11 hours ago
Bitcoin vs Gold Mining: The Future of Scarcity is Digital
James Butterfill — Head of Research
Max Shannon — Research Analyst
Matthew Kimmell — Research Analyst
The Bitcoin mining industry has largely shrugged off halving-related concerns, with network hashrate accelerating sharply in Q4 2024. This surge was driven by a combination of favourable political developments and a strong price rally, propelling the hashrate to a record high of 900 Eh/s. We now expect the network to reach the symbolic 1 Zettahash/s milestone by July 2025.
At the same time, valuation multiples across the sector have compressed, reflecting a growing investor consensus that Bitcoin mining increasingly resembles a zero-sum game — where gains in one miner’s capacity directly erode the market share of others. In response, several miners are strategically diversifying into data centre infrastructure and high-performance computing (HPC) hosting, seeking to capture more stable and scalable revenue streams beyond Bitcoin.
The weighted average cash cost to produce one bitcoin among publicly listed miners increased to approximately US$82,162 in Q4 2024, up from US$55,950 in Q3 — a 47% rise. Excluding Hut 8, whose figures were skewed by a significant tax expense, the average cost stood at US$75,767, representing a 35% quarter-over-quarter increase. Despite this, most mining operations remained profitable at the prevailing market price of around US$82,000 per bitcoin. When factoring in non-cash costs such as depreciation and stock-based compensation, the total average cost surged to US$137,018.
We had projected that Bitcoin’s network hashpower would reach 765 Eh/s by the end of 2024. In reality, it slightly exceeded expectations, ending the year at 778 Eh/s. This was largely driven by robust price action, which encouraged miners to accelerate hardware deployment. Our latest projections now indicate that the long-anticipated 1 Zettahash (Zh/s) threshold could be reached as early as July this year, with hashpower expected to climb to 1.28 Zh/s by year-end and reach 2.0 Zh/s by early 2027.
In prior publications we have detailed our methodology on hashrate prediction using a piecewise exponential model — found here.
Hash prices, a critical proxy for miner revenue potential, have shown a modest rebound this year. Nonetheless, our proprietary forecasting model indicates a gradual structural decline, with prices likely to remain range-bound between US$35 and $50 per PH/day through to the 2028 halving cycle. Notably, we anticipate the average hash price will fall below US$40 by Q1 2026, reflecting ongoing efficiency gains and competitive pressure across the mining sector.
Gold and Bitcoin are often compared as scarce, non-sovereign assets. But while much has been written about their investment case as stores of value, fewer draw the comparison at the production level. Both assets rely on mining — one physical, the other digital — to introduce new supply. Both industries are defined by cyclical economics, high capital intensity, and deep ties to energy markets.
Yet the mechanics and incentives of bitcoin mining differ in nuanced ways to gold mining, which ultimately have an outsized effect on how the economics and strategies of industry participants are both set up and play out. This report will walk you through a few of their similarities, but perhaps more valuably, their material differences.
Asset scarcity is derived from physical and computational mining
Gold mining is a centuries-old process of extracting and refining metal from the earth. It requires identifying viable deposits, securing permits and land access, and using heavy machinery to remove ore from the ground — followed by chemical treatment to isolate the metal for downstream distribution.
Bitcoin mining, by contrast, involves repeating a computational process in a race to settle batches of Bitcoin transactions and earn newly issued coins plus fees. This process, known as creating Proofs of Work, involves procuring rack space, electricity, and specialized hardware (ASICS) to efficiently run the calculations — followed by an internet connection to broadcast the results to the Bitcoin network.
In both systems, mining is an unavoidably costly process that underpins the scarcity of each asset: Bitcoin’s is enforced by code and competition; gold’s is by physical and geological location. But the way that scarcity is extracted, the economics of the producers — and how each evolves over time — has few similarities.
Bitcoin mining economics are shaped by competition, hardware cycles, and multiple revenue streams
Gold mining has relatively forecastable economics. Companies model reserve estimates, ore grades, and extraction timelines with reasonable accuracy, though initial estimates can still be far off the mark: about one in five developed gold mining projects proves profitable over its lifecycle. Major costs — labour, energy, equipment, regulatory compliance, and remediation — are well projected in advance. Depreciation mostly amounts to the ordinary wear and tear of machinery or reserve depletion. The primary short-to-medium term uncertainty is oftentimes the notoriously steady market price of gold itself. And on top of that, almost all of these input costs can be effectively hedged.
Bitcoin mining, by contrast, is much more dynamic and unpredictable. Company revenues depend not only on the relatively volatile market price of bitcoin, but on their share of the global hashrate (read: global competition). If others expand their operations more aggressively, your relative output can decline even if your mining operations do not change. It’s an ongoing variance to consider for operators.
Thus, our first difference is that, unlike gold mining, where production forecasts are quite stable, bitcoin miners face the uncertainty of their production being privy to the entry, exit, and strategies of other industry participants.
One of the most important costs of bitcoin mining companies is depreciation, and more specifically the depreciation of ASIC machinery. These chips in bitcoin mining machines keep rapidly improving in their efficiency, urging companies to upgrade to remain competitive, long before such equipment naturally wears out. The implication being depreciation takes place on a timeline of technological progress rather than physical deterioration. It is a major expense — albeit a non-cash one — and marks a key distinction from gold, where mining equipment remains useful longer, having already experienced the majority of efficiency gains in their machinery.
The combined effect of bitcoin production changing at the whims of industry competition and short depreciation cycles is a constant pressure to reinvest in new hardware to maintain production levels — a trap professionals often call the “ASIC hamster wheel”.
Another, this time positive, fundamental difference is in revenue structure. Gold miners earn only by extracting and selling the unreleased supply stored in reserves. Bitcoin miners however earn both from extracting unreleased supply and transaction fees. Transaction fees offer a revenue stream of already released supply to miners, which varies based on the demand to transfer bitcoin. As bitcoin nears its 21 million supply cap, fees will become an increasingly important revenue stream — a dynamic that gold miners simply don’t have.
**Note: y-axis is partially shown with an 80% bottom range.
Last, a mainly long-term advantage of mining bitcoin is the ability to repurpose a by-product of operations: heat. As electricity passes through their machines, it generates substantial thermal energy that can be captured and redirected for secondary use — like industrial processes, greenhouse farming, or residential and district heating. It opens up entirely new revenue streams. The impact of reusing heat will likely grow as mining machines commodify and depreciation timelines lengthen. Likewise, gold miners can benefit from saleable by-products like silver or zinc — elements typically identified as part of project planning and treated as an offset to gold production costs.
Bitcoin mining offers a more promising environmental future than gold
It is well known that gold mining is inherently extractive and leaves a lasting physical footprint: deforestation, water contamination, tailing ponds, and disruption of ecosystems. In many regions, it also raises concerns around land rights and worker safety.
Bitcoin mining, on the other hand, is not physically extractive and entirely electricity-dependent. This opens the door for integration — not conflict — with local infrastructure. Because miners are mobile and interruptible, they can serve as grid stabilizers and monetize otherwise wasted or stranded energy sources like flared gas, overbuilt hydro, or curtailed wind and solar.
Unbeknownst to many, bitcoin mining also shows promise as a subsidy for clean energy and as a way to justify grid connection. By co-locating with renewable or nuclear generation, miners can improve project economics before grid connection — without relying on publicly funded subsidies.
Lastly, while now well documented, it’s worth noting that compared to traditional industries, Bitcoin’s carbon profile is lower on average and more transparent. Arguably, it is even necessary for a smooth transition to renewable-heavy electric grids.
We have almost no rise in energy consumption since the 2024 peak, this is due to ever increasing efficiency in new mining hardware, which is now averaging just 20W/Th, a 5 fold improvement in efficiency since 2018.
Faster cycles and technology reputation define Bitcoin mining investment profiles
Both industries are cyclical and sensitive to the price of the asset they produce. But while gold miners typically operate on multi-year timelines, bitcoin miners can scale operations up or down more quickly in response to market conditions. It makes bitcoin mining more flexible, but also more volatile.
Public Bitcoin mining companies have tended to trade like high-beta tech stocks, reflecting their sensitivity to both bitcoin’s price and broader risk sentiment. In fact, several market data providers classify listed bitcoin miners in the Technology sector, not under traditional energy or materials groupings.
Gold mining companies, however, are longer-established and often hedge future production, which can mute exposure to gold’s price upside and downside. They are generally categorized within the Materials sector and evaluated more like traditional commodity producers.
Capital formation also differs. Gold miners typically raise capital based on reserve estimates and long-term mine plans. Bitcoin miners often raise capital more opportunistically — in recent history often through direct or convertible equity issuance — to fund rapid hardware upgrades or data centre expansions. As a result, they are more dependent on market sentiment and cycle timing, and tend to operate with shorter reinvestment timelines.
Bitcoin mining offers exposure to energy, compute, and future financial networks
Gold and Bitcoin may converge long-term to serve similar macroeconomic roles, but their production ecosystems are structurally different. Gold mining is slower-moving, physical, environmentally damaging and resource-depleting. Bitcoin mining is faster, more modular, and likely to be increasingly integrated with modern energy systems.
For investors, that means bitcoin miners are an imperfect digital analogy of gold miners. Rather they represent a new class of capital-intensive infrastructure, blending exposure to commodity cycles, energy markets, and technology disruption. Those investing with long-time horizons should consider them as a distinct, emerging asset class with unique fundamentals, particularly as transaction fees grow in importance and energy partnerships evolve.
Understanding these nuances is necessary, in our opinion, to make informed investment decisions in the distributed financial system that society is increasingly trending towards.
As an investment, bitcoin miners offer exposure to scarcity, but also a play on the growth of data centre infrastructure, energy markets, and the monetization of compute — a convergence traditional mining simply doesn’t capture.
- Please refer to previous reports where where we explain our methodology.
Hut 8’s reported tax expense per bitcoin was US$281,000, primarily due to a US$93.0 million deferred tax liability related to unrealised gains on the fair value of its bitcoin holdings, alongside elevated taxable income for the full year 2024. Interest expense also rose, driven by the US$150 million Coatue convertible note and additional borrowings from its credit facility with Coinbase. The firm further recognised amortisation of debt issuance costs and expenses tied to promissory notes.
In Q4, Hut 8 signed a hosting agreement with BITMAIN, securing capacity for 15 EH/s of ASICs at its Vega site. To finance this, Hut 8 pledged 968 bitcoins (purchased at US$101,761, pledged at US$104,258) as collateral for approximately 30,000 Antminer S21+ ASICs, priced at US$15/TH. This is expected to boost self-mining hashrate to ~10.3 EH/s, with average fleet efficiency improving to 20.5 J/TH. If the company exercises its purchase option under the BITMAIN agreement, it anticipates increasing total self-mining hashrate to ~25.1 EH/s, with fleet efficiency at 16.0 J/TH.
CleanSpark, Iren, and Cormint were the only miners to lower their cost of revenues per bitcoin quarter-over-quarter, decreasing by 8%, 39%, and 44%, respectively.
CleanSpark’s cost reduction was driven by a four percentage-point improvement in uptime (from 94% to 98%), a ~20% boost in fleet efficiency (from 22 J/TH to 18 J/TH), and a 56% increase in deployed hashrate (22 EH/s to 34 EH/s). Additionally, SG&A, depreciation and amortisation (D&A), and stock-based compensation (SBC) per bitcoin all declined. Tax benefits and a modest rise in interest income also supported profitability. Overall, cash costs fell by 15%, and all-in costs declined by 13%.
Cormint’s Q4 performance rebounded after a challenging Q3. Power prices dropped to 1.8¢/kWh, with a full-year average of 2.3¢/kWh. Operating costs remained tightly managed, allowing scale benefits to flow directly to the bottom line.
Iren’s operating hashrate surged from an average of 12.2 EH/s to 22.6 EH/s, driven primarily by the expansion of the Childress facility. Installed capacity rose from 21 EH/s to 31 EH/s during Q4. Although the company initially targeted 50–52 EH/s in H1 2025, it has since paused these plans, suggesting this milestone is now unlikely.
The switch to spot electricity pricing at Childress reduced electricity costs to 3¢/kWh, cutting net electricity cost per bitcoin mined by 39% (from US$35.4k to US$21.4k). Fleet efficiency improved to 15 J/TH, supporting both profitability and energy savings.
Energy Portfolio MW Growth
Bitcoin miners are increasingly reallocating energy resources towards AI infrastructure, reflecting greater revenue opportunities, access to capital, and investor interest.
Core Scientific (CORZ) and Cipher Mining (CIFR) are leading this transition:
- CORZ has committed 43% of its capacity to AI (either active or contracted).
- CIFR plans to allocate 35% of future capacity to AI, supported by a US$50 million investment from SoftBank.
- CORZ remains the only miner with significant completed AI infrastructure, with over 500 MW under contract.
Q1 results could disappoint as hash price continued its decline due to Bitcoin trading in a narrow range between US$80,000 and US$90,000. Q2 results may show deterioration, as tariffs on imported mining rigs range from 24% (Malaysia) to 54% (China). Miners relying on older or less efficient rigs face higher exposure to these tariffs.
- CORZ is better insulated, as it transitions to HPC. Its capital expenditure remains fixed at US$1.5 million/MW, with new sites expected to incur costs ~20% higher, in line with industry peers.
- BTDR, which manufactures its own equipment, may experience margin pressure on external US sales.
Valuation multiples have compressed, reflecting a growing market consensus that Bitcoin mining is a zero-sum game, where growth by one miner diminishes others’ market share. As a result, miners are pivoting towards data centre infrastructure and HPC hosting, which offer more predictable and diversified revenue streams.
The “Store of Value” narrative is gaining momentum in the US. A Q1 2025 proposal for a Bitcoin Strategic Reserve may be implemented in the future in a budget-neutral way. However, according to Bessent, gold will not be sold or devalued to finance Bitcoin purchases.
Key state-level initiatives include:
- Arizona: Legislation to allocate 10% of reserves (~106k BTC) has reached its final legislative stage.
- Oklahoma: A similar proposal (equivalent to ~17.8k BTC) is at stage four.
- Texas: Considering an allocation of ~2.2k BTC using 1% of reserves.
Combined, these initiatives could result in up to US$10.3 billion in cumulative buying pressure at current prices.
Overall, we believe that most potential macroeconomic scenarios off the back of ‘Liberation Day’ remain supportive of Bitcoin. The introduction of reciprocal tariffs is likely to fuel inflation in both the United States and its trading partners. US trade partners could face elevated inflation while simultaneously contending with growth headwinds. This dynamic may compel them to adopt more accommodative fiscal and monetary policies — measures that often lead to currency debasement, ultimately enhancing Bitcoin’s appeal as a non-sovereign, inflation-resistant asset.
In the United States, the outlook is more ambiguous. Both Trump and Bessent have signalled a preference for lower long-term yields, particularly on the 10-year Treasury. While the underlying motivations can be speculated — such as reducing debt servicing burdens or boosting asset markets — this stance typically supports interest rate-sensitive assets like Bitcoin. At this current moment, the opposite has occurred. US 10-year yields have fallen below 4%, but has since reversed higher back up to 4.5% and now sitting around 4.3% after suspicions of the basis trade unwind, a tarnished US reputation and an increasingly precarious position for the US Dollar as the world’s reserve currency, and entrenched beliefs that Trump will not back down from his tariffs and further spike inflation. However, this crisis is a manufactured crisis and can be quickly reversed with tariff concessions and deals.
However, these signals may also reflect softening forward earnings expectations for equities, raising concerns about an impending slowdown. This presents a key risk for broader markets — and by extension, Bitcoin — if the asset is still perceived by investors as a high-beta, risk-on instrument. In a global downturn, such sentiment could lead to Bitcoin trading in line with equities, despite its longer-term store-of-value narrative.
Despite this, Bitcoin has held up relatively well compared to equity markets since “Liberation Day”. This resilience underscores its unique qualities: a globally tradeable, government-neutral asset with a fixed supply that is accessible 24/7/365. As a result, market participants are increasingly recognising Bitcoin as a credible long-term store of value.
