The market assumes a mining stock is a levered bet on Bitcoin. That assumption is now a structural error.
Tom Lee, the veteran strategist, recently published a ranking of 17 crypto-related stocks by their 90-day rolling correlation to Bitcoin and Ethereum. The list placed MicroStrategy at the top for Bitcoin, with a 78% correlation. Core Scientific, a major mining operator, appeared near the bottom, with a mere 16% correlation to Bitcoin. The data is clean. The implication is brutal.
This is not a statistical anomaly. It is a structural break.
The ranking, published on a financial news platform, covered companies with market capitalizations above $2 billion. The methodology was straightforward: measure the 90-day rolling correlation coefficient between each stock and the spot price of BTC and ETH. The results were then ranked from highest to lowest correlation. The intention was to help investors identify which equities provide the most effective crypto exposure. But the output tells a different story.
The context here is not technical — it is industrial.
During the 2020 DeFi Summer, I modeled the correlation between Uniswap V2 liquidity depth and global M2 money supply. That analysis taught me that crypto liquidity is derivative of traditional finance. The same principle now applies to mining stocks. Their price action is no longer a derivative of Bitcoin. It is a derivative of AI demand, power contracts, and data center utilization.
Bitcoin mining is an energy-intensive process. Miners consume vast amounts of electricity to run ASICs, solving cryptographic puzzles to secure the network. Their revenue is denominated in Bitcoin. For years, the logic was simple: Bitcoin price goes up, miner revenue goes up, stock price goes up. That correlation held. But the business model has shifted.
The core insight is buried in the earnings reports.
Mining companies have pivoted from mining to hosting. They own cheap power and warehouse facilities. They are now renting that capacity to AI companies, which need massive compute for training and inference. The revenue from AI hosting is more stable and often more profitable than Bitcoin mining. Core Scientific, TeraWulf, and IREN have all reported AI-related revenue exceeding 50% of total income.
This shift is not accidental. It is a strategic response to the halving cycle and the commoditization of mining hardware. The 2024 halving cut the block reward in half, squeezing margins. The rise of AI compute demand in 2025 provided an off-ramp. Miners have taken it.
The consequence is a systematic decoupling. As AI revenue grows, the stock price becomes less sensitive to Bitcoin. The 90-day correlation data confirms this. Core Scientific at 16%, Riot Platforms at 31%, IREN at 33% — these are not outliers. They are the new norm.
The contrarian angle is that the market has not yet repriced these assets.
Investors still buy mining stocks assuming they hold crypto beta. But the underlying driver has changed. The stock is now a hybrid: part crypto infrastructure, part AI data center. The valuation model should shift from a mining multiple to an infrastructure multiple. That means lower volatility, but also lower upside in a Bitcoin rally.
This is not a new phenomenon. In 2022, during the Terra collapse, I delayed my analysis until I had confirmed on-chain evidence of the death spiral. The market was still pricing LUNA as a stablecoin, but the structural break had already occurred. The same delay is happening now. The 90-day correlation data is a lagging indicator. The break already happened. The market just hasn't reflected it yet.
For the Bitcoin bull, the message is clear: MicroStrategy is a better proxy. MSTR is a treasury company, not a mining company. It holds Bitcoin on its balance sheet and issues debt to buy more. Its correlation to Bitcoin is 78%, the highest in the study. It is a levered Bitcoin play, not a data center play.
For the Ethereum bull, the story is different. BitMine and Coinbase show high correlation to ETH — 80% and 74% respectively. But BitMine comes with a caveat: Tom Lee is its chairman. The ranking is not independent. Coinbase is a regulated exchange, but its revenue is tied to trading volume, which is cyclical and regulatory-sensitive.
The geometry of trust in a permissionless system is being redrawn.
Mining stocks were once the bridge between traditional finance and crypto. They were listed, regulated, and audited. They offered a way to gain exposure without holding digital assets. That bridge is now collapsing. The structural break means that the old assumption — that a mining stock is a Bitcoin proxy — is no longer valid.
Where code enforcement meets regulatory ambiguity, the gap between perception and reality widens. The silence before the algorithmic deleveraging is deafening. Investors who do not adjust their asset allocation will find themselves holding AI infrastructure stocks when they thought they held Bitcoin proxies.
The takeaway is not that mining stocks are bad. It is that they are different.
If your goal is Bitcoin exposure, use spot ETFs or MicroStrategy. If your goal is exposure to the AI infrastructure boom, then Core Scientific, TeraWulf, or IREN may be appropriate, but only after analyzing their power contracts, customer concentration, and capital expenditure plans.
The 90-day correlation data is a snapshot. It will change as the market reprices these stocks. The next earnings season will be the first real test. When mining companies report Q2 2026 results, the market will see the revenue split. If AI revenue dominates, the decoupling will accelerate. If Bitcoin prices surge and AI demand softens, the correlation may temporarily return. But the structural trend is clear.
Decoding the signal within the noise of volatility requires a willingness to look beyond price action. The signal here is that the business model of Bitcoin mining has fundamentally changed. The noise is the old narrative that mining stocks are crypto proxies.
I have seen this pattern before. In 2017, I audited EOS and 10x Network whitepapers, applying stochastic calculus to token emission schedules. The market was buying hype; the math said otherwise. In 2024, I analyzed the institutional inflow data after the ETF approval, predicting the altcoin bear market. The pattern repeats: the crowd is late to recognize structural breaks.
Today, the break is in the mining sector. The stocks are no longer what they seem. The market will eventually adjust. But until then, the gap between perception and reality is an opportunity — for those who understand the geometry of trust in a permissionless system.
