If you asked me to bet on the direction for the next round that could take over from AI chips, I wouldn’t chase GPUs anymore. I’m going to focus on a segment that many people haven’t really paid attention to yet: AI power grids + electrical equipment.
Among them, the one I care most about is $ETN.US N Eaton.
Why?
In the past two years, the biggest question for the AI industry has been: Is there enough GPU?
But as data centers keep getting built, the problem is gradually shifting to another one—Is there enough power?
Microsoft, Meta, Google, and Amazon can keep spending money to build data centers. NVIDIA can keep selling more GPUs. But in the end, servers still have to land on one real-world question: Where does the electricity come from, and how is it delivered in a stable way? That’s exactly where Eaton fits in. Transformers, distribution equipment, switchgear, UPS systems, power management… From feeding power into the data center via the grid, all the way to distributing it to server racks, there are countless pieces of equipment you can’t bypass. That’s why I look at ETN—not just because of the “AI needs more electricity” story, but because once AI capex truly comes to earth, the power infrastructure has to expand alongside it.
If you break down the AI market over the past few years, it’s roughly like this:
First round: $NVDA, $AVGO — compute power
Second round: $MU, $SK Hynix, $SNDK — storage
Third round: $LITE, $COHR, $ANET — optical communications and networking
For the next round, I’d rather look at power.
My current priority order is: ETN > PWR > GEV
ETN leans more toward electrical equipment, PWR is more focused on grid construction, and GEV covers power generation and grid infrastructure. Of course, good logic doesn’t necessarily mean the stock price will definitely rise. What I care about more is whether demand continues to be realized, whether orders keep growing steadily, and whether the valuation has returned to a reasonable range.
As AI spreads into the next stage, the market may not always be chasing the next chip. Sometimes, the real opportunity may be hidden in an unremarkable transformer. @K线猎人-
Among them, the one I care most about is $ETN.US N Eaton.
Why?
In the past two years, the biggest question for the AI industry has been: Is there enough GPU?
But as data centers keep getting built, the problem is gradually shifting to another one—Is there enough power?
Microsoft, Meta, Google, and Amazon can keep spending money to build data centers. NVIDIA can keep selling more GPUs. But in the end, servers still have to land on one real-world question: Where does the electricity come from, and how is it delivered in a stable way? That’s exactly where Eaton fits in. Transformers, distribution equipment, switchgear, UPS systems, power management… From feeding power into the data center via the grid, all the way to distributing it to server racks, there are countless pieces of equipment you can’t bypass. That’s why I look at ETN—not just because of the “AI needs more electricity” story, but because once AI capex truly comes to earth, the power infrastructure has to expand alongside it.
If you break down the AI market over the past few years, it’s roughly like this:
First round: $NVDA, $AVGO — compute power
Second round: $MU, $SK Hynix, $SNDK — storage
Third round: $LITE, $COHR, $ANET — optical communications and networking
For the next round, I’d rather look at power.
My current priority order is: ETN > PWR > GEV
ETN leans more toward electrical equipment, PWR is more focused on grid construction, and GEV covers power generation and grid infrastructure. Of course, good logic doesn’t necessarily mean the stock price will definitely rise. What I care about more is whether demand continues to be realized, whether orders keep growing steadily, and whether the valuation has returned to a reasonable range.
As AI spreads into the next stage, the market may not always be chasing the next chip. Sometimes, the real opportunity may be hidden in an unremarkable transformer. @K线猎人-
