Over the last 90 days, the correlation between Bitcoin and the largest mining stocks has collapsed to as low as 16% for Core Scientific, 31% for Riot Platforms, and 33% for IREN. Meanwhile, MicroStrategy (MSTR) still holds a 78% correlation. If you bought mining stocks thinking you were buying Bitcoin exposure, you are now holding a very different asset. The math doesn't care about your conviction — it cares about structural changes in revenue models.
Tom Lee, the veteran strategist, recently published a ranking of 17 crypto-related stocks based on their 90-day rolling correlation to Bitcoin and Ethereum. The goal was simple: help investors find the best equity proxies for crypto exposure. But the data revealed something more profound. The top of the list was dominated not by miners, but by treasury companies and exchanges. BitMine led ETH correlation at 80%, Coinbase followed at 74%, and MicroStrategy sat at 78% for Bitcoin. The mining companies — Core Scientific, TeraWulf, IREN, Riot — all landed in the bottom half, with correlations barely above 20%.
Here's the catch: Tom Lee is the chairman of BitMine. That doesn't invalidate the data, but it demands a higher standard of skepticism. As I've learned from years auditing DeFi protocols and managing yield strategies, conflicts of interest are the most common source of blind spots. The data is likely correct, but the interpretation needs independent verification.
The core insight is not the ranking itself, but the structural shift it reveals. Mining companies are no longer pure plays on Bitcoin's hash rate. They are becoming AI infrastructure landlords. The evidence is in their financial statements. Core Scientific reported that AI revenue now accounts for over 50% of its total income. TeraWulf's CFO explicitly stated that future earnings will be driven by long-term contracts with AI firms, not by Bitcoin's price. IREN, which still has the highest BTC correlation among miners at 33%, is rapidly building out data centers for AI compute.
This is not a narrative. It's a reclassification of assets. The business model of a miner has shifted from 'convert electricity into Bitcoin' to 'lease power and compute capacity to AI companies at a margin.' The revenue stream has changed from volatile block rewards to recurring contract income. Consequently, the stock's price drivers have shifted from Bitcoin's price to AI demand, power contracts, and data center utilization rates.
Audits don't guarantee safety — and neither do correlation scores. A 90-day correlation is a snapshot, not a promise. It captures the recent past, not the structural future. The market has partially priced in this shift, but the majority of retail investors still treat mining stocks as leveraged Bitcoin ETFs. This is the largest cognitive disconnect in the current market.
Let's look at the numbers. MicroStrategy's business model is straightforward: buy and hold Bitcoin, raise debt, buy more Bitcoin. Its stock price is a levered function of Bitcoin's price. The correlation of 78% is expected and sustainable. Coinbase's correlation to Ethereum at 74% reflects its core revenue from trading fees, staking, and custody. These are direct crypto exposures. But a mining company like Core Scientific, with a 16% correlation to Bitcoin, is effectively a different asset class. Smart money is already repositioning. The hedge funds I track are increasing their allocations to MicroStrategy for Bitcoin exposure and using miners only as a play on AI infrastructure.
The contrarian angle is this: Most investors still believe mining stocks are a leveraged crypto trade. They see the stock price lagging Bitcoin and think it's a buying opportunity — a discount. But the discount is structural, not cyclical. The stock is not lagging; it's decoupling. If Bitcoin rallies 50% from here, there is no guarantee that Core Scientific or TeraWulf will follow. They might, but only if the AI narrative also strengthens. Conversely, if AI hype fades, miners could lose both their AI premium and their Bitcoin exposure, creating a double hit.
The fastest way to lose money is to confuse narrative with fundamentals. The narrative is that miners are pivoting to AI. The fundamentals are that some have already executed this pivot at a cost. Mara and CleanSpark, two miners that aggressively pursued AI transformation, have collectively lost $851 million in the process. The pivot is not free. It requires massive capital expenditure, dilutive equity offerings, and operational risk. The market is currently rewarding the narrative, but the financials are mixed.
Consider the risk matrix. If you are a Bitcoin bull, the last thing you want is an asset that responds to AI demand rather than Bitcoin's price. If you are an AI infrastructure bull, miners offer a cheaper entry point than traditional data center REITs, but with higher execution risk. The investors who will suffer most are those who buy miners for the wrong reason — expecting a tight Bitcoin correlation — and then hold through a divergence.
The wormhole is a bridge, not a bank. In this context, the mining stock is a bridge to AI infrastructure, not a bank for Bitcoin exposure. The investor who wants pure Bitcoin exposure should use the simplest tool: the spot ETF, or MicroStrategy. The investor who wants to bet on the AI-crypto convergence should pick miners with the strongest contract visibility and lowest debt.
What does this mean for your portfolio? First, re-evaluate your holdings. Check the 90-day correlation of any mining stock you own. If it's below 40%, ask yourself: Why am I holding this? Is it for Bitcoin exposure, or for AI exposure? Second, monitor the next earnings season. The real test will be whether AI revenue growth can offset the decline in mining revenue from Bitcoin's halving. If the answer is yes, the decoupling will accelerate. If no, the stocks will suffer a double penalty.
Takeaway: The market is in the process of repricing mining stocks from crypto beta to AI infra alpha. This repricing is incomplete. The risk is not that the correlation is low today — it's that investors will refuse to update their mental models. The next time you see a mining stock trade at a discount to Bitcoin's price, pause. It might not be a discount. It might be a different asset entirely.
Your exit liquidity is someone else's alpha. Don't be the one holding a Bitcoin proxy when the market has already reclassified it as an AI data center REIT. The math doesn't care about your previous assumptions. The only question that matters: Are you positioned for the reality of the next 90 days, or the memory of the last 90?