The Weekly Hash, 1/25/21: Despite falling Hash Spreads™ and rising difficulty, mining economics remain robust
Bitcoin dropped 10.7% in a volatile week of trading to end at $32,216 at midnight UTC on 1/24. Meanwhile, Difficulty increased 1.05% on 1/23, with a slowdown in blocks mined toward the end of the Difficulty epoch. This resulted in a target Hashrate of 149EH/s, a new high, but slightly below our projected 150EH/s+ estimate last week. Total BTC earnings per PH/s are now ~6.45mBTC / PH/s, down from 6.58mBTC last week. (1mBTC or milliBTC = 1/1000 BTC.)
Transaction Fees (Tx fees) fell 89 bps week-on-week (WoW) to 6.3% of miner rewards. We have observed a strong negative 69% correlation between daily blocks mined and Tx fees, as rising block count leads to less network congestion and shorter confirm times.
Bitcoin mining revenue fell to $208 / PH/s per day and $227/MWh, as price declines combined with higher difficulty and declining Tx Fees.
The BitOoda North American Hash Spread™ eased off 14.2% WoW to $195, having peaked at $311 on 1/9.
We define the BitOoda Hash Spread™ as the difference between the cost of power per MWh and the Bitcoin mining revenue per MWh. This gives miners a quick sense of the surplus generated by their business to cover personnel, overhead, depreciation, and profit. EIA data shows a weighted average peak/off-peak U.S. wholesale industrial power price of $31.76 / MWh, leading to an aggregate spread of $195 across 8 power markets, which is 14.2% lower than last week.
Older generation S9-class devices, which are much more sensitive to price fluctuations, saw their hash spread fall 22% to $32 / MWh. S17-class devices, the bulk of the current installed base, see a hash spread of about $134 / MWh.
Spreads remain robust and mining investments remain a compelling opportunity. We examine a hypothetical scenario of outfitting an existing building with power infrastructure and S19-class rigs for a 25MW project. Under our assumptions, it would take 180 days to the mine the first BTC with a $32mm investment. By day 398 (218 days of mining), the rig investment would be recovered, with an IRR of ~80% and a 524-day payback period including the long-lived power infrastructure investment.
We assume prices are static at $36,000 for the project duration, while difficulty could rise 6–8x before daily revenue dips below operating expenses. If price appreciates along with Difficulty, returns would be healthier still, with shorter payback periods.
At current Difficulty levels, it takes $4508 in power expense to mine 1 BTC using S19 rigs, and $16,031 using S9 rigs at the national average power price. Thus, excluding labor (appx $1000 per BTC on S19s), mining margins range from 50% to 85%.
Any power purchase agreements that lower the price, or that deliver incentives such as rebates for curtailment as a demand response load, can further reduce the cost of power. The below chart shows the impact of power price on the cost to mine 1BTC.
The marginal cost of 1BTC is about $4500, with about 50% margins even with old, depreciated S9 machines.
Long-term price growth is supported by news flow surrounding accelerating fund allocations to the asset class. While investors need to navigate price and difficulty risk that can pose existential threats to unhedged mining operations, we are optimistic about the long-term outlook for well-capitalized, appropriately-hedged businesses.
Email us at firstname.lastname@example.org to discuss how we can help manage your risk or gain exposure to the space.