Debt Capital Markets in Bitcoin Mining (Part I)
On the genesis of debt in bitcoin mining and common debt structures
Originally published on the Braiins blog, this two-part series provides a high-level overview of the past, present, and future of debt in bitcoin mining. Part I explores the genesis of debt in bitcoin mining and the most common structures. Part II discusses key considerations for borrowers and lenders, the effects of leverage on returns, and how the market might evolve in the future.
The History of Debt in Bitcoin Mining
Disclaimer: this series is biased toward western markets (specifically North America). While China was the Mecca of bitcoin mining long before industrial-scale projects arose in the US, the author and audience alike are likely more focused on the western hemisphere. As such, we will focus exclusively on the debt capital markets in North America, despite the fact that many similar products were offered in China and other countries.
Thanks for reading Cathedra Bitcoin’s Substack! Subscribe for free to receive new posts and support my work.
In the 14 years since the Bitcoin network launched, we have seen bitcoin miners evolve from cryptography enthusiasts hashing with CPUs, to profit seekers investing in more and more efficient hardware (i.e., the lowest energy consumption per hash). From CPUs, miners quickly moved to GPUs, FPGAs, and finally to their own application-specific hardware, bitcoin mining ASICs.
In these early days, there was little to no debt financing available to bitcoin miners. The reasons were various:
The useful life of mining hardware was very short, with certain models remaining competitive for only 6 – 18 months before newer, more efficient models rendered them economically obsolete.
The short lifespan of ASICs paired with bitcoin’s price volatility led to highly volatile returns for mining investments, making assets and projects very difficult to underwrite.
Most lenders didn’t understand bitcoin or bitcoin mining. And with bitcoin mining being such a small market (annual mining revenue didn’t exceed $1 billion until 2017), there was little incentive for lenders to pay attention.
The “industry” was still immature; participants were aggressive, and often unprofessional. Many of the new mining companies and hardware manufacturers proved to be outright scams (I leave the reader to judge how much this has changed).
For the above reasons and others, bitcoin mining has tended to lag the rest of the “crypto” industry in terms of debt products. Debt has long been available for traders on offshore exchanges looking to lever up 100x on perpetual swaps, but it took until 2017 for bitcoin-backed debt to become available en masse to retail and institutional borrowers in the US. Retail bitcoin-backed debt was pioneered by companies like Unchained Capital and BlockFi, while trading houses like Genesis and Galaxy Digital provided the early collateralized debt products for institutions. In this era, most large bitcoin-backed loans were bespoke trades offered to hedge funds and sophisticated high net worth individuals, but the ultra-low interest rate environment and bitcoin’s utility as pristine collateral attracted more capital and lenders to the asset class. The price appreciation of bitcoin (“number go up”) also played a role, as more individuals and institutions looked to borrow against their bitcoin.
In the mining world, the 2017/2018 era also saw the first publicly traded bitcoin miners, with the likes of Riot Blockchain (which pivoted from a pharmaceutical technology company to bitcoin mining in 2017), DMG Blockchain, Hive Blockchain (which pivoted from a minerals exploration company in 2017), Hut 8, and Fortress Blockchain (now Cathedra Bitcoin) leading the charge. Others soon followed, with Bitfarms going public in 2019 and Marathon Digital completing its pivot in 2020 (after an initial bitcoin mining pilot in 2017). Outside of public markets, this era also saw the rise of then-private, “institutional” bitcoin miners in North America like Argo, Compute North, Core Scientific, Griid, Greenidge Holdings, and others.
These companies were primarily financed with equity; however, debt financing for bitcoin miners also began to emerge. In 2018 Galaxy Digital issued a $16m term loan to Hut 8 to finance new mining infrastructure, collateralized by bitcoin and infrastructure. The year 2018 also saw the founding of Arctos Capital and Blockfills, early pioneers in ASIC-backed debt in North America. Shortly thereafter DCG and Galaxy Digital both spun up their own mining businesses each with their own financing products for miners (Foundry (2019) and Galaxy Digital Mining (2020)).
From 2019 through 2020, a variety of smaller debt transactions were executed, as capital-deprived miners sought whatever forms of financing they could during the bear market. The lower velocity of the bear market also provided lenders the opportunity to fine-tune their structures, underwriting, and servicing capabilities.
The growth of debt during this period was enabled by a variety of factors:
Commoditized hardware. The 2018 crash drove many hardware manufacturers out of business (or at least atrophy into insignificance). By summer 2019, the hardware market was dominated by Bitmain, Canaan, and MicroBT. This greatly simplified the buying and selling of machines, with “rigs” becoming increasingly fungible. These manufacturers released roughly similar machines (in terms of efficiency) on roughly similar timelines.
Longer ASIC lifespans. The Bitmain Antminer S9 was first released in 2016 but was still in wide use up until the 2020 halvening (arguably, it still is). This suggested that ASICs were now investible for the long-term, and that step-changes in hardware efficiency were perhaps a thing of the past. Tangentially, these first two factors also led to an increasing focus on power costs in mining, as access to the latest hardware no longer provided a durable competitive advantage.
The growth of mining in North America. Due to its permissionless nature, anyone can mine bitcoin anywhere in the world. However, the United States and Canada’s robust capital markets, abundant energy, and respect for property rights led to a meaningful increase in mining investments here from 2017 onward. This growth accelerated in 2021 following a bitcoin mining ban in China, as miners sought refuge in the regulatory environment of North America.
Near-zero interest rate environment. As mining became more and more common in the epicenter of the global financial system, pockets of yield-starved capital salivated at the prospect of 8 – 30% debt (despite the obvious risks).
Industry “maturation.” The emergence of the aforementioned “institutional” mining companies gave the sense that this was now a real industry that could be trusted with credit. Many of these companies were run by converts from the 2017 bull run, bringing with them corporate sensibilities and relationships from other industries. Many thought the days of Shenzhen “chicken shack” mining were over, the network instead to be dominated by “Big Box Miners,” with hyperscale data centers built atop freshly waxed floors.
Number go Up (“NgU”). After bouncing back from the bottom of $3k in Q4 2018, it became increasingly clear that bitcoin was here to stay. And the rally in 2019 gave many hope that they would see the moon once again (soon™).
By the end of 2020, bitcoin mining debt capital markets had expanded to include a variety of competing lenders (Foundry, Galaxy Digital, NYDIG (which acquired Arctos), BlockFi, and others) offering a variety of debt products.
The foundations were laid. So, when bitcoin price took off in Q4 2020/Q1 2021, a boom in mining investments followed. Throughout 2021, equity capital flocked to bitcoin miners, pumping mining stonks in the process. These rising asset and equity values enabled miners to access increasing amounts of debt. Highlighting this point, several large equity raises of 2021 were accompanied by large debt facilities. As this period also saw the rise of ASIC “futures orders,” it even became possible to secure debt financing on machines that had not yet been delivered. The “old” generation of “institutional miners'' used this as an opportunity to catapult themselves to hyperscale. Simultaneously a variety of new entrants jumped onto the scene with multi-billion-dollar SPACs on zero deployed hash rate.
One can’t help but admire the supporting macro factors that contributed to the perfect storm of this capital binge for North American miners. The COVID stimulus left many markets flush with capital. Energy prices remained low as the commodity bear cycle played out its final act. Bored Patagonia-clad PE bros with a recently piqued interest in “crypto” sought to take advantage of the public/private market arbitrage of hash rate valuations, in hopes their returns might bankroll yachts and bottle service at E11even. All while China banned bitcoin mining, suppressing hash rate during bitcoin’s scenic route from $60k down to $30k and then back up to $69k
But like all binges, the bitcoin mining hype cycle of 2021 was followed by a brutal hangover. The bear market of 2022 has seen many darlings of the last few years turn into distressed assets. Several large-scale public miners have been forced to restructure their balance sheets or file for bankruptcy.
And while debt may have been a new tool for bitcoin miners, credit markets have been around for thousands of years, and the key principles of debt are largely unchanged. In this next section, we will explore these principles, before evaluating the current array of bitcoin mining debt products.
Principles of Debt
If you consider yourself fluent in the concepts and parlance corporate fuh-nance, feel free to skip this section.
Wikipedia defines “debt” as “an obligation that requires one party, the debtor, to pay money or other agreed-upon value to another party, the creditor.” Such obligations can take many forms, but there are a few key principles that underlie any piece of debt.
Capital stack overview
To understand debt, first one must understand the idea of the “capital stack.”
For our purposes, all forms of financing can be grouped into two buckets: debt, and equity. Debt holders are considered “contractually senior” to equity holders. This means that debt holders have first claim to the assets and are therefore paid out first in the event of a bankruptcy or liquidation. As a result, debt is considered “less risky” than equity, and therefore has a lower expected return. The debt and equity portions of the capital stack can be segmented further; in order of seniority: senior debt, junior or mezzanine debt, preferred equity, and equity. There could also be multiple types of debt or equity that are the same level of seniority, or pari pasu, with each other.
The main reason you would try to fund an enterprise with debt instead of equity is to enhance returns. If you want to buy a $10 project, you can either use $10 of your own money, or you can borrow $5, and—so long as you service the debt—can keep all profits (after debt service) for yourself (on a lower denominator of invested capital). We will explore the impact of leverage on bitcoin mining returns in more detail in a later section.
Lenders structure debt facilities with the goal of getting paid back full principal and interest regardless of other exogenous factors. To do so, lenders seek to give themselves as much protection as possible; “protection” meaning tools that allow them to minimize the risk of not getting their money back. These tools are frequently referred to as “credit enhancements” and come in several forms.
The lender will first evaluate the overall creditworthiness of the borrower. Creditworthiness is a measure of the borrower’s ability to pay back the lender. Less “creditworthy” borrowers will typically have additional credit enhancements to increase the lender’s comfort in getting repaid.
The simplest form of “credit enhancement” is collateral. Lenders often require borrowers to pledge assets to “collateralize” the loan. This means that if the borrower were to fail to repay (or “service”) the debt, the lender could claim and liquidate the collateral. For example, if you buy a house using debt financing from a bank (take out a mortgage), you pledge your home to the bank. If you fail to make good on your debt payments, the bank can reclaim your house (foreclose).
Debt that is backed by a specific asset (a car, house, or bitcoin mining ASIC) is typically referred to as “asset-backed,” though like all finance jargon, the boundaries of the definition can be murky. In general, if debt is being used to finance the purchase of an asset, lenders will require that asset to be part of the collateral, but lenders might also request additional collateral. Additionally, just because a loan is “asset-backed” doesn’t mean the creditworthiness of the borrower is no longer relevant; asset-backed lenders focus on the overall creditworthiness of the borrower plus the valuation of the underlying assets.
The value of collateral relative to the amount of debt is very important. This ratio has a variety of names, but the most common is Loan-to-Value or “LTV” for short (other variations include “collateralization ratio” and “advance rate”). LTV is calculated as the size of the outstanding principal (the “L”) divided by the value of the collateral (the “V”). This provides a simple way for lenders to calculate their protection in the event they had to foreclose on the loan. For example, if a lender had to foreclose on an 80% LTV loan and liquidate the collateral, the asset price could fall by 20% before the lender took a hit. In other words, the higher the LTV, the riskier the loan. This example also demonstrates the importance of collateral liquidity and volatility; the more liquid and less volatile the collateral, the less risky the loan.
Some loans do not require collateral to be pledged to the lender; these are called “unsecured” loans as the debt is not “secured” by any collateral. Lenders will still consider the borrower’s assets and credit worthiness when sizing and pricing unsecured loans, the lender just has no specific legal claim (or “perfected lien”) against any assets. These loans are typically only available to people and organizations with strong credit credentials, as a loan secured by specific collateral has a lower risk of loss than an unsecured loan. Such loans are rarely available to bitcoin miners.
As means of further protection, lenders often also stipulate specific rules around what a borrower can and can’t do; such rules are called “covenants.” Covenants might stipulate the exact use of an asset, certain rules around other activities of the borrower, or might require certain actions from the borrower should things happen that are entirely outside of their control (e.g., margin calls or uptime requirements).
Cost of capital
The above will all help determine the cost of debt (the interest rate). This cost of debt stems from the time value of money. The lender needs to be compensated for giving their capital to the borrower. The amount they need to be compensated is based on the probability of repayment. The riskier the debt is perceived to be, the more expensive that debt will be. In the words of Art Lyon, "There's no bad risk, just a bad price.” Notably, the price of risk varies dramatically with market conditions.
Likewise, a lender’s cost of capital is also important in determining the interest rate a borrower pays. A lender can only offer loans at a cost of capital that is higher than the cost they pay. In fact, a lender’s whole business model is lending out money at a higher interest rate than their own cost of capital; the difference between their interest income and interest expense is their profit, or in debt-speak their “net interest margin” (“NIM” for short). For this reason, in emerging debt capital markets, the cost of capital is typically high as the first lenders likely have higher cost of capital themselves. As the market matures, larger lenders with lower cost of capital often push out these pioneer species by undercutting their rates.
To this end, the existence of a secondary market for loans in an asset class can influence the cost of debt for those loans. If a lender can sell loans to other investors, then they can lock in a return and free up capital to generate more loans, which can drive down the cost of debt for borrowers. This is the principle behind securitization markets, and is a large reason why certain forms of debt like car loans and mortgages are so cheap. The larger the secondary market for these loans, the easier it is for lenders to originate and distribute, earning net interest margin in the process.
Debt Products for Bitcoin Miners
The above principles underlie all debt, but most industries have a limited array of readily available debt products, typically tried and true structures that have stood the test of time. We will now walk through the existing universe of debt products available to bitcoin miners.
As a reminder, asset-backed debt refers to loans backed by a specific asset. There are three main categories to consider: bitcoin-backed, ASIC-backed, and infrastructure-backed.
Bitcoin-backed debt is one of the oldest structures available to bitcoin miners. The logic is simple: a miner can deposit bitcoin as collateral to the lender and receive a fiat-denominated loan in exchange. Bitcoin has many properties that make it particularly attractive collateral; it’s a digital bearer asset that trades 24/7 on a variety of highly liquid venues. As a result, it typically offers the lowest cost of capital for bitcoin miners. However, bitcoin’s price volatility means LTVs can fluctuate dramatically; for this reason, lenders often have margin call provisions (if bitcoin falls below a certain price, the borrower needs to contribute additional collateral or repay part of the loan). The interest rates for these loans vary depending on the lender and market conditions, ranging from 4 – 12%. The loans have varying terms, but they typically don’t amortize, meaning there may not be any mandatory principal payments until the maturity date.
ASIC-backed debt emerged as the structure du jour during the last bull market. These loans typically had 18-month terms for an all-in cost of capital ranging from 13 – 28%. This is notably more expensive debt than bitcoin-backed loans, but it gave new miners without large bitcoin treasuries access to a new pocket of capital. In some transactions, miners could even finance machines before they were delivered (e.g., if a miner entered into a futures order, they could secure debt financing before the machines were deployed). However, once the financed machines are plugged in, the loans start amortizing (meaning principal is paid down gradually over the term). Similar to bitcoin-backed debt the value of the collateral (ASICs) can be quite volatile, resulting in similar LTV fluctuations. Unlike bitcoin-backed debt, however, it’s hard to employ a margin call on ASICs. This has caused some lenders to offer low LTVs (sub 50% in some cases), but even still, the market turmoil of H2 2022 has resulted in many such loans becoming undercollateralized (LTVs > 100%). In terms of covenants, these loans might stipulate certain uptime requirements, prohibit the use of firmware and immersion, or other operational rules to ensure the value of the collateral is preserved.
Infrastructure-backed loans for bitcoin miners also appeared in the last 24 months, allowing miners to gain financing on the supporting infrastructure at a mining site. The loans range from utilities financing substations to mining savvy lenders providing leverage against containerized data centers. Infrastructure-backed lending (or “project financing”) is quite common in other industries (including and especially energy), however these traditional lenders typically are not comfortable lending to bitcoin miners. Legacy industries are often able to borrow against infrastructure at higher LTVs, lower interest rates, and over longer terms (up to five years). But given the volatility of bitcoin mining, miners likely only have access to such loans at materially worse terms than traditional industries like energy. The primary lenders currently serving bitcoin miners with infrastructure-backed debt are either the small subset of legacy institutions who have gotten comfortable with bitcoin mining risk (e.g., certain utilities and energy infra lenders) or smaller lenders who are comfortable taking a bitcoin miner’s power or data center infrastructure as collateral. Infrastructure-backed debt is similar to ASIC-backed debt in that the collateral is a physical asset that can be liquidated (in theory at least), but there is much wider variance in the terms. These loans have yet to become a meaningful portion of the market, but they do exist.
Instead of borrowing against a specific asset, miners also might choose to borrow money at the corporate level. A common example of this would be a “Senior Secured Term Loan,” which typically involves an “all asset lien,” meaning that the lender would take substantially all the borrower’s assets as collateral. These loans can also have more complex features, such as rules around accessing some proceeds only after certain operational achievements. There are a few benefits from the lender’s perspective: first, the structure is simpler as there is little risk of multiple pieces of debt that may be in conflict with each other; second, the lender has a claim to all cash flows from the business not just specific assets; and third, if the borrower does default, the lender can just take ownership of the existing operations without being forced to sell collateral on a potentially illiquid market. From a borrower’s perspective, these loans are a simple way to get debt financing, but they offer less flexibility. These loans are typically only available to large miners and may have unique covenants and rules around unlocking proceeds. It’s worth calling out that such loans have not been the primary forms of debt financing for bitcoin miners, but they may become more common in the future.
There are other types of corporate debt as well (e.g., convertible debt and other bespoke structure), but they are not common enough to warrant evaluation in this essay.
It is worth flagging that this table is obviously prone to fluctuations in the market (bitcoin price, mining economics, broader macro factors, etc.) as well as borrower-specific considerations. This overview of different structures should be viewed as a high-level map, outlining the current landscape of debt for bitcoin miners (and perhaps provide context for some of the recent news in bitcoin mining).
Next, in Part II, we will look at some high-level considerations for lenders and borrowers alike, play with some math to see the effects of leverage on bitcoin mining operations in a variety of market conditions, and lastly reflect on how the market could evolve in the near to medium term.
This article was written for the Braiins blog by Drew Armstrong. Drew is the Chairman, President, and COO of Cathedra Bitcoin, a company that believes sound money and abundant energy are the keys to human flourishing. Prior to joining Cathedra, he was a founding member of Galaxy Digital’s bitcoin mining team and helped build out Galaxy's mining equipment finance product. Drew began his career at Barclays' investment bank, where he focused on the origination of esoteric securitized products, such as data center securitizations and collateralized fund obligations. His views here do not reflect those of any of his past, present, or future employers. Follow Drew on Twitter.
Thanks for reading Cathedra Bitcoin’s Substack! Subscribe for free to receive new posts and support my work.