1

Wall Street is paying up for Bitcoin miners’ AI infrastructure before most of it is built

Why leased megawatts suddenly look like gold

Here’s the new weirdness: a megawatt that’s already leased to an AI or high-performance computing tenant is being priced way higher by investors than a megawatt sitting in a miner’s “coming soon” pipeline. In plain English: signed deals for AI capacity are trading like premium real estate while the stuff that hasn’t been delivered yet is still treated like vacant land.

Analysts who track miner finances have started valuing delivered AI-ready power at many times the simple energy output metric. Firms with solid, signed leases for AI or HPC customers attract valuation multiples that dwarf what pipeline-heavy miners get. The math behind that is straightforward — stable, recurring lease cash flows are less scary than unpredictable coin production — and Wall Street is rewarding the stability.

To put numbers on the fantasy land: some valuation models peg operating income per megawatt at roughly $1.5 million, then slap on an infrastructure-style multiple (think mid-teens). Gross value per megawatt can jump into the tens of millions, then you subtract the build cost (commonly estimated in the low tens of millions per megawatt) and suddenly you’ve got a meaningful chunk of equity value — assuming the project actually gets finished on time and on budget.

That last part — actually finishing on time and on budget — is the spicy bit. Bump construction costs by a few million per megawatt, or add a year to the schedule, and the shiny equity math gets notably less shiny. And because creditworthiness varies by tenant, a leased megawatt to a big hyperscaler is valued with a much friendlier cost of capital than one leased to a smaller, riskier cloud provider.

The giant funding gap and how miners might bridge it

So here’s the headline grabber: the industry faces a multibillion-dollar funding hole to turn all those announced AI sites into real, humming data halls. Estimates put the near-term shortfall in the tens of billions, with long-run needs swelling into the hundreds of billions if every plan is chased to completion. That’s a lot of cable, concrete, and coffee for the crews.

How do you fill that hole? The options are familiar but messy. Project finance and construction debt can get the build rolling, but they also add fixed obligations to companies that historically rode the rollercoaster of mining margins. Some miners may monetize Bitcoin reserves to raise cash, which is effective but changes the investment story for shareholders. Other routes include strategic partners, tenant prepayments, or equity raises — all of which dilute existing owners or hand a slice of future upside to whoever pays now.

Governance matters here in a way it didn’t when the business was just about hashing. Running billion-dollar infrastructure programs needs management teams and boards that can allocate capital at that scale. Weak governance that was fine during the mining era becomes a real liability when you’re selling long-term power capacity to corporate tenants.

There are two obvious scenarios. In the bull case, demand from hyperscalers stays hot, financing markets open for creditworthy projects, delivered megawatts start producing lease revenue, and the market permanently re-rates the best-executed miners toward data-center landlord multiples. In the bear case, capex overruns and pricey debt force equity issuance or asset sales, and many early shareholders end up funding the build only to see the upside seep away to new owners or lenders.

At the end of the day, Wall Street already seems to be pricing miners more like AI infrastructure landlords than pure crypto miners. The wild card is whether those price tags reflect real, recurring cash flow that will arrive on schedule — or whether they’re just a bet on paper promises that still need tens of billions of actual dollars and a lot of construction crews to become reality. Either way, it’s a fascinating pivot: from hashing boxes in a warehouse to long-term power deals and very expensive real estate for silicon-hungry customers.

Short version: investors love the idea of AI-ready power. Delivering it — on time, on budget, and to reliable tenants — is the plot twist that will decide who wins and who gets ghosted by the markets.