Key insights
- The article discusses Bitcoin miners transitioning into AI infrastructure plays, highlighting that their existing power and site infrastructure gives them a competitive advantage. Hut 8 and Iris Energy are mentioned as potential genuine AI infrastructure plays, while KEEL is viewed as overvalued. The key takeaway is that access to power and infrastructure, rather than just computing power, is the real bottleneck for AI, making some miners attractive.
I keep seeing Bitcoin miners marketed as “AI infrastructure” stocks. Some of that feels like buzzword dressing. Some of it might be real.
Large scale AI datacenters need a few boring but hard things cheap power, grid capacity, cooling, and industrial sites you can actually run 24/7. Miners accidentally built a lot of that..
A few quick, high level takes:
- Hut 8 (HUT) – Leaning into AI and HPC, with existing sites and power contracts that can be repurposed rather than built from scratch. * Iris Energy (IREN) – Looks closer to a datacenter operator with strong renewable power access than a pure miner. * KEEL – More optionality than proof so far. The stock already prices in a lot of future “AI infra” without much track record.
You can always buy more GPUs if you have the cash. You cannot quickly permit new substations, secure long term power, or pour concrete for massive cooling facilities. That time lag is what makes some of these miners interesting as AI plays, not the fact they mentioned “AI” on one earnings call...
This is not a pitch for any of them. I am more interested in the framing.
Are there other miners you see as genuine AI infrastructure plays rather than just high beta Bitcoin exposure with an “AI” label on top?
Wenn du willst, mache ich dir als Nächstes direkt den „AI’s real bottleneck: power, not compute?“ Post in der gleichen Schärfe klar.