The Great Decoupling: Why Bitcoin Mining Stocks No Longer Track BTC

Core Scientific’s 90-day BTC correlation just hit 16%. That’s not a rounding error. It’s a structural signal. In the same study, MicroStrategy sits at 78% and Coinbase at 74%, while most mining stocks barely register above 30%. The narrative that buying a mining stock gives you leveraged Bitcoin exposure is dead. The market just hasn’t priced in the obituary. Tom Lee, a well-known crypto bull, recently ranked 17 crypto-related stocks by their 90-day rolling correlation to Bitcoin and Ethereum. The study was intended to help investors choose the best equity proxy for crypto exposure. Instead, it revealed a hard truth: the proxy is broken. The data set covers companies with market caps above $2 billion, including miners, exchanges, and treasury firms. The results are clean, but the implications are messy. Here’s the context. The crypto equity universe has historically been divided into three buckets: pure-play treasury companies (MicroStrategy), exchanges (Coinbase), and miners (Riot, Mara, Core Scientific, etc.). Investors used these stocks as beta plays on Bitcoin. If Bitcoin rallied, miners were supposed to rally harder due to operational leverage. That logic held during the 2020-2021 bull run. It does not hold today. Let’s examine the core data. The correlation table shows a clear hierarchy. MicroStrategy leads with 78% BTC correlation. BitMine (a tiny ETH miner) tops the ETH list at 80%. Coinbase follows at 74% ETH. Then the drop-off. Core Scientific: 16% BTC. Riot: 31%. IREN: 33%. TeraWulf: 31%. Even the DJT (Trump Media) correlation is higher than most miners. That’s not a joke. The data set shows that mining stocks now have lower BTC correlation than a meme stock. Why? The answer lies in the income statement. Mining companies are pivoting hard to AI compute. Core Scientific now generates over 50% of its revenue from AI hosting. TeraWulf’s CFO explicitly stated that future revenue will be driven by recurring contracts, not Bitcoin mining. IREN is building data centers for AI workloads. The business model is shifting from “extract Bitcoin” to “rent compute.” That changes the valuation driver from BTC price to AI demand, power contracts, and data center utilization. s immutable logic. If a company’s revenue is tied to AI compute, its stock price will correlate with AI infrastructure stocks, not Bitcoin. The data confirms this. The 90-day correlation between mining stocks and BTC is inversely proportional to the share of AI revenue. Core Scientific, with the highest AI revenue share, has the lowest BTC correlation. Mara and CleanSpark, which are still primarily mining, have higher correlation but are also bleeding cash from AI experiments. The trend is clear: the more a miner pivots to AI, the less it behaves like a crypto asset. This is where the contrarian angle bites. The market still treats mining stocks as Bitcoin proxies. Retail investors buy Riot or Mara thinking they are getting leveraged BTC exposure. They are not. They are buying a hybrid asset that is part crypto, part AI infrastructure, and part power company. The real beta is no longer to Bitcoin. It’s to hyperscaler capex and electricity prices. The conflict of interest in Tom Lee’s study only amplifies the confusion. Lee is the chairman of BitMine, which ranks first in ETH correlation. That does not invalidate the data, but it demands skepticism. The ranking may be accurate, but the source has a clear incentive to promote BitMine as the best ETH proxy. Investors should treat that conclusion as a hypothesis, not a fact. From my experience auditing smart contracts and trading through the 2020 DeFi summer, I learned that the most dangerous assumption is that past correlations will hold. The Terra collapse taught me that systemic risk is always predictable through code analysis. The same principle applies here. The code is the business model. If the business model changes, the correlation changes. The data is not lying. The market is just slow to update its mental models. What does this mean for the average investor? First, if you want Bitcoin exposure, buy Bitcoin. Or buy MicroStrategy, which is essentially a Bitcoin treasury with a software wrapper. MSTR’s correlation is high because its asset base is 90% Bitcoin. Second, if you are holding mining stocks thinking they are crypto plays, you are misallocated. You are now exposed to AI demand, power contracts, and execution risk. Third, be wary of rankings that come from conflicted sources. Cross-check the data with independent analysis. The takeaway is actionable. The next time you see a mining stock rally, ask yourself: is it because Bitcoin pumped, or because a hyperscaler signed a new hosting deal? The answer will tell you which asset class you are actually trading. The era of the mindless mining stock as a Bitcoin proxy is over. The market is repricing these assets into a new category. Whether that repricing leads to higher multiples or lower ones depends on AI’s trajectory, not Bitcoin’s. Forward-looking thought: If AI demand continues to grow, mining stocks will decouple further. They will trade like data center REITs, not crypto equities. If AI demand stalls, they will lose both the AI premium and the crypto beta, creating a double hit. The smart money is already adjusting. The question is whether retail will catch up before the correction.