Posted 2024-10-19T21:50:54, updated 2024-10-24T22:09:44 20 0.9489 stocks

AI clusters are completely different than legacy data centers. Legacy centers have tons of legacy low power density properties. That's great for stuff like delivering Netflix. AI compute is the exact opposite. They haven't been multibaggers because they're like legacy auto businesses trying to ramp up an EV production line. They're not a pure play. That's why the AI data center darling is a converted Ethereum miner called Corweave (private company unfortunately). They incredibly jumped from one gold rush to another. They were in the business of dense GPU compute just as Ethereum switched from Proof of Work and AI ramped up. They made a landmark hosting deal with Core Scientific which is a publicly traded Bitcoin miner which would be the closest proxy bet if they keep adding more AI contracts. Other miners like Iren have smaller deals with AI firms like Poolside. You really want power dense assets that aren't currently doing AI but could transition from a lower margin business and massively reprice up. Right now Bitcoin miners are the most obvious play I know of. The miners have hoarded real land, energy, and infrastructure assets which can have a *huge* re-rating if they get an AI compute contract like CORZ did with Coreweave, APLD with NVDA, or get bought out by a hyperscaler. WULF popped recently after the nuclear deals due to their proximal nuclear assets. They also have major convex exposure to BTC. That's because you're paying for their mining margin with these stocks. For example, if they are making 10k per BTC mined at 50K, and BTC spikes to 100k, they are suddenly making 60k per BTC mined. Or 6x the margin for a 2x BTC price increase. There are risks of course. Execution risk, dilution risk, geographic risk, BTC price risk, etc. But when their energy infra valuations are near book value during an energy gold rush and BTC is lagging the QQQ's currently they have huge upside asymmetry on two fronts.

WULF NVDA