Tom Lee just published a ranking of 17 crypto-related stocks. The data reveals a shocking truth: the most popular mining equities share a correlation with Bitcoin as low as 16%. That is not a typo. Core Scientific, a company that once mined thousands of BTC, now moves in lockstep with AI demand, not Bitcoin. The market's collective panic should be immediate: the crypto equity proxy is broken.
Ignore the headline. Look at the latency spike. I audited the numbers myself. This isn't a temporary decoupling; it's a structural reclassification of an entire asset class.
Context: The Ranking That Backfired
Tom Lee, the well-known crypto bull and head of research at Fundstrat, recently released a 90-day rolling correlation analysis of 17 crypto-related stocks with market caps above $2 billion. The stated goal was to help investors identify the best equity proxies for Bitcoin and Ethereum exposure.
But there's a critical conflict: Lee also serves as chairman of BitMine, a company that topped his list for ETH correlation at 80%. That doesn't automatically invalidate the data, but it demands independent verification. And when I ran my own checks, the pattern held: MicroStrategy (MSTR) leads BTC correlation at 78%, followed by Coinbase (COIN) at 74% for ETH.
The real story, however, is what the ranking didn't intend to reveal: the collapse of mining stocks as crypto proxies. The market didn't crash; it woke up. The average investor still thinks Riot Platforms or Core Scientific offer Bitcoin exposure. The data says otherwise.
Core: The Business Model Drift
Let's get granular. The 90-day correlation for key mining stocks against Bitcoin:
- Core Scientific (CORZ): 16%
- TeraWulf (WULF): 19%
- Riot Platforms (RIOT): 31%
- IREN (IREN): 33%
Compare these to MicroStrategy's 78% or even to a meme stock like DJT (which had a 58% BTC correlation in the same period). The mining stocks are now less correlated to Bitcoin than a political meme.
Why? Because their revenue structure has shifted.
In 2024, Core Scientific derived over 50% of its revenue from AI computing services, not Bitcoin mining. TeraWulf's CFO recently stated that the company's future will be driven by recurring contract income from AI data centers, not volatile block rewards. MARA and CleanSpark collectively lost $851 million in 2024 while pivoting to AI infrastructure.
Based on my work as a trading signal strategist, I've seen sector rotations before. But this is different. The mining companies are not just adding a new business line; they are fundamentally reclassifying themselves. They own cheap power, industrial warehouses, and fiber-optic connections. AI companies need exactly that. So the miners are becoming landlords for compute, not producers of digital gold.
The Mechanism: From Hash Rate to Power Purchase Agreements
A traditional mining stock's value was driven by: Bitcoin price, network difficulty, hash rate, and energy costs.
A post-pivot mining stock's value is driven by: AI contract terms, data center utilization rates, electricity price arbitrage, and customer concentration (e.g., Core Scientific's deal with CoreWeave).
These are completely different risk factors. The correlation table from Tom Lee's report is a lagging indicator of this shift. I've audited the 10-K filings of six of these companies. The narrative is consistent: “We are an AI infrastructure company.”
But the market hasn't repriced them yet. The average retail investor still buys Riot or CORZ thinking it's a Bitcoin proxy. That's a dangerous mispricing.
Contrarian: The Unreported Angle
Here's what the mainstream analysis misses: this decoupling is not a bug; it's a feature.
If you believe AI demand will continue to grow at 30%+ CAGR, then these mining stocks are actually underpriced as infrastructure assets. They trade at 1x-2x book value, while pure-play data center REITs like Equinix trade at 5x-6x. The market is still pricing them as volatile miners, not as stable power landlords.
But there's a trap: the reclassification cuts both ways. If AI hype deflates, these stocks will crash harder than BTC. Why? Because they lose both the AI premium and the Bitcoin correlation. They become orphan assets.
And the conflict with Tom Lee is a signal, not a bug. BitMine's high ETH correlation might be a self-serving promotion, but it also highlights a key gap: there is no pure ETH equity proxy except Coinbase (which has regulatory risks).
My contrarian take: The smart money should be shorting the mining stocks as Bitcoin proxies and buying MicroStrategy or spot BTC. If you want AI infrastructure exposure, buy the mining stocks directly—but only if you've audited their power contracts and customer concentration. Don't confuse the two.
Takeaway: The Next Watch
The next quarterly earnings will be the catalyst. If mining companies report AI revenue exceeding 60% of total, the market will be forced to reprice them. That will create a gap for the prepared.
But for the average investor holding MARA or RIOT as a “Bitcoin play,” the moment of awakening is coming. The market's collective panic will set in when they realize their crypto exposure is actually a bet on Nvidia's data center demand.
s collective panic.
The question you should ask before buying any mining stock: Are you speculating on Bitcoin's hash rate or on the utilization rate of a GPU cluster? The answer will determine your portfolio's survival.