The Great Decoupling: Why Bitcoin Mining Stocks Are No Longer Your Crypto Proxy

The myth that buying a mining stock is a reliable way to bet on Bitcoin has been a cornerstone of the “crypto equity” thesis for years. The logic seemed flawless: miners produce Bitcoin, their revenue is tied to the Bitcoin price, and their stock should move in lockstep. But as I sit here in Mumbai, reviewing the latest correlation data from Tom Lee’s ranking of 17 crypto-related stocks, I see a different story. Over the past 90 days, the correlation between Bitcoin and the largest mining companies has dropped to as low as 16% for Core Scientific and 31% for Riot Platforms. Meanwhile, MicroStrategy, a company that doesn’t mine at all, boasts a 78% correlation. This is not a temporary blip. It is a structural decoupling, driven by a fundamental shift in the business models of mining firms. They are no longer pure Bitcoin plays; they are becoming AI infrastructure landlords. And for investors who thought they were buying crypto exposure, this is a wake-up call that demands a reevaluation of the entire “crypto equity” category. To understand why this decoupling is happening, we need to look beyond the price charts and into the balance sheets of these companies. The traditional mining model was simple: you buy ASICs, you pay for electricity, you mine Bitcoin, and you sell it to cover costs. The stock price was a leveraged bet on Bitcoin’s price—if Bitcoin doubled, the stock often tripled. But that model is under pressure. The 2022 bear market, the collapse of Terra, and the rise of AI have changed the incentives. In 2023, I watched as Core Scientific, a company that went through Chapter 11 bankruptcy, re-emerged with a new strategy: instead of dedicating all its power to Bitcoin mining, it started renting out its data center capacity to AI companies. This pivot is not unique. During my 2020 DeFi Trust Bridge work, I learned that the most resilient communities are those that adapt their infrastructure to serve the most pressing needs. The miners are doing the same. They have cheap power, physical facilities, and cooling systems—assets that AI companies desperately need. As a result, companies like TeraWulf and IREN are now generating a significant portion of their revenue from AI compute services, not from mining. TeraWulf’s CFO recently stated that their business will increasingly be driven by recurring contract income, not by the volatile price of Bitcoin. This shift is visible in the numbers: the higher the share of AI revenue, the lower the correlation with Bitcoin. Core Scientific, with a large AI book, has a Bitcoin correlation of just 16%. IREN, which has the lowest AI share among the miners analyzed, has a correlation of 33%, still far below MicroStrategy’s 78%. The core insight here is that the market is in the process of reclassifying mining stocks. They are moving from the “crypto beta” bucket to the “AI infrastructure” bucket. This is not a gradual drift; it is a tectonic shift in how these companies create value. From code audits to community heartbeats, I have always argued that understanding the underlying incentives is more important than the surface-level data. The incentive for mining CEOs is now clear: AI contracts offer more stable, higher-margin revenue than Bitcoin mining. In the current market, renting out a GPU to an AI startup can generate three times the profit per kilowatt-hour compared to hashing Bitcoin. This economic reality is driving a strategic pivot that will only accelerate. The 90-day correlation data is a snapshot of this transformation. But we must be careful: the correlation is backward-looking. It tells us what happened, not what will happen. If Bitcoin prices surge again, the correlation could temporarily rise as miners who still have some hash power benefit. But the structural trend is clear: the mining industry is becoming a hybrid sector, part crypto, part AI. For investors, this means that the old heuristic of “buy miners to get leveraged Bitcoin exposure” is no longer valid. Building bridges where DeFi once built walls, we need to build new investment frameworks that recognize this hybrid nature. Now, let me introduce the contrarian angle—the blind spots that most analyses miss. The first blind spot is the assumption that the AI pivot is a positive development for all miners. The data shows that the transition is expensive. MARA and CleanSpark, two companies that have aggressively moved into AI, have collectively lost $851 million in the process. The capital expenditure required to retrofit data centers, acquire GPUs, and secure long-term power contracts is massive. Some of these companies are carrying high debt loads, and the AI revenue is not yet enough to cover the costs. This is a classic technology adoption curve: the pioneers often get the arrows, not the profits. The second blind spot is the conflict of interest in the very ranking that brought this decoupling to light. Tom Lee, the analyst who published the correlation ranking, is the chairman of BitMine, a company that ranked first in ETH correlation. BitMine’s ETH correlation is 80%, but with Lee’s dual role, investors must question whether the ranking is designed to promote his own company. Trust is not a protocol, it is a practice, and I have seen too many conflicts erode the credibility of otherwise useful data. The third blind spot is the assumption that correlation is stable. The data is a 90-day rolling calculation, which means it can change rapidly as market conditions shift. In a strong Bitcoin bull run, miners might temporarily corrrelate more, creating a false sense of security for those who think they are buying a crypto proxy. But the underlying business model shift is permanent. The mining industry is structurally different from what it was two years ago. So what is the takeaway? For investors seeking pure Bitcoin exposure, the most effective vehicles remain Bitcoin spot ETFs, direct holdings, or MicroStrategy. MicroStrategy’s model is simple: it holds Bitcoin on its balance sheet and uses debt to amplify returns. It does not have the operational complexity of a mining company that is also trying to be an AI data center. For those who want to bet on the AI infrastructure theme, mining stocks can be a compelling play, but only if you understand the risks. The key signals to watch are the share of AI revenue in quarterly reports, the quality of long-term contracts, and the free cash flow. If a miner’s AI revenue exceeds 50% of total revenue, it is no longer a Bitcoin proxy. It is an AI infrastructure company. The market is still pricing some miners as if they are crypto plays, creating opportunities for mispricing. But the biggest risk is asset misallocation—thinking you are long Bitcoin when you are actually long AI. As I often say, the audit was just the beginning of the bond. The real work is in understanding the evolving nature of the assets we hold. In this sideways market, where chop is the dominant theme, positioning is everything. Do not let the old labels fool you. The mining stocks are rewriting their stories, and it is up to us to read the new chapters.