BTC mining dynamic

Bitcoin's security model is elegantly simple: miners expend real-world resources, energy and compute, to secure the network, and the protocol compensates them through block rewards and transaction fees.

Difficulty adjusts every two weeks, recalibrating the computational bar to maintain a steady block rate regardless of how many miners are competing. While this looks like quick recalibration, longer capex cycles make it far less reflexive in practice.

Different than before

While miners have weathered difficult periods in 2018 and 2022, this cycle is different. Rather than comparing hashrate and mining difficulty, we make this argument from first principles. Revenue (block rewards + fees) and input costs (energy + compute) show that margin compression was cyclical in previous epochs. Today it feels structural.

Hashrate and difficulty have scaled to a point where the protocol's self-correcting mechanism can no longer offset the economic headwinds. We're at the structural ceiling, not a cyclical trough.

As the exhibit below shows, both sides of the equation are compressing in the first half of the epoch, which is usually when we see expansion:

• Revenue: For the first time across any epoch, BTC has yet to deliver a 2x return on a 4-year rolling basis, sitting at just 1.15x, the minimum needed to offset the halving's 50% subsidy cut and keep block reward revenues intact.
• Margins: This epoch's gross margins peaked at ~30%, a level that marked the floor of prior bear market cycles.

Exhibit 1 - Revenue and gross margin of miners across epoch 3-5

As a result, we are seeing miners expanding and pivoting into high-performance computing for AI applications. Several factors are driving this shift:

• Unit Economics: HPC delivers meaningfully higher yield per GW/h relative to BTC mining, improving margins at a time when they are under structural pressure
• Market Rerating: Miners making the pivot are being rewarded with multiple reratings in public markets, unlocking access to cheaper capital through equity issuances and convertible debt at more favorable terms
• Optionality: HPC infrastructure offers greater operational flexibility than ASIC-dependent mining setups, giving operators the ability to adapt as market conditions evolve

BTC mining is a structurally rigid business model, and the current margin environment is forcing players to identify levers they can pull. While the AI pivot is a compelling one, it is also a drastic and capital-intensive step.

We believe active balance sheet management represents another lever one that is often underexplored that miners can and should be pulling in parallel.

Revenue: The 2x problem

Block rewards halve every four years by design. If BTC doesn't at minimum double over that same window, miner revenue is in structural decline. Fees are the only other offset. In every prior epoch, price delivered multiples that made this a non-issue. Epoch 3 returned 20x+, Epoch 4 north of 10x. Epoch 5 is sitting at ~1.15x. Fees haven't structurally kicked in, and the revenue squeeze follows directly.

The break is structural, not cyclical. Bitcoin is maturing. Volatility has compressed as institutional participation has deepened and BTC increasingly trades as a macro risk asset. The ETF approvals and corporate treasury adoption that the industry celebrated as validation are, by definition, the forces that dampen explosive return cycles. A more liquid, more institutionally-held asset does not produce 20x four-year returns. For miners whose business model was implicitly underwritten by hyperbolic price appreciation, that is a regime change, not a bad quarter.

Exhibit 2 - Network revenue and 4-year rolling BTC return across epochs

The top panel shows network revenue tracking below prior epochs at the equivalent stage of the cycle. The bottom panel makes the return compression explicit: Epochs 3 and 4 were registering 5-20x at this point in their respective cycles. Epoch 5 sits just above 1x, the 2x breakeven uncrossed.

The natural question is whether fees fill the gap. The narrative is intuitive: as block subsidies compress toward zero, rising network usage should organically grow fee revenue to replace them. The data tells a different story.

Exhibit 3 - Daily BTC miner transaction fees and fee share of total miner revenue

Fee revenue is episodic, not structural. The spikes, driven by Ordinals activity and periodic mempool congestion, are visible but transient.

Outside those windows, fees contribute low single digits as a percentage of total miner revenue. A business cannot be underwritten on recurring congestion.

Exhibit 4 - Gross margin per epoch with and without transaction fee revenue

The dashed lines, representing margin inclusive of fees, barely separate from the solid lines across all three epochs. In Epoch 5, where block reward margins are already thin, fees are not widening that gap in any meaningful way. The protocol's second revenue mechanism is not functioning as a structural backstop.

Margins

The cost structure of Bitcoin mining is uniquely simple. Unlike any other commodity, input costs reduce to just two variables: energy and compute. No raw materials, no logistics, no labour intensity at scale. That simplicity is also the trap. With so few levers to pull, when revenue compresses, margin compresses with it.

Exhibit 5 - Network revenue and energy cost (top) and gross margin (bottom) across epochs

The wedge has been narrowing. The bottom panel translates this into gross margin percentage, and the long-run compression is unambiguous. Peak margins in each successive epoch are lower than the prior, and bear market troughs cut deeper. Epoch 5 has peaked at levels that in prior cycles represented the trough, not the ceiling.

Equally telling is what is absent. Prior epochs exhibited clear mid-cycle margin recovery, driven by price appreciation in the years following the halving. Epoch 5 is showing no such seasonality. The recovery arc that gave miners breathing room to reinvest, refinance, and extended runway has not materialized.

Exhibit 6 - Gross margins per epoch overlaid

The epoch overlay makes the compression impossible to dismiss. Epoch 3 miners operated through extended periods of 70-80% gross margins. Epoch 4, even through its bear market trough, recovered meaningfully before the next halving. Epoch 5's red line is tracking structurally below prior cycles from day one.

The Levers

Repurposing

The AI infrastructure boom has created an entirely new monetization path for an asset class the market had written off as structurally challenged. Hyperscalers and AI compute operators are in a race to secure power at scale, and they need it faster than traditional data center development timelines allow. Grid interconnection queues in the US stretch five to ten years in many markets. Permitting, construction, and community opposition add further friction.

Bitcoin miners spent years building large-scale power infrastructure in low-cost energy markets. They now find themselves sitting on exactly what the AI industry needs most urgently and cannot easily replicate. The logical response is a pivot toward HPC hosting, and for miners making that move, the market re-rating is immediate and significant.

Exhibit 7 - Equity value creation per watt from recent BTC-to-DC conversion transactions

Sites that traded at $1-7 per watt under Bitcoin mining economics are clearing at multiples of that once reoriented toward contracted AI compute. Value creation per watt has nearly tripled since mid-2025, reaching ~$18/watt by December, anchored by the HUT/Google/Anthropic deal. Not a mining company anymore. An infrastructure company.

As a result, we are seeing miners expanding and pivoting into high-performance computing for AI applications. Several factors are driving this shift:

• Unit Economics: HPC delivers meaningfully higher yield per GW/h relative to BTC mining, improving margins at a time when they are under structural pressure
• Market Rerating: Miners making the pivot are being rewarded with multiple reratings in public markets, unlocking access to cheaper capital through equity issuances and convertible debt at more favorable terms
• Optionality: HPC infrastructure offers greater operational flexibility than ASIC-dependent mining setups, giving operators the ability to adapt as market conditions evolve

BTC mining is a structurally rigid business model, and the current margin environment is forcing players to identify levers they can pull. While the AI pivot is a compelling one, it is also a drastic and capital-intensive step. We believe active balance sheet management is the most underutilized lever available to miners and one that deserves far greater strategic attention.

The Case For Active Balance Sheet Management

Bitcoin miners are collectively carrying close to 1% of the total BTC supply, a legacy of the HODL era. While the industry is shifting toward active divestiture, the full toolkit of treasury management remains largely untapped.

Yield generation is an ethos long embraced in DeFi, but one Bitcoin has been slower to adopt. Given the scale of BTC sitting idle on miner balance sheets, this is a structural inefficiency the sector can no longer afford to ignore.

Many miners today carry elevated leverage with meaningful near-term debt obligations, particularly through convertible notes. The sensitivity table below maps the yield that needs to be generated on held BTC, across spot price and treasury allocation scenarios, to fully cover those obligations.

Exhibit 8 - Required yield given BTC price and % of treasury allocated to cover debt service.

Debt service is only one component of total liquidity needs, alongside opex, capex, and working capital requirements. But this exercise illustrates a broader point: efficient BTC treasury management can meaningfully offset liquidity pressure, reducing operational fragility and improving the resilience of the business through the cycle.

How to generate yield

Within the miners we face, approaches broadly fall into two categories:

• Active management: monetizing market risk through derivatives structures. Covered calls, cash secured puts, and more sophisticated overlays. Requires ongoing management, market access, and derivatives expertise.
• Passive management: monetizing credit risk through decentralized lending. Deploy BTC into a lending protocol, earn interest, and know your terms upfront. Set and forget relative to the options approach. One example is the recently launched WBTC market on Wildcat.

Depending on their capabilities and risk appetite, miners should look to combine active and passive treasury management to optimize yield generation. In an industry that feels increasingly like a race to the bottom, any marginal edge meaningfully improves survival odds.

Conclusion

Epoch 5 is the first cycle where the protocol's self-correcting mechanisms are not enough. Difficulty adjusts, but it cannot fix a fee market that hasn't structurally arrived, a BTC price that has failed to deliver the multiples miners implicitly underwrote their businesses against, or an energy cost curve that keeps compressing margins from below.

The AI pivot is real, and the market is pricing it aggressively. But it is a solution available to a minority, those with the right site quality, balance sheet, and operational bandwidth to execute a fundamental repositioning. For the rest, survival depends on optimizing what they already have.

That means active treasury management. Miners are sitting on close to 1% of the total BTC supply. In prior epochs, HODLing was rational. In this one, letting that treasury sit idle is a cost. Covered calls, cash secured puts, on-chain lending, the tools exist. The miners who treat their BTC holdings as a working asset rather than a passive reserve will carry a structural edge into the next halving.

The industry is being forced to grow up. The ones who do it fastest will still be here for Epoch 6.