The correlation data is not the story. The story is what's driving the correlation.
Tom Lee's recent ranking of 17 crypto-related stocks was supposed to be a simple guide for investors seeking Bitcoin and Ethereum exposure through equities. The methodology was straightforward: calculate 90-day rolling correlations with BTC and ETH. The results, however, revealed a structural rupture that most market participants have not yet internalized.
MicroStrategy (MSTR) sits at the top with a 78% correlation to Bitcoin. Coinbase (COIN) follows with 74% to Ethereum. But then the numbers get strange. Core Scientific (CORZ) — a company that runs one of the largest Bitcoin mining fleets in North America — shows a mere 16% correlation with BTC. Riot Platforms (RIOT) manages 31%. IREN, once a pure-play miner, checks in at 33%. Meanwhile, BitMine, a company chaired by Tom Lee himself, claims 80% correlation with ETH.
The market is living in a narrative lag. The data is already in the future.
Context: The Old Playbook Is Broken
For years, the logic was simple: if you wanted Bitcoin exposure without holding the asset directly, you bought mining stocks. These companies earned revenue in BTC, so their stock prices tracked the coin. The model worked because miners were essentially BTC production factories. Their income was a direct function of Bitcoin price times hash rate minus electricity costs.
That model is now decomposing. The catalyst is not a market crash or a regulatory crackdown. It's a business model pivot that has quietly reclassified mining companies from "crypto miners" to "AI infrastructure providers." The shift began in 2022 when the bear market squeezed margins and the Terra-Luna collapse wiped out many overleveraged players. In response, surviving miners looked for alternative revenue streams. They found AI computing.
Mining companies possess two scarce assets: cheap power and large-scale data centers. AI companies need both. The result is a wave of contract agreements where miners lease out their facilities to AI firms for GPU-based computing. Core Scientific, for example, has signed multi-year deals with AI companies, and its CEO now explicitly states that AI hosting is the primary growth driver. TeraWulf's CFO recently noted that the business is becoming "more recurring contract-driven." IREN is building out its data center capacity for AI workloads.
The data from Q1 2025 earnings confirms this. For Core Scientific, AI-related revenue already accounts for over 60% of total income. TeraWulf is approaching 50%. Even Riot Platforms, which has been slower to pivot, is exploring AI partnerships. The result is a fundamental shift in what drives these companies' stock prices.
Core: The Structural Decoupling
Let's be precise about what is happening. The 90-day rolling correlation between a mining stock and Bitcoin is not falling because of noise or market inefficiency. It is falling because the underlying business has changed. The correlation is a symptom, not a cause.
Consider the causal chain. In the old model: BTC price increases → mining revenue increases (because BTC-denominated income is worth more in USD) → miner profitability rises → stock price rises. The chain is direct and fast.
In the new model: BTC price increases → mining revenue increases (still true) BUT AI hosting revenue is independent of BTC price → total revenue mix dilutes BTC sensitivity → stock price reacts more to AI demand signals, power contract renewals, and data center utilization rates than to BTC price movement.
This is not a temporary anomaly. It is a permanent reclassification of an asset class. The market is slowly pricing mining stocks as hybrid assets: part crypto beta, part AI infrastructure beta. The problem is that most investors still think they are buying pure crypto exposure.
Based on my experience auditing cross-border payment systems, I have seen this pattern before. When a technology company shifts from a pure payment processing model to a diversified software-as-a-service model, its correlation with the underlying transaction volume declines. The same principle applies here. The revenue stream changes, and the correlation follows.
Contrarian: The Double-Edged Sword of AI Pivot
The conventional narrative is that the AI pivot is a smart move that will stabilize revenue and reduce volatility. That is partially true. AI contracts are longer-term and less volatile than Bitcoin mining margins. But the contrarian perspective is more nuanced.
First, the pivot is expensive. MARA and CleanSpark, two companies that aggressively pursued AI, have posted combined losses of $851 million. The capital expenditure required to build out GPU clusters is enormous. Companies are taking on debt and diluting shareholders to fund the transition. The risk is that the AI demand cycle peaks before they achieve positive returns.
Second, the pivot creates a new set of dependencies. Mining stocks now depend on the health of the AI industry, which is itself subject to hype cycles, regulatory scrutiny (especially around training data and energy consumption), and competitive pressure from hyperscalers like AWS and Google Cloud. If AI demand softens, mining stocks could lose both the AI premium and the BTC linkage, creating a double-whammy.
Third, the correlation data itself is backward-looking. The 90-day window captures recent market behavior, but it might not predict future dynamics. If Bitcoin enters a sustained bull run, mining stocks could temporarily regain correlation as the BTC component of their revenue becomes dominant again. But the structural trend is clear: as AI revenue share grows, the correlation will continue to erode.
You're not buying Bitcoin. You're buying a data center lease in disguise.
Takeaway: What This Means for Your Portfolio
If your goal is pure Bitcoin exposure, the most efficient vehicles remain Bitcoin spot ETFs, MicroStrategy, or direct BTC holdings. MicroStrategy's 78% correlation is the highest among public equities, but even that is not perfect. The company's leverage, financing costs, and market sentiment introduce additional variables.
If your goal is Ethereum exposure, Coinbase and BitMine show higher correlations, but both come with caveats. Coinbase's revenue is tied to trading volumes, which are volatile and subject to regulation. BitMine's chairman is the same person who published the ranking — a conflict of interest that demands independent verification.
If your goal is AI infrastructure exposure, some mining stocks may offer attractive entry points, but you need to evaluate them as data center operators, not as crypto proxies. Look at their power contracts, customer concentration, capital expenditure plans, and free cash flow. The old metrics — hash rate, BTC per share, mining difficulty — are no longer sufficient.
This is not a correlation breakdown. It's a business model revolution. The market is slow to reprice, but it will. The question is whether you are positioned for the new reality or still operating under the old assumptions.
In crypto, the biggest risk is not volatility — it's misclassification.
