Bitcoin miners stumbled into the AI gold rush. During the 2021 hash rate migration from China, companies like Marathon, Riot, and CleanSpark locked down land and interconnection agreements while Bitcoin margins were high. Brandon Bailey noted on TFTC that these miners inadvertently built a pipeline of energized real estate that hyperscalers cannot replicate quickly.
That pipeline is now the fastest path to new compute. AI hyperscalers are hitting a wall with grid interconnection wait times, making existing, powered mining sites prime targets. Bailey argues that by simply changing the offtaker of their power from a Bitcoin miner to an AI firm like Microsoft or Google, these companies can trigger a massive valuation arbitrage. Bitcoin mining trades between four and six times EBITDA. Traditional data center giants like Equinix trade at 20 to 24 times EBITDA.
"Bitcoin miners transitioned to AI compute because a 10-15 year hyperscaler lease offers guaranteed cash flow, unlike volatile mining income."
- Brandon Bailey, TFTC: A Bitcoin Podcast
The AI energy bottleneck is tightening globally. On Macro Voices, Lyn Alden argued that while the U.S. leads in chip design, China holds a structural advantage in the power systems required to run them, thanks to its massive industrial base. Eric Townsend pointed to a looming social conflict in the U.S. between AI energy demand and consumer utility bills.
The economics are staggering. John Tinsman, speaking on TFTC, highlighted Elon Musk’s Colossus facility, built in 112 days on a $4 billion investment, which could generate $15 billion annually in leasing revenue. Marty Bent cited Morgan Stanley data showing AI spending has made the U.S. economy inelastic to interest rates; when a project yields a 100% return on capital, a 5% interest rate is a rounding error.
On Forward Guidance, Tyler Neville argued this capex boom is rerating Big Tech multiples. Firms like Google and Microsoft have moved from buyback-rich cash cows to leveraged entities trapped in an AI capex hamster wheel, comparing them now to the volatile, capital-intensive oil industry.
The pivot isn't simple engineering. Bitcoin mining rewarded speed and cost-cutting; AI training requires five-nines uptime, massive redundancy, and sophisticated cooling. Bailey described a two-phase re-rating process: signing a lease creates the first valuation pop, and successfully financing and delivering construction removes the remaining execution discount. Core Scientific, with 590 megawatts leased to CoreWeave, has emerged as the pioneer.
"Hyperscalers are being re-rated from cash-flow rich to leveraged businesses, capping their upside."
- Tyler Neville, Forward Guidance
The energy squeeze will push Bitcoin mining toward stranded power. Bailey concluded the AI compute rush will force mining toward distributed and stranded energy sources, potentially creating a more durable era for mining economics. Dave Bennett, on Bitcoin And, proposed an even more fringe model: using autonomous robots to gasify forest debris to power Bitcoin miners, solving fire risk while creating a revenue stream. The model highlights the creative extremes miners are considering as the energy race intensifies.


