Everyone talks about how many GPUs AI needs. But there’s a problem: You need enough electricity to actually turn them on. AI data centres require MASSIVE amounts of power, and securing enough electricity is becoming one of the biggest bottlenecks of the AI boom. So one of the biggest opportunities might not just be the companies making the chips. It could be the companies powering them. Here’s how I think about the AI power trade: 1. Power Generation ⚡ More data centres = more electricity demand. Companies like $VST and $CEG generate huge amounts of electricity, while $BE offers onsite power generation that can reduce reliance on new grid capacity. Data centres don’t just need more electricity. They need huge amounts of it, 24/7 and in specific locations. The bet: More AI → more data centres → more electricity demand. 2. Natural Gas 🔥 AI companies can’t always wait years for new power infrastructure. Natural gas can provide large amounts of reliable electricity, creating opportunities from producers like $EQT, to pipelines like $WMB, to turbine manufacturers like $GEV. The bet: AI needs power NOW → natural gas helps fill the gap. 3. Nuclear ☢️ Data centres need massive amounts of reliable, 24/7 electricity. That’s helped put power producers like $CEG, uranium companies like $CCJ, and nuclear technology companies like $OKLO in the AI power conversation. The bet: AI needs reliable power around the clock → nuclear becomes increasingly valuable. 4. The Grid 🔌 You can generate all the electricity in the world… But you still have to get it to the data centre. That requires transformers, substations, transmission lines and other electrical equipment. Companies like $ETN, $GEV and $PWR help build the infrastructure connecting new power demand to the grid. The bet: More electricity demand → more spending on the infrastructure that moves it. 5. Power + Cooling ❄️ Getting electricity to the building still isn’t enough. AI servers consume enormous amounts of power and generate enormous amounts of heat. Companies like $VRT and $ETN provide the cooling, power management and electrical infrastructure needed to keep those GPUs running. The bet: More powerful AI chips → more infrastructure needed around every chip. So the AI power trade isn’t one stock. It’s an entire chain: AI → Compute → Data Centres → Electricity → Grid → Power + Cooling Everyone is trying to figure out whether $NVDA, $GOOGL, $MSFT, $META or the next AI model will be the biggest winner. But underneath all of them is the same requirement: Power. You don’t necessarily have to predict who wins AI. Sometimes you can invest in what every winner needs.
82 views
2 Comments
Blossom
@blossom · 39m
Blossom
@blossom · 39m
See the full comment section 👀Sign up for the full Blossom experience!
