are afraid to chase it. But my view is: the AI capital expenditure cycle has just started, and it will take several more years for 2nm and GAA processes to be fully rolled out. The visibility of equipment demand is much stronger than that of most tech stocks.

Of course, the risks are real: the valuation isn't cheap, export controls are a potential landmine at any time, and insiders have been selling recently.

My own strategy is to hold without adding to the position, waiting for a pullback before considering more.