In June, the Bitcoin network saw a 10% reduction in mining difficulty—a textbook tailwind for miners. Yet the three U.S.-listed mining companies that reported first—CleanSpark, BitFuFu, and Canaan—all produced fewer coins than in May. The numbers are stark: CleanSpark mined 614 BTC (down 8.5%), BitFuFu 125 BTC (down 29.4%), and Canaan 64 BTC (down 28.9%).
This is not a glitch. It is a structural signal.
Difficulty dropped because some miners likely turned off inefficient machines. But the survivors failed to capture the freed-up rewards. When I read the filings, I didn’t see market noise. I saw three different failure modes—each revealing a distinct weakness in the industry’s current operating model.
Context: The Post-Halving Squeeze
Bitcoin’s fourth halving occurred in April 2024, cutting the block reward from 6.25 to 3.125 BTC. Miners who relied on older hardware (e.g., S19 series) suddenly faced negative margins at prevailing hash prices. The logical response is to upgrade or shut down. The network difficulty adjusts every two weeks to reflect the total computational power. A 10% drop means a significant amount of hash power left the network.
Normally, a difficulty drop should boost the profitability of remaining miners: they compete against fewer computers for the same block reward. June’s data suggests the opposite happened for these three public firms. Why?
Core: Code-Level Breakdown of the Three Failure Modes
I’ve spent years auditing contract-level logic and on-chain data. Here I treat each miner’s production report as a stack trace: the cause must be found in the operational layer, not the consensus layer.
1. CleanSpark: Operational Hashrate Decline CleanSpark claimed an average operational hashrate of 43 EH/s in June, down from 46 EH/s in May (a 6.5% drop). This is not a catastrophic decline, but it directly explains the 8.5% drop in BTC mined. The company attributed it to “managing our fleet during the summer heat and curtailment events.” Translation: high ambient temperatures force machines to throttle or shut down, especially air-cooled facilities. CleanSpark’s fleet is relatively modern (mostly S21s and M60s), but even efficient hardware cannot run at full capacity in 40°C conditions without adequate cooling infrastructure. This is a predictable, recurring issue, not a one-off.
2. BitFuFu: Collapsing Hosted Hashrate BitFuFu’s total hashrate dropped from 19.5 EH/s to 15 EH/s, a 23% decline. The company explicitly stated this was “due to a reduction in hosted hashrate.” Meanwhile, its self-mined hashrate actually increased to 3.5 EH/s. This tells me BitFuFu is a middleman: it leases hashpower from third-party hosting providers. Those providers likely raised prices or left the network when margins compressed. The company’s own hash rate grew, but not enough to offset the loss of rented power. This exposes a fundamental fragility in the “light asset” model: if you don’t own the machines and the power contracts, your hash rate is a variable, not a constant.
3. Canaan: Grid Maintenance as a Proxy for Infrastructure Weakness Canaan produced only 64 BTC in June, down from 90 BTC in May. The company cited “grid maintenance at some mining sites.” This sounds like a one-off, but it signals a deeper problem: Canaan operates mine sites where the power utility is unreliable. As a hardware manufacturer that also self-mines, Canaan should have the incentive to secure rock-solid power agreements. The fact that it didn’t indicates either poor execution or a reliance on cheap, unstable power sources. “Ghost in the audit: finding what wasn’t there,” as I often say. The grid maintenance is a symptom of a larger operational gap.
Contrarian: The Difficulty Drop Was a Trap, Not a Gift
Conventional wisdom says: lower difficulty = easier to find blocks = more BTC per hash. That’s true in a static model. But in practice, the difficulty drop is a lagging indicator of distress. The 10% drop likely came from two sources: old generation machines (S9, S17) that became unprofitable above $50k BTC, and overleveraged miners who couldn’t fund operations after the halving. The hash rate that left wasn’t idle—it was the marginal, insecure capacity. What remained was the core of professional miners who could survive, but they faced a different problem: competition from new, high-efficiency machines deployed by well-capitalized players (e.g., Riot, Marathon). In June, the network difficulty may have dropped, but the composition of the hash rate shifted towards more efficient, lower-cost hardware. This means the “effective” cost per hash didn’t fall proportionally. Miners with older fleets or higher electricity costs still struggled.
The real insight is that the difficulty drop didn’t help these three because their own operating issues—heat, hosting dependence, grid reliability—are orthogonal to network difficulty. They are internal, not external. “Trust is math, not magic: stripping away the myth,” that difficulty is the only variable. It’s not.
Takeaway: The Next Danger Is Not BTC Price, It’s Capital Structure
Reading between the lines: CleanSpark’s production drop is manageable, but BitFuFu’s dependency on hosted hash rate is a ticking time bomb. If Bitcoin stays below $60k for another quarter, hosting providers will continue to churn, and BitFuFu may lose half its effective capacity. Canaan’s grid issues could appear in other sites, especially as summer heat peaks in July-August. The common thread is that all three companies will likely need to raise capital—either through debt or equity—to upgrade infrastructure and secure long-term power agreements. This dilutes existing shareholders. The market has already priced in some of this: CLSK, FUFU, and CAN are down 20-40% from their May highs. But I suspect the worst is not yet priced. “Silence speaks louder than the proof,” as the saying goes. The real signal will come in the Q2 earnings calls when management reveals their all-in production cost per BTC. If that number is above $40k, they are living on borrowed time.
For now, the data is clear: reducing difficulty is not a cure-all. The miners who survive the post-halving winter will be those who own their machines and their power—and have the balance sheet to weather the storm. The rest will become statistics.