If you’re watching mining stocks, AI compute, or the data-center business, lock in one split that’s already happening: the identity “Bitcoin miner” is fracturing into two completely different companies, and their valuation logic and risk exposure are moving in opposite directions.

One species keeps betting purely on mining — earning block rewards, living and dying with BTC’s price and the halving cycle. It’s essentially a leveraged proxy for Bitcoin’s price. The other species is converting the power capacity and data-center sites it built up over years of mining into infrastructure that can host AI/HPC compute, leasing it to AI cloud providers under 5-to-15-year contracts for a fixed hosting fee — a business with far more predictable cash flow than mining, one that deserves a data-center/cloud-infrastructure valuation, not a Bitcoin-leverage one.

Core Scientific is the most complete transformation on record. It signed a 12-year, $10-billion-plus compute-hosting contract with AI cloud provider CoreWeave, and its HPC revenue share is projected to jump from roughly 5% in 2024 to roughly 71% in 2026. At that ratio, its core cash flow is now driven mainly by fixed AI-hosting income, and its sensitivity to Bitcoin’s price has dropped sharply — it should be valued as an emerging AI data-center company, not a miner.

But the detail most worth remembering is a strategic reversal: CoreWeave had proposed to acquire Core Scientific outright in 2025 — and that merger agreement was terminated on October 30, 2025 (the hosting contract itself stayed in force). Going from “we’re going to buy you” to “we’re staying a customer” is a reminder that the AI-transition narrative is sexy, but execution is far messier than a press release.

The real obstacle is money. Institutions including VanEck estimate the mining industry’s AI transition faces a near-term funding gap of roughly $50 billion, with long-term capital needs potentially reaching $221 billion. The market is shifting from trading transition headlines to scrutinizing something much plainer: who can actually raise the money, build the data center on schedule, and run it reliably. Signing a contract is easy. Delivering a data center is a different order of difficulty.

So stop judging transitioning miners by hashrate scale (EH/s) — that metric is already outdated. The real dividing line is three new capabilities: fundraising ability, engineering-delivery capability, and customer-relationship management. Whoever proves first that its power assets can reliably convert into AI-compute revenue completes the leap from “Bitcoin-price leverage” to “infrastructure company.”

Is the AI pivot in mining a real second growth curve, or just this bear market’s favorite story for capital markets?

— Adapted from Crypto Sector Leaders, Chapter 5: Mining and Hashrate — From Core Mining Business to AI/HPC Compute Hosting

#BitcoinMining #AICompute #DataCenters