Over the past 90 days, Core Scientific’s correlation to Bitcoin dropped to 16%. That’s not a typo. It’s a data point that obliterates the long-held assumption that buying a Bitcoin miner gives you leveraged exposure to the underlying asset. Tom Lee’s recent ranking of 17 crypto-related stocks was supposed to be a guide for traditional investors seeking crypto exposure. Instead, it became an autopsy of a broken strategy.
Lee’s report, covered by major outlets, sorted stocks by their 90-day rolling correlation to Bitcoin and Ethereum. The top of the list is predictable: MicroStrategy at 78% for BTC, BitMine at 80% for ETH. But the bottom of the list tells a different story. Core Scientific, Riot Platforms, and IREN all sit below 35% correlation to Bitcoin. These are the names that most retail investors still buy as ‘crypto stocks.’ The gap between perception and reality is a chasm.
Context: Why This Matters Now
This isn’t just a statistical anomaly. It’s a structural shift in the business models of publicly traded miners. Over the past two years, the narrative has quietly pivoted. Miners are no longer pure-play Bitcoin producers. They are becoming AI infrastructure landlords. The shift started when energy-rich, facility-heavy miners realized they could earn more stable revenue by leasing compute power to AI companies than by gambling on hashprice. The data confirms it: Core Scientific’s AI revenue now dominates its income statement. TeraWulf’s CFO explicitly stated that future revenue will be driven by recurring contracts, not Bitcoin block rewards. IREN is building data centers designed for AI, not ASICs.
Strategic pivots aren’t always transparent to the market. The market still prices these stocks as if they are crypto proxies. But the correlation numbers from Lee’s own ranking show the truth: the linkage is decaying. The 90-day rolling correlation for the average miner is now closer to a tech data center than to Bitcoin.
Core: The Data Tells a Story of Asset Reclassification
Let’s break down the numbers. MicroStrategy remains the gold standard for BTC equity exposure. Its 78% correlation is high, but that’s because the company is essentially a Bitcoin treasury with a leveraged balance sheet. You don’t buy MicroStrategy for its business; you buy it for its BTC holdings. That’s a clean proxy.
Coinbase sits at 74% for ETH. Its revenue is tied to trading volume, which is correlated with crypto activity. That’s a valid proxy for the broader ecosystem, but it carries regulatory and competitive risk.
Now look at the miners: Core Scientific at 16% BTC correlation. Riot at 31%. IREN at 33%. These numbers are not just low; they are lower than the correlation of DJT (the Trump Media stock) to Bitcoin. That’s right. The stock of a social media company is more correlated to Bitcoin than the stocks of companies that mine it.
This is not a bug. It’s a feature of the business model shift. When a miner sells 40% of its revenue to AI companies, its stock price becomes driven by AI demand, power contracts, and data center utilization rates. The Bitcoin price becomes a secondary factor. In my experience analyzing the 2020 Compound liquidity crisis, I saw a similar pattern: financial engineering can disconnect a token from its underlying asset. Here, the disconnect is driven by real economic activity. Liquidity doesn’t flow to where it’s expected; it flows to where the contracts are signed.
The ranking also reveals a hidden conflict of interest. Tom Lee is the chairman of BitMine, which sits at the top of the ETH correlation list. While the data may be accurate, the independence of the ranking is compromised. Any investor using this list as a buy signal must audit the source. I’ve seen this pattern before in the 2021 Yuga Labs pivot: a narrative that serves the promoter’s own portfolio. The market is efficient, but it’s also manipulated by incentives.
Contrarian: The Unreported Blind Spot
The conventional wisdom is that mining stocks are a good way to get crypto exposure without holding the asset directly. That wisdom is now dangerous. The contrarian truth is that the mining sector has undergone a silent asset reclassification. These stocks are no longer crypto beta; they are AI infrastructure beta.
Investors who bought miners as a Bitcoin proxy are now exposed to a completely different set of risks: energy price volatility, AI demand cycles, and the execution risk of capital-intensive data center builds. The MARA and CleanSpark example is instructive: they have collectively lost $851 million in their AI pivot. The transition is expensive, and not all will succeed.
Moreover, if the AI narrative cools, miners could face a double hit: losing the AI premium while their Bitcoin correlation remains low. They would be value traps in both narratives. The 90-day correlation window also masks this risk. Over a longer horizon, the correlation may re-emerge if Bitcoin price surges, but the structural shift in revenue composition means the link is permanently weakened.
Takeaway: The Next Watch
You don’t buy an AI data center when you want Bitcoin. If your goal is crypto exposure, the instruments are clear: spot Bitcoin, ETFs, or MicroStrategy. If you buy a miner, you are making a bet on AI infrastructure, not on the next halving. The market will eventually reclassify these stocks. The question is whether that reclassification happens before the next bear cycle forces a margin call. Will the market adjust its pricing models in time, or will investors continue to buy a story that no longer fits the data? The answer will determine who gets caught holding the wrong bag.
Signatures: - Liquidity doesn’t lie, but narratives do. - Strategic pivots aren’t always transparent to the market. - You don’t buy an AI data center when you want Bitcoin.