Cathie Wood: The impact of this tech revolution may exceed the Industrial Revolution! I’m more concerned about where the money for AI actually goes.
Personally, I care less about the conclusion of “exceeding the Industrial Revolution,” and more about whether AI can truly improve production efficiency and corporate profits.
If AI only boosts efficiency, it’s an industrial upgrade; if it further changes the cost structures of R&D, manufacturing, software, logistics, and finance, then that’s the real revolution in productivity.
So I’m focusing on three tracks right now:
**① Compute infrastructure:** GPUs, high-speed networks, optical modules, HBM, advanced packaging—these are the most direct “shovels” for the AI expansion.
**② Power and cooling:** Data centers are consuming ever more electricity. Infrastructure such as power grids, energy storage, liquid cooling, and UPS may become the new bottleneck for AI expansion.
③ AI applications:** Robotics, autonomous driving, AI agents, enterprise software. I focus on who can truly turn AI into orders, revenue, and profits.
But I need to remind everyone: a technological revolution is real, but it doesn’t mean AI stocks are automatically cheap right now.
The easiest mistake the market makes is to price all of next few years’ growth into the stock price in advance.
In my view, the next phase of the AI cycle will gradually shift from “trading technology and concepts” to “trading performance and trading cash flow.” The companies worth laying in may not be the ones best at telling AI stories, but those that can continuously capture AI capital expenditures and convert technology into profits.
Technology determines the room for growth, performance determines the trend, and valuation determines the entry point.
Personally, I care less about the conclusion of “exceeding the Industrial Revolution,” and more about whether AI can truly improve production efficiency and corporate profits.
If AI only boosts efficiency, it’s an industrial upgrade; if it further changes the cost structures of R&D, manufacturing, software, logistics, and finance, then that’s the real revolution in productivity.
So I’m focusing on three tracks right now:
**① Compute infrastructure:** GPUs, high-speed networks, optical modules, HBM, advanced packaging—these are the most direct “shovels” for the AI expansion.
**② Power and cooling:** Data centers are consuming ever more electricity. Infrastructure such as power grids, energy storage, liquid cooling, and UPS may become the new bottleneck for AI expansion.
③ AI applications:** Robotics, autonomous driving, AI agents, enterprise software. I focus on who can truly turn AI into orders, revenue, and profits.
But I need to remind everyone: a technological revolution is real, but it doesn’t mean AI stocks are automatically cheap right now.
The easiest mistake the market makes is to price all of next few years’ growth into the stock price in advance.
In my view, the next phase of the AI cycle will gradually shift from “trading technology and concepts” to “trading performance and trading cash flow.” The companies worth laying in may not be the ones best at telling AI stories, but those that can continuously capture AI capital expenditures and convert technology into profits.
Technology determines the room for growth, performance determines the trend, and valuation determines the entry point.