The asset was always the megawatts
Step back from the Bitcoin-versus-AI framing and a simpler picture appears: what the miners spent years acquiring was cheap, permitted, energized power capacity. Bitcoin mining was the workload that justified building it. Now a second workload — AI and high-performance computing — wants that same input, and wants it faster than new supply can be built. Bernstein's label for the miners is apt: the "power landlords" of the AI boom.
Why the lead time is the moat
The advantage is time, not just cost. Standing up new AI-ready capacity means securing land, contracting power, and waiting through grid interconnection and energization — a process that routinely runs to years, not months, on top of substantial capital. A miner that already operates energized sites has effectively pre-paid that delay. That is why AI developers are signing large, long-dated agreements with miners rather than building everything themselves, and why Bernstein projects the sector's AI revenue could climb from roughly $1.2 billion in 2026 to $10.7 billion by 2030.
Where Hut 8 fits
Hut 8 has leaned into exactly this identity — an infrastructure operator turning energized capacity into contracted AI infrastructure rather than a pure miner. Under the power-landlord thesis that is the higher-value position, and it aligns with Signal21's constructive standing view on the stock. The discipline that matters now is the quality of the contracts: capacity is only worth its "landlord" premium once it is leased, for years, to counterparties that can pay. Announced potential is not yet contracted revenue. This is market analysis, not a recommendation to buy or sell any security.