Borrowing Against Your Stack

Today the magnitudes are smaller because bitcoin's trailing four-year CAGR has compressed from its early-period highs. But the structural argument is identical. The reader making this decision today, with 2 BTC at $110,000 facing a $50,000 medical bill, is in exactly the same structural position as the reader years ago with 100 BTC at $100. The asymmetry is the same: the upside of borrowing-and-retaining is the entire future appreciation of the bitcoin that would have been sold. The downside is the interest cost of the loan§Regulatory note · Interest deductibilityWhether interest on a bitcoin-backed loan is tax-deductible depends entirely on what the proceeds funded (the IRS “loan tracing” rules). Deductible against investment income if proceeds went to taxable investments; deductible as a business expense if used for business capital. Not deductible if used for personal consumption (lifestyle spending, vacations, debt consolidation). Keep documentation tying each dollar of borrowed money to its use — the burden of proof is yours., structured against the tail risks that the rest of this page exists to detail.

The point isn't that borrowing is always better than selling. The point is that the decision can't be evaluated honestly without weighing both paths. The reflexive “just sell some” answer treats the sale as if it has no cost beyond the immediate fiat received. It does. Selling forecloses the future of the coin sold.

What this page actually is

This page isn't promoting borrowing against bitcoin. It's also not against it. The page exists because the decision is consequential enough — and the landscape complex enough — that blind entry into this strategy is the worst possible outcome. We've seen what blind entry produces: Celsius depositors becoming unsecured creditors, BlockFi customers locked for two years, Genesis funds tied up in bankruptcy proceedings, retail holders surprised to discover that the lender they trusted was lending out their bitcoin to leveraged third parties they had no idea existed.

The page is built around a specific editorial mission. The reader walks away in one of two states. Either they've identified the lowest-risk, highest-integrity version of borrowing — the version that meets the standards described below — and concluded the trade is worth making for their specific need. Or they've concluded, on reflection, that the trade isn't worth it for them, and they continue to HODL.

Both are good outcomes. The bad outcome — the one this page exists to prevent — is the reader sleepwalking into one of the wrong versions because they didn't know what to look for.

A practical framing that helps the self-assessment: distinguish need from want. The structural argument for borrowing rather than selling is strongest when the alternative is selling to fund a real life-event need — surgery, education, a primary home, business capital that will generate income, a tax obligation. The argument weakens substantially when the alternative is waiting — when the immediate liquidity would fund something you could equally fund later, in fiat, after the bitcoin has had more time to appreciate. Selling some bitcoin in a few years to buy the same Ferrari is structurally a different trade than borrowing against the stack to buy it now and incurring new risks before then. The patience option is usually free. The borrowing option carries new risks that didn't exist before you originated the loan.

If you can wait, wait. The strategy on this page is for the cases where you genuinely can't.

Up front: what we recommend, and what we don't

The cautionary record below is detailed. Before walking through it, here's what the page actually concludes — so the failure cases don't leave you running for the exits before we've had the chance to point you at the responsible versions.

Approaches that meet a bar of serious consideration:

• Collaborative-custody multisig (Unchained, Anchor Watch, Onramp) — your bitcoin sits on-chain in a multisig you or anyone can verify at any moment. The highest standard.
• Tier-1 non-rehypothecating CeFi (Ledn, APX Lending, Strike, Salt, Arch, Aven, Figure, Coinbase's Morpho-integrated lending) — full custody but explicit, contractual, regulated§Regulatory note · State licensingMany US states now require crypto lenders to hold specific licenses to operate — New York mandates a BitLicense; California's DFAL took effect July 2026. As of 2026, 19 states have laws criminalizing unlicensed crypto lending operations (New York's CRYPTO Act treats >$1M annual volume as a Class C felony). The implication for borrowers: whether a given Tier-1 platform can serve your state depends on its licensing footprint, not its willingness. Confirm coverage in your jurisdiction before originating., transparent. Convenient and credible, with the trade-off of full counterparty trust.
• Bitcoin-native Layer 2 lending (Citrea, Stacks, Rootstock) — small but architecturally promising; where the long-term trajectory points.
• Bitcoin-collateralized residential mortgages — Milo, Better/Coinbase's Fannie-Mae-conforming product, Horizon. Long-duration, much higher overcollateralization, structurally different enough from the other categories that they get their own treatment on a dedicated page — see Bitcoin-Backed Mortgages →. The category is bitcoin-only at the GSE level (no crypto, no gold), which is itself a meaningful indicator of where regulated finance has landed on collateral quality.

Approaches we cover briefly for completeness but don't take seriously:

• DeFi via wrapped BTC (Aave, Morpho, Compound, Euler, etc.) — the April 2026 Kelp DAO / Aave / Arbitrum cascade demonstrated definitively that these systems are centralized in practice. Detailed below.
• Rehypothecating CeFi (Nexo, Binance Loans, Matrixport) — the structural failure mode that produced every catastrophic collapse of 2022.
• Predatory-LTV platforms (YouHodler-style 90/97 LTV) — included only as a concrete example of what aggressive terms look like on paper.

If you find the cautionary record concerning, that's the appropriate response. The point isn't to scare you out of the entire topic. It's to scare you out of the wrong versions of it.

Two strategies, one mechanism

Borrowing against bitcoin can be one of two distinct things, and conflating them is the most common mistake retail readers make.

The tactical use

You need liquidity for a specific event — a home down payment, business capital, a tax bill, a medical emergency. Selling bitcoin would trigger capital gains and end your position in those coins permanently. Borrowing lets you fund the need while keeping the stack intact§Regulatory note · §1259 constructive saleBorrowing is tax-deferred only as plain unhedged borrowing. Combining the loan with offsetting positions that eliminate price risk (a short, a full hedge) can trigger a “constructive sale” under IRS §1259 — the gain you were trying to defer comes due immediately. The PARITY Act of 2026 confirmed §1259 applies to digital assets. The trap is hedged structures, not borrowing itself., with the loan repaid eventually from external income, refinancing, or — if the situation worsens — eventually from the stack itself. This is the framing most readers start with, and for the use cases the previous section called needs, it's often the right one.

The strategic use

The asset compounds faster than the cost of debt. Debt-to-value ratio declines over time even as you keep borrowing. The stack never has to be sold. Under current US tax law, your heirs inherit at step-up basis, eliminating the capital gain on the appreciated portion entirely. This is the Buy-Borrow-Die strategy that has shaped how wealthy families relate to appreciating assets for over a century — the NYC real-estate dynasties leveraged a version of it; Michael Saylor articulated the bitcoin version explicitly through MicroStrategy (now Strategy) for years.

The math of perpetual borrowing is real, but it isn't easy. If asset growth rate G exceeds your borrowing rate R sustainably, a loan taken today at LTV of X/V₀ becomes X(1+R)/[V₀(1+G)] one year later. When G > R, that fraction shrinks. Compound across decades and the LTV trends down, not up, even as you draw further loans against the appreciating collateral.

The historical examples support the principle, though with more nuance than the folklore version. The NYC dynasties (Astors, Vanderbilts, Rockefellers) did hold real estate across generations using a version of this strategy — but they also sold selectively, used trusts, and benefited from estate planning machinery far beyond borrowing. Jeff Bezos has used Amazon-stock-backed credit, but he has also been a sizeable net seller of stock over the years.

The Saylor case is recent and revealing. After five years of explicit “never sell” public commitments — “Sell a kidney if you must, but keep the BTC,” “BTC is the exit strategy” — Strategy's Q1 2026 earnings call on May 5 included an explicit reversal. CEO Phong Le confirmed the company will likely sell some bitcoin to fund preferred dividend obligations, framing it as “math over ideology.” Saylor himself, in a follow-up interview, reframed his prior “never sell” rhetoric as “strategic brushback aimed at short sellers” — i.e., positioning years of public commitments as marketing rather than literal policy.

This is the most committed institutional practitioner of the perpetual-borrowing thesis acknowledging that the strategy has limits when balance-sheet obligations require it. The thesis isn't undermined by this — Strategy is still a buyer on net, the long-run logic still holds — but it becomes more honest about itself. Even Strategy hedged. Anyone modeling this strategy for themselves should know that.

Which strategy you're actually running depends on whether your asset's compound growth rate exceeds your interest rate across your time horizon — and, more practically, whether you can survive the drawdowns along the way. The Power Law channel and a conservative LTV are the two levers that decide the answer. The Calculator on the next tab makes this concrete.

The four categories of bitcoin lender

Not every lender operates with the same ethos. The bitcoin-aligned approaches — those built around verifiable on-chain segregation, no rehypothecation, transparent collateral, and bitcoin-only focus — sit on one side of the market. The crypto-ethos approaches sit on the other, with material structural differences in how your collateral is held and what happens when anything goes wrong.

Bitcoin-aligned approaches

Collaborative-custody multisig. The structural gold standard. Your collateral is held in a 2-of-3 multisig (Unchained's flagship model) or a Multi-Institution Custody arrangement spreading keys across independent qualified custodians (Onramp, Anchor Watch). In every version of this model, your bitcoin remains on-chain, identifiable, and provably yours — not commingled with the lender's other operations, not lent out to third parties, not legally transferred to the platform on deposit. The collateral's presence and segregation is verifiable by anyone, at any time, cryptographically. This is the only category in which bitcoin's foundational rule — don't trust, verify — actually applies to the rehypothecation question. Anchor Watch goes further with Trident time-locked recovery paths and Lloyd's-of-London insurance up to $100 million against key loss or physical coercion. The cost of this category is higher (typical APRs run 12–16%) precisely because the lender can't subsidize rates by re-lending your collateral.

Tier-1 non-rehypothecating CeFi. The convenient entry point. Ledn, APX Lending, Strike, Salt, Arch Lending, Aven, Figure, and Coinbase's Morpho-integrated lending product all operate with full custody but explicit no-rehypothecation policies. Your bitcoin sits in segregated accounts at qualified custodians (BitGo, Fidelity Digital Assets, Anchorage Digital). The trade-off is that the no-rehypothecation guarantee is contractual — backed by a written policy, regulatory oversight, and bankruptcy law — but not verifiable in the way an on-chain multisig is. You're trusting the platform's operational integrity to honor what it has promised. Ledn's S&P BBB- rating on its 2026 securitization is the strongest institutional endorsement any of these have achieved.

A late-April 2026 development inside this tier worth singling out. At the Bitcoin 2026 Conference, Strike announced — in partnership with Tether — a “volatility-proof” bitcoin-backed loan structure paired with a $2.1 billion Tether-backed credit facility, a lending proof-of-reserves mechanism that lets borrowers verify their collateral is segregated in a distinct on-chain address, and rate compression to roughly 7.49–10.5% APR (lower end on loans above $5M, upper on loans below $250K) — meaningfully below the historical no-rehypothecation tier pricing. The structural claim is that forced liquidation from price moves alone has been removed; the mechanism is proprietary and not yet publicly documented, but the most plausible inference is that the Tether credit facility absorbs the price-volatility risk on Strike’s balance sheet rather than passing it to the borrower as a margin call. This is the same end-state the Coinbase / Better / Fannie-Mae-conforming mortgage product achieves through a different route — the GSE wrapper plus a 60-day-delinquency-only liquidation trigger. Two distinct 2026 products both removing the margin-call mechanic, via structurally different paths: one through legal structure and federal regulation, the other through private balance-sheet absorption.

What Strike’s product brings to the BAS use case is the operational feel of a credit facility rather than a discrete loan event — you draw, repay, draw again, at CeFi rates that have begun to compete with traditional lines of credit. It sits at the operational end of the BAS spectrum, with the Coinbase-Better mortgage product at the life-event end. Both belong to the Tier-1 non-rehypothecating CeFi category structurally; they differ in use-case texture and underwriting weight — the mortgage product is a 30-year commitment around a defined home purchase; Strike’s offering is closer to always-on liquidity that the user reaches for whenever cash flow needs it. The structural-opacity caveat applies. Unlike BBM’s publicly documented Fannie-Mae-conforming features, Strike’s “volatility-proof” is a private contractual feature backed by Tether’s balance sheet; several sources at announcement noted that “the technical details are still unclear.” The contractual structure for what happens if Tether itself faced balance-sheet pressure is the piece worth watching as the product matures.

Bitcoin-native Layer 2 lending. The architecturally promising frontier. Citrea (a ZK-rollup that launched mainnet in January 2026, using BitVM2 for trust-minimized exits), Stacks with its sBTC primitive, Rootstock with its long uptime record. The structural argument: bring smart-contract functionality to bitcoin without ever wrapping it onto a foreign chain. Still small in 2026 — TVL across the category remains modest compared to Ethereum-based DeFi — but the trust-minimized exit primitive (BitVM2) addresses the structural failure mode that has defined cross-chain bridges for the last six years. Worth monitoring as the next several years unfold.

Crypto-ethos approaches (covered for completeness only)

DeFi via wrapped BTC. Aave, Morpho Blue, Spark, Compound V3, Euler V2 are the most prominent. The marketing claim is that smart contracts replace human custody, oracles replace platform operations, code replaces fiduciary responsibility — that the system is trustless and decentralized.

The April 2026 Kelp DAO / Aave / Arbitrum cascade definitively refuted that claim through the protocols' own behavior. An attacker exploited a 1-of-1 LayerZero verifier configuration, forged a cross-chain message, and minted 116,500 unbacked rsETH (~$292 million). Within 46 minutes, Kelp DAO's emergency pauser multisig froze rsETH contracts across Ethereum mainnet and 20+ L2s. Within hours, Aave's Protocol Guardian froze rsETH markets on V3 and V4, then proactively froze WETH markets on Core, Prime, Arbitrum, Base, Mantle, and Linea. SparkLend, Fluid, Compound V3, and Euler independently froze their rsETH markets. Ethena paused its LayerZero bridges. Lido paused earnETH deposits. Three days later, the Arbitrum Security Council unilaterally froze 30,766 ETH (~$71M) connected to the exploiter's wallet, moving it into a governance-controlled custodial holding account. An industry-wide $300 million coordinated backstop, branded DeFi United, was organized to make affected holders whole. Weeks later, Aave's DAO voted to deliberately adjust oracle parameters to create a deficit inside the attacker's positions — manipulating the price feeds the protocol depends on to make forced liquidations against a specific actor technically possible.

That sentence makes the case better than this page can rephrase it. What the 2026 cascade demonstrated isn't that DeFi has occasional security incidents — it's that the entire stack is built on admin keys, guardian multisigs, security councils, and discretionary governance interventions, and that all of those mechanisms are deployed instantly when the system is under stress. For an audience that understands what bitcoin's actual decentralization means, evaluating these systems as serious lending venues for the bitcoin stack is a category error. The infrastructure that competes with bitcoin-aligned lending is centralized — just centralized without KYC, and without the regulatory accountability that comes with explicit custody.

This page treats DeFi as a confirmation of dismissal rather than a risk to manage. The category's failures aren't the reason to avoid it; the response to those failures is.

Rehypothecating CeFi. Nexo, Binance Loans, Matrixport. These platforms subsidize their headline rates (Nexo advertises borrowing rates as low as 2.9%) by lending out user collateral to third parties — generating yield on assets that don't belong to them, with the profit funding the lower stated rates. This is the structural failure mode that took down Celsius, BlockFi, Genesis, and Voyager between 2022 and 2023. Subsidized rates aren't a benefit; they're a signal that the lender is making money somewhere structural — and that somewhere is your collateral commingled with other obligations.

Predatory-LTV platforms. YouHodler offers up to 90% opening LTV with a 97% liquidation threshold — a ~7% drawdown to liquidation. These platforms exist legally and have customers; the only useful purpose served by mentioning them on this page is to make concrete what aggressive terms look like on paper, so readers can pattern-match against them when they appear elsewhere.

The single biggest variable: rehypothecation

If you remember nothing else from this page, remember this: rehypothecation is the structural axis on which every other risk turns.

Every catastrophic failure of bitcoin-backed lending in the last decade — Celsius, BlockFi, Genesis, Voyager — was a rehypothecation failure at its core. The lender re-lent depositor collateral to third parties to generate yield, those third parties were over-leveraged, the contagion propagated, and depositors discovered they'd been operating in a black box where the lender owed them their bitcoin back but no longer had it.

What the post-2022 regulatory environment has done is eliminate most of the explicit rehypothecation models from the regulated tier. What it hasn't eliminated is the underlying temptation. Lenders that can't lend out collateral can't subsidize rates. The “no-rehypothecation premium” — the higher APR you pay for the safer structure — is the visible cost of buying out of the systemic risk Celsius exposed.

For a bitcoin-ethos audience, there's a clear hierarchy of confidence:

The hidden-complexity warning extends beyond direct lenders. Structured products that wrap bitcoin or bitcoin-adjacent equities (covered-call ETFs, leveraged BTC notes, certain MSTR-derivative products) often disclose leverage and effective rehypothecation deep in their prospectus fine print — page-40 territory, where retail buyers rarely read. The structural pattern is the same: implicit leverage that's not visible on the surface, structured to be discovered only when something has already gone wrong.

The cautionary record

A few words on what the historical record actually shows. Every catastrophic failure of bitcoin-backed lending in the last decade was a failure of the rehypothecating-CeFi or DeFi categories described above — not the bitcoin-aligned approaches the page recommends. Celsius, BlockFi, Genesis, Voyager all engaged in the practice that defines the bottom tier of the rehypothecation hierarchy. The DeFi cascade detailed earlier was an architecture-level failure of the wrapped-BTC category. None of these were collaborative-custody multisig failures. None were Tier-1 non-rehypothecating CeFi failures. The collapses on record are the predictable failures of the structural choices the page already tells you to avoid.

That framing matters because the same failures, read without it, can produce the wrong takeaway — that lending against bitcoin is dangerous in general, rather than the actual lesson: that the specific structural choices these companies made were dangerous, and avoiding those choices is the point of this page.

The four lessons that emerge across the cluster:

• Ownership is a legal fiction without the keys. Depositors on centralized platforms are legally categorized as unsecured creditors in insolvency — your “ownership” is a contractual claim, not a property right. (This is what collaborative-custody multisig structurally avoids.)
• Yield is the price of tail risk. Aggressive yields and aggressively low borrow rates are lagging indicators of aggressive rehypothecation. If the platform's economics don't make sense at face-value rates, they're being made up somewhere structural.
• Technical liquidation is a certainty during stress. Even healthy positions can be force-liquidated when oracles fail, networks congest, or platforms go down — the Infrastructure Failure Gap where you have the funds but can't physically deliver them.
• The only zero-counterparty position is self-custody. Every other configuration is a bet on someone else's integrity, solvency, or code.

When borrowing makes sense

Two structural conditions decide whether borrowing against bitcoin works for you. Both are addressed concretely by the calculator on the next tab; they're framed here as the conceptual question the calculator answers.

The first is your channel position. The Power Law channel — bitcoin's structural floor, trend, and upper bound across fifteen years of price history — is the cleanest available frame for when to originate a loan. Borrowing when bitcoin is trading near the channel floor (a deeply oversold position relative to long-run trend) gives you the maximum drawdown buffer before liquidation. Borrowing near the upper bound puts you at high risk of being liquidated by routine mean-reversion, even if your long-run G > R thesis is intact.

This isn't market timing in the speculative sense. It's structural-risk timing — the same channel that the Disciplined Rebalancing page uses to identify high-percentile sell zones is what tells you, here, that originating a loan in a 90th-percentile zone is taking on additional liquidation risk you don't have to take. The calculator on the next tab makes this visible: it places your liquidation price directly on the Power Law channel chart, so you can see at a glance whether your loan is structurally exposed or structurally buffered.

The second is your LTV. Tier-1 CeFi platforms recommend 30–40% LTV as the conservative range; collaborative-custody multisig lenders typically operate in the same band or slightly more conservative. At 30% LTV, bitcoin can fall roughly 60% from origination before hitting a typical 80% liquidation threshold — enough buffer to survive even bitcoin's worst historical drawdowns. At 50% LTV, the buffer shrinks to roughly 35%, which has been breached multiple times in the past decade. The lower your LTV, the longer your strategy survives.

A useful self-check the calculator surfaces directly: calculate your liquidation BTC price under your proposed loan. If that price is below the Power Law floor, you have strong structural buffer — bitcoin would need to break the channel for you to be liquidated. If your liquidation price is above the trend, you're betting on continued strength rather than buying yourself room to be wrong. Both can be rational positions, but only one is conservative.

What HODL actually wins

For most readers, most of the time, the answer to “should I borrow against my bitcoin?” remains probably not.

HODL — unencumbered, self-custodied, zero-counterparty — captures the full Power Law upside without any of the structural risks this page has spent a lot of words detailing. There's no margin call. No liquidation event. No lender to fail. No oracle to manipulate. No wrapping bridge to exploit. No prospectus footnote to surprise you in year three. The bitcoin grows or it doesn't, and you continue to hold it either way.

The decision to borrow against your stack is a decision to trade some of that absolute safety for liquidity — and the trade is worth making only when both the use case is a genuine need and the structural conditions support survival. The calculator on the next tab lets you test, against your inputs, whether your specific proposed loan clears that bar.

What HODL gives up is liquidity — and for many readers, especially those facing genuine life events (surgery, education, a primary home, business capital), that liquidity matters in ways the “just HODL” answer can't address. The point of this page isn't to talk anyone out of borrowing. It's to give you the structural literacy to do it survivably, if you decide to do it at all — or to walk away from the strategy entirely, on reflection, knowing exactly what you're walking away from.

Both endings are good endings. The bad ending is the one where the reader sleepwalks in.