Executive summary
This report analyzes a recursive leverage strategy on entity[“cryptocurrency”,”Bitcoin”,”BTC”]: post BTC as collateral, borrow entity[“cryptocurrency”,”USD Coin”,”USDC”], acquire more BTC, re-pledge it, and repeat. The core balance-sheet math is valid, but two present-day constraints materially change the practical picture on current urlCoinbase docsturn18view1: the BTC-backed loan product is currently capped at $5 million, and Coinbase’s help center says loan proceeds cannot be used for trading on Coinbase. So the math below is a rigorous leverage model and a protocol-level analysis, but it is not the same thing as a currently documented, fully supported in-app “infinite loop” workflow on Coinbase. citeturn18view1turn18view0
The borrower-facing mechanics are clear in the reviewed docs. On Coinbase, BTC-backed loans are enabled through Morpho on Base; the collateral BTC is wrapped into cbBTC and locked in a Morpho smart contract; the loan has no due dates or minimum payments; LTV includes principal plus accrued and unpaid interest; liquidation occurs at 86% LTV; and Coinbase documents an additional 4.38% liquidation penalty. Morpho’s own docs explain that liquidation occurs when a position’s health factor falls to 1 or below, typically because collateral value falls or debt rises through accrued interest. citeturn8view1turn8view0turn8view2turn8view3turn9view2turn9view3
The loop is mathematically convergent, not truly infinite. If you start with collateral value C, current debt D, and target LTV L, then the additional amount you can recursively borrow-and-repost is X = (L*C - D)/(1 - L). If you start with 1 BTC and no debt, the fully looped exposure multiplier is 1/(1-L): at 50% LTV you reach 2.00 BTC of exposure, at 60% you reach 2.50 BTC, and at 67% you reach 3.03 BTC. The tradeoff is liquidation distance: with Coinbase’s 86% LLTV, the drawdown buffer is 1 - L/0.86, which is about 41.9% at 50%, 30.2% at 60%, and 22.1% at 67%. citeturn10view0turn8view1
For your personal baseline, modeled from your prompt as 22.89 BTC, $1,234,245 debt, and 86% liquidation LTV, the strategy is highly path-dependent. At $83k–$86k BTC, you are already around 63%–65% LTV, so a target of 50% or 60% is not reachable by borrowing more; it requires deleveraging. At 67%, each extra loop at those prices only adds roughly $117k–$257k of borrow capacity. At much higher BTC prices, the theoretical loop becomes enormous—but current Coinbase product limits bind far earlier, and the trading restriction means the direct in-app loop is not documented as supported. citeturn18view0turn18view1turn8view0
If the governing objective is “keep liquidation at or below $65,000 BTC,” then the correct control variable is not a fixed LTV. It is a price-dependent target LTV: L_target = 0.86 * 65,000 / BTC_price. That means the allowable LTV falls as BTC rises: about 67.35% at $83k, 65.00% at $86k, 11.18% at $500k, and 5.59% at $1M. This “floor-protection” framing is the right way to think about a long-horizon BTC-collateral strategy if liquidation price matters more than nominal borrowing capacity. citeturn8view1turn10view0
Mechanics on Coinbase and Morpho
According to urlCoinbase Help: Crypto-backed loans – Introductionturn18view1, eligible users can borrow USDC against crypto collateral through Morpho on Base. The rate is variable, tied to market supply and demand, and shown in the app. The same article lists a current $5M borrowing limit for BTC, states that a one-time processing fee is charged each time you borrow, says there are no minimum-payment requirements or due dates, and warns that you should monitor LTV to avoid liquidation. It also states, importantly, that Coinbase’s terms prohibit using loan proceeds for trading on Coinbase. citeturn18view1
The collateral path is also explicit. Coinbase’s borrower-facing docs say that when you borrow against BTC, the BTC is wrapped as cbBTC and locked on Morpho; when the loan is fully repaid, the collateral is returned automatically. The consumer-facing borrow page adds that you create a smart wallet, your BTC is withdrawn from the Coinbase account, converted to cbBTC, and transferred onchain into a Morpho smart contract. Coinbase’s wrapped-asset page says cbBTC is backed 1:1 by BTC held in custody by Coinbase. citeturn8view1turn8view3turn19view0
Morpho’s market model matters because it defines how the loan behaves once it is onchain. Morpho describes its variable-rate markets as isolated and immutable: one collateral asset, one loan asset, each market with fixed risk parameters such as an LLTV and an interest-rate model. Morpho also states that borrower collateral is specific to each market, is not cross-margined at the market level, and the collateral itself does not generate yield while posted. citeturn8view7turn8view9turn10view0
Interest accrual is a first-order risk input, not a footnote. Morpho says the borrow rate is dictated by the market’s interest rate model, primarily as a function of utilization, and explains that interest continuously increases borrower debt even if collateral value is flat. The onchain state is only updated when _accrueInterest is triggered by a market interaction, so a precise integration must account for borrow growth since lastUpdate. Coinbase’s loan health article is consistent with that framing: LTV includes accrued unpaid interest, and LTV rises when interest accrues, when you borrow more, or when collateral value falls. citeturn9view2turn8view0
Liquidation is also split across protocol and product language. Morpho says a position becomes liquidatable when LTV >= LLTV, or equivalently when health factor is <= 1; a liquidator can repay some or all debt and seize collateral at a bonus. Morpho’s generic docs say that for an 86% LLTV market, the liquidator incentive is about 5%. Coinbase’s borrower-facing help article, however, is what a Coinbase user should plan against: it states liquidation at 86% LTV and a 4.38% penalty. For a live Coinbase loan, the Coinbase docs should govern your operational assumptions. citeturn9view3turn9view1turn8view1
Coinbase also offers “loan protection,” which can auto-top-up collateral when LTV reaches a chosen trigger. But Coinbase explicitly warns it is not a guarantee against liquidation; the protection is one-time, expires after a year, and during system-wide drops prioritization may still leave some users unprotected. It is a backup aid, not a substitute for conservative LTV management. citeturn8view0
flowchart TD
A[BTC held on Coinbase] --> B[Borrow USDC through Coinbase]
B --> C[BTC converted to cbBTC and locked on Morpho]
B --> D[USDC credited to Coinbase balance]
D --> E{Is the trade leg supported by your venue and terms?}
E -- No / not documented --> F[Direct in-app loop stops here]
E -- Conceptual leverage model only --> G[Acquire more BTC]
G --> H[Post additional BTC as collateral]
H --> I{LTV below policy cap?}
I -- Yes --> B
I -- No --> J[Wait, repay, or top up]The flow above reflects the reviewed documentation: collateral wrapping and onchain custody are documented; the direct recursive “borrow then trade on Coinbase” leg is not, because Coinbase’s help center says loan proceeds cannot be used for trading on Coinbase. citeturn18view1turn8view3
Loop math
Morpho’s own example code shows the correct LTV and health-factor structure. In its worked example, current LTV is computed as borrowedAmount / collateralValue, and health factor is computed as (collateralValue * LLTV) / borrowedAmount. Coinbase’s help center matches that borrower-facing definition: LTV is total outstanding loan, including accrued interest, divided by the current market value of collateral. citeturn10view0turn8view0
Once you assume that every borrowed dollar is immediately used to acquire more BTC and that BTC is immediately re-posted, the recursive loop reduces to simple algebra:
- Start with collateral value
Cand debtD. - Borrow
X. - Buy
Xdollars of BTC. - Re-post that BTC as collateral.
Then final debt is D + X, and final collateral value is C + X. To end exactly at target LTV L:
L = (D + X) / (C + X)
Solve for X:
X = (L*C - D) / (1 - L)
That is the exact loop-capacity formula.
If you start from 1 BTC and zero debt, then C = P where P is BTC price, so:
X = L*P / (1 - L)
The purchased BTC is X/P, so final BTC exposure is:
1 + X/P = 1 + L/(1-L) = 1/(1-L)
That is the loop multiplier. It is the cleanest way to see why the loop converges:
- at 50%, multiplier =
2.0000x - at 60%, multiplier =
2.5000x - at 67%, multiplier =
3.0303x
Liquidation distance follows directly from the same identities. If your setup LTV is L and liquidation happens at LLTV, then the liquidation price as a fraction of entry price is:
P_liq / P_0 = L / LLTV
So:
P_liq = P_0 * L / LLTV
and the downside buffer is:
buffer = 1 - L/LLTV
With LLTV = 86%, the core tradeoffs are:
| Target LTV | Exposure multiplier | Liquidation at % of setup price | BTC downside buffer |
|---|---|---|---|
| 50% | 2.0000x | 58.14% | 41.86% |
| 60% | 2.5000x | 69.77% | 30.23% |
| 67% | 3.0303x | 77.91% | 22.09% |
A short worked example makes the algebra concrete. Suppose 1 BTC is worth $83,000 and you target 67%. Then X = 0.67*83,000 / 0.33 = $168,515, final BTC stack is 1 + 168,515/83,000 = 3.0303 BTC, final debt is $168,515, and liquidation price is 83,000 * 0.67 / 0.86 = $64,663. The loop adds a lot of upside exposure, but it leaves only about 22.1% room before liquidation. citeturn10view0turn8view1
Scenario analysis
Assumptions
The scenario tables below are author calculations using the formulas above and the following explicit assumptions: the personal baseline is the user-supplied case of 22.89 BTC collateral and $1,234,245 current debt; liquidation LTV is 86%; BTC and cbBTC are valued 1:1; each incremental borrow is assumed to be converted into BTC instantly at the scenario price and immediately re-posted as collateral; the core tables ignore new processing fees, slippage, trading fees, taxes, and intra-loop interest accrual; and where the target LTV is below the current LTV, additional borrowing is impossible, so the table shows the cash repayment required to reach the target while keeping BTC unchanged. Rows whose total debt exceeds Coinbase’s current $5M BTC loan limit are theoretical, not currently executable on Coinbase as documented. citeturn18view0turn8view1turn8view0
Conceptual 1 BTC case
Target LTV 50%
| BTC | Extra borrow | Final BTC | Final debt | Liq price |
|---|---|---|---|---|
| $83k | $83.0k | 2.0000 | $83.0k | $48.3k |
| $84k | $84.0k | 2.0000 | $84.0k | $48.8k |
| $85k | $85.0k | 2.0000 | $85.0k | $49.4k |
| $86k | $86.0k | 2.0000 | $86.0k | $50.0k |
| $500k | $500.0k | 2.0000 | $500.0k | $290.7k |
| $1M | $1.000M | 2.0000 | $1.000M | $581.4k |
| $10M | $10.000M | 2.0000 | $10.000M | $5.814M |
| $21M | $21.000M | 2.0000 | $21.000M | $12.209M |
| $55M | $55.000M | 2.0000 | $55.000M | $31.977M |
Target LTV 60%
| BTC | Extra borrow | Final BTC | Final debt | Liq price |
|---|---|---|---|---|
| $83k | $124.5k | 2.5000 | $124.5k | $57.9k |
| $84k | $126.0k | 2.5000 | $126.0k | $58.6k |
| $85k | $127.5k | 2.5000 | $127.5k | $59.3k |
| $86k | $129.0k | 2.5000 | $129.0k | $60.0k |
| $500k | $750.0k | 2.5000 | $750.0k | $348.8k |
| $1M | $1.500M | 2.5000 | $1.500M | $697.7k |
| $10M | $15.000M | 2.5000 | $15.000M | $6.977M |
| $21M | $31.500M | 2.5000 | $31.500M | $14.651M |
| $55M | $82.500M | 2.5000 | $82.500M | $38.372M |
Target LTV 67%
| BTC | Extra borrow | Final BTC | Final debt | Liq price |
|---|---|---|---|---|
| $83k | $168.5k | 3.0303 | $168.5k | $64.7k |
| $84k | $170.5k | 3.0303 | $170.5k | $65.4k |
| $85k | $172.6k | 3.0303 | $172.6k | $66.2k |
| $86k | $174.6k | 3.0303 | $174.6k | $67.0k |
| $500k | $1.015M | 3.0303 | $1.015M | $389.5k |
| $1M | $2.030M | 3.0303 | $2.030M | $779.1k |
| $10M | $20.303M | 3.0303 | $20.303M | $7.791M |
| $21M | $42.636M | 3.0303 | $42.636M | $16.360M |
| $55M | $111.667M | 3.0303 | $111.667M | $42.849M |
Personal baseline case
Your present baseline from the prompt is fundamentally different from the 1 BTC clean-sheet case because you already carry debt. That existing debt “uses up” part of the loop capacity. As BTC price rises, the drag from the pre-existing debt matters less, so the final BTC stack approaches the asymptotes implied by the loop multiplier: about 45.78 BTC at 50%, 57.23 BTC at 60%, and 69.36 BTC at 67%. Those are the high-price limits, not immediate numbers. The tables below incorporate your starting debt explicitly.
Target LTV 50%
| BTC | Current LTV | Extra borrow | Repay needed | Final BTC | Final debt | Liq price |
|---|---|---|---|---|---|---|
| $83k | 64.96% | — | $284.3k | 22.8900 | $949.9k | $48.3k |
| $84k | 64.19% | — | $272.9k | 22.8900 | $961.4k | $48.8k |
| $85k | 63.44% | — | $261.4k | 22.8900 | $972.8k | $49.4k |
| $86k | 62.70% | — | $250.0k | 22.8900 | $984.3k | $50.0k |
| $500k | 10.78% | $8.977M | — | 40.8430 | $10.211M | $290.7k |
| $1M | 5.39% | $20.422M | — | 43.3115 | $21.656M | $581.4k |
| $10M | 0.54% | $226.432M | — | 45.5332 | $227.666M | $5.814M |
| $21M | 0.26% | $478.222M | — | 45.6625 | $479.456M | $12.209M |
| $55M | 0.10% | $1.256B | — | 45.7351 | $1.258B | $31.977M |
Target LTV 60%
| BTC | Current LTV | Extra borrow | Repay needed | Final BTC | Final debt | Liq price |
|---|---|---|---|---|---|---|
| $83k | 64.96% | — | $94.3k | 22.8900 | $1.140M | $57.9k |
| $84k | 64.19% | — | $80.6k | 22.8900 | $1.154M | $58.6k |
| $85k | 63.44% | — | $66.9k | 22.8900 | $1.167M | $59.3k |
| $86k | 62.70% | — | $53.1k | 22.8900 | $1.181M | $60.0k |
| $500k | 10.78% | $14.082M | — | 51.0538 | $15.316M | $348.8k |
| $1M | 5.39% | $31.249M | — | 54.1394 | $32.484M | $697.7k |
| $10M | 0.54% | $340.264M | — | 56.9164 | $341.499M | $6.977M |
| $21M | 0.26% | $717.949M | — | 57.0781 | $719.184M | $14.651M |
| $55M | 0.10% | $1.885B | — | 57.1689 | $1.887B | $38.372M |
Target LTV 67%
| BTC | Current LTV | Extra borrow | Repay needed | Final BTC | Final debt | Liq price |
|---|---|---|---|---|---|---|
| $83k | 64.96% | $117.2k | — | 24.3018 | $1.351M | $64.7k |
| $84k | 64.19% | $163.6k | — | 24.8382 | $1.398M | $65.4k |
| $85k | 63.44% | $210.1k | — | 25.3620 | $1.444M | $66.2k |
| $86k | 62.70% | $256.6k | — | 25.8737 | $1.491M | $67.0k |
| $500k | 10.78% | $19.497M | — | 61.8834 | $20.731M | $389.5k |
| $1M | 5.39% | $42.734M | — | 65.6235 | $43.968M | $779.1k |
| $10M | 0.54% | $460.996M | — | 68.9896 | $462.230M | $7.791M |
| $21M | 0.26% | $972.206M | — | $69.1855 | $973.440M | $16.360M |
| $55M | 0.10% | $2.552B | — | 69.2956 | $2.554B | $42.849M |
The immediate takeaway is simple. In your baseline, 50% and 60% are deleveraging targets at $83k–$86k, not looping targets. Only 67% leaves room for extra borrowing in that range, and even then the added borrow is modest compared with the downside compression in liquidation price. At higher BTC prices, the theoretical loop becomes very large, but those rows quickly exceed Coinbase’s present $5M BTC loan cap. citeturn18view0
Floor-protection method
If your governing thesis is “liquidation should never rise above $65,000 BTC,” then the target LTV is not fixed. It must solve:
target LTV = 0.86 * 65,000 / BTC_price
That is the mathematically correct way to anchor the loan to a BTC floor rather than to an arbitrary constant LTV. Under that rule, the maximum allowable LTV falls as BTC rises. Because your current debt is already close to the floor-protected debt ceiling at very high BTC prices, additional borrow capacity under this method converges to only about $45k–$51k, even when BTC goes to extreme nominal prices. That is not a bug; it is exactly what floor protection means. citeturn8view1turn10view0
| BTC | Floor-protected LTV | Current LTV | Extra borrow | Final BTC | Final debt |
|---|---|---|---|---|---|
| $83k | 67.35% | 64.96% | $138.8k | 24.5618 | $1.373M |
| $84k | 66.55% | 64.19% | $135.4k | 24.5023 | $1.370M |
| $85k | 65.76% | 63.44% | $132.3k | 24.4469 | $1.367M |
| $86k | 65.00% | 62.70% | $129.4k | 24.3952 | $1.364M |
| $500k | 11.18% | 10.78% | $51.0k | 22.9920 | $1.285M |
| $1M | 5.59% | 5.39% | $48.0k | 22.9380 | $1.282M |
| $10M | 0.56% | 0.54% | $45.6k | 22.8946 | $1.280M |
| $21M | 0.27% | 0.26% | $45.4k | 22.8922 | $1.280M |
| $55M | 0.10% | 0.10% | $45.4k | 22.8908 | $1.280M |
xychart-beta
title "Floor-protected max LTV for a $65k liquidation line"
x-axis ["83k","84k","85k","86k","500k","1M","10M","21M","55M"]
y-axis "LTV %" 0 --> 70
line [67.35,66.55,65.76,65.00,11.18,5.59,0.56,0.27,0.10]xychart-beta
title "Personal current LTV by BTC price"
x-axis ["83k","84k","85k","86k","500k","1M","10M","21M","55M"]
y-axis "LTV %" 0 --> 70
line [64.96,64.19,63.44,62.70,10.78,5.39,0.54,0.26,0.10]The two charts make the controlling insight visual: if “$65k is the floor,” your allowable LTV must compress as BTC rises. Fixed-LTV thinking is the wrong frame for that objective.
Fees, interest, and present-day Coinbase feasibility
The clean loop formulas above are upper bounds because they ignore fees. Coinbase’s current processing-fee page says each new draw adds a one-time fee to principal, using a tiered schedule of 2% on the first $250,000 and 1% above $250,000; interest then accrues on the borrowed amount plus that fee. Coinbase instructs users to confirm the exact fee on the loan review screen before borrowing. citeturn7view0turn18view0turn18view1
If the incremental fee rate on a new draw is f, and you borrow gross cash amount B, you only get B to buy BTC but your debt rises by B*(1+f). The fee-adjusted loop-capacity formula becomes:
B = (L*C - D) / (1 + f - L)
For the zero-debt, 1 BTC case, the fee-adjusted exposure multiplier becomes:
multiplier_fee = (1 + f) / (1 + f - L)
So even a modest fee reduces the loop slightly. At 67%, the no-fee multiplier is 3.0303x; with a 1% fee it falls to about 2.9706x; with a 2% fee it falls to about 2.9143x. The fee does not destroy the strategy, but it absolutely moves the safe-borrow line inward.
Interest does the same thing over time. If borrow APR were constant at r and BTC price stayed flat, debt would grow roughly as D_t = D_0 * e^(r t) in a continuous approximation, so LTV would also drift upward at that rate. The time to hit 86% from starting LTV L_0 is approximately:
t_liq = ln(0.86 / L_0) / r
That produces the following “BTC flat” time-to-liquidation estimates:
| Start LTV | 5% APR | 10% APR | 15% APR |
|---|---|---|---|
| 50% | 10.85 years | 5.42 years | 3.62 years |
| 60% | 7.20 years | 3.60 years | 2.40 years |
| 67% | 4.99 years | 2.50 years | 1.66 years |
But if BTC appreciates, LTV can still fall despite interest. Over one year, a simple approximation is:
LTV_1y ≈ LTV_0 * (1 + borrow_apr) / (1 + BTC_return)
Starting from 67%, that means:
| Borrow APR | BTC flat | BTC +30% | BTC +60% |
|---|---|---|---|
| 5% | 70.35% | 54.12% | 43.97% |
| 10% | 73.70% | 56.69% | 46.06% |
| 15% | 77.05% | 59.27% | 48.16% |
This is why the strategy is seductive: if BTC compounds faster than debt, the position can “heal” over time. But the path still matters. A 30% or 60% annual BTC gain does not prevent liquidation if the path includes a fast 20% drawdown first. Morpho and Coinbase both explicitly warn that falling collateral value and rising accrued interest are enough to liquidate a position. citeturn9view2turn8view0turn9view3
The present Coinbase feasibility constraint is even harder than fees and APR: Coinbase’s help center says you cannot use loan proceeds for trading on Coinbase, and its loan-setup docs list a current $5M BTC-backed loan cap. That means the direct “borrow USDC on Coinbase, buy BTC on Coinbase, re-pledge it there, repeat” loop is not described as a supported workflow in the documentation I reviewed. It also means many of the large-number scenario rows are only balance-sheet theory under current Coinbase limits. citeturn18view1turn18view0
For the personal baseline specifically, the current Coinbase $5M cap begins to bind at approximately:
- $272.4k BTC if your target is 50%
- $199.5k BTC if your target is 60%
- $161.5k BTC if your target is 67%
Those breakpoints are derived directly from the cap and the loop formulas above. In other words, under today’s documented Coinbase cap, the “$500k, $1M, $10M…” rows are theoretical long before the math itself runs out.
Failure modes and historical stress lessons
The first failure mode is the obvious one: price-path risk. You can be directionally right on BTC over five years and still get liquidated next week if the path includes a violent drawdown. A recent illustration comes from urlChaos Labsturn17view0 and the related urlAave governance postturn17view1: a 10–20% market drawdown over seven days triggered about $202.47M of collateral seized against $193.12M of repaid debt, with one thin-liquidity collateral position still realizing a deficit. The lesson is not that liquidations fail every time; the lesson is that normal crypto volatility is already large enough to stress lending systems hard. citeturn17view0turn17view1
The second failure mode is the true liquidation cascade: too many positions approaching their thresholds at the same time. urlSpark’s liquidation-cascade case studyturn17view2 points to Black Thursday on March 12, 2020, when ETH fell about 43% in 24 hours, some Maker liquidations cleared at zero bids, and the protocol was left with around $4.5M in bad debt. The take-away for a BTC looper is that liquidation risk is not just about your collateral; it is also about blockspace, liquidator competition, and market depth during stress. citeturn17view2
The third failure mode is smart-contract exploit risk. Morpho’s own risk docs say there is always a non-zero possibility of a vulnerability or bug, even with audits, formal verification, and bug bounties. A broad DeFi reminder is the Euler incident: urlChainalysisturn17view3 describes the March 2023 exploit in which roughly $197M was stolen from Euler after a flash-loan-assisted attack. That event was not a liquidation event, but it is exactly why “protocol risk” must sit next to “market risk” in any looping strategy. citeturn17view4turn17view5turn17view3
The fourth failure mode is oracle and counterparty risk. Morpho warns that no oracle is immune to manipulation and that faulty oracles can cause liquidations or bad debt. It also warns that centralized governance over an asset can blacklist a user or even Morpho itself, creating asset-loss risk. Those warnings matter more, not less, when your collateral is a wrapped or centralized representation rather than native BTC living on the Bitcoin chain. citeturn20view0
The fifth failure mode is venue/product risk. Coinbase itself flags crypto-price volatility, protocol liquidity issues, protocol security vulnerabilities, and smart-wallet issues as loan risks. It also warns that Coinbase cannot stop Morpho from liquidating your collateral once parameters are breached. Operationally, that means there is no magical “customer-service override” if you let the position drift too close to the line. citeturn18view1turn8view3
Operating rules and live checklist
Numbered rules
- Treat liquidation price, not nominal LTV, as the primary risk variable.
If you have a hard BTC floor thesis, computeL_target = 0.86 * floor / priceand manage to that. A fixed 67% policy only makes sense near the price level at which it corresponds to your acceptable liquidation floor. - Do not confuse “no due dates” with “no urgency.”
Coinbase gives you flexibility on repayment timing, but interest still accrues continuously enough to raise LTV, and liquidation can still occur with no monthly invoice warning. citeturn8view2turn8view0turn9view2 - Use 67% as a redline, not a cruising altitude.
The clean math says 67% gives 3.03x exposure, but it also leaves only ~22.1% downside to liquidation. For most live management, 50%–60% is the more robust zone. - Only expand the loop when your current LTV is materially below your policy cap.
In your baseline at $83k–$86k BTC, you are already above 50% and 60%, so those are deleveraging targets, not borrowing targets. - Keep external repayment liquidity.
The strategy is not “never repay.” It is “never be forced to repay by selling BTC at the wrong time.” Maintain cash or USDC that can reduce debt quickly if needed. A simple external-cash repayment formula is:Repay needed = Debt - target_LTV * collateral_value - Precompute collateral top-up amounts.
If you prefer not to repay cash, a BTC top-up target is:Additional BTC needed = Debt / (target_LTV * BTC_price) - current_BTC - Use loan protection only as a secondary defense.
It can help, but Coinbase explicitly says it is not guaranteed, can fail in technical stress, is one-time, and expires after one year. citeturn8view0 - Assume fees and APR make every theoretical number a little worse.
Processing fees are added to principal and then earn interest against you. Every fresh draw slightly raises the “true” LTV relative to no-fee math. citeturn7view0turn18view1 - Respect product constraints.
Under the current docs, Coinbase caps BTC-backed borrowing at $5M and says loan proceeds cannot be used for trading on Coinbase. Do not build a risk plan around a workflow the venue does not document as supported. citeturn18view0turn18view1 - Monitor more often as LTV rises.
A practical cadence:- below 50%: daily is usually enough
- 50% to 60%: at least daily plus BTC price alerts
- 60% to 67%: intraday checks and ready cash
- above 67%: active defense mode
Live checklist
Keep the following six numbers visible at all times:
- Current debt, including accrued interest
- Current collateral value
- Current LTV
- Current liquidation price
- Cash repayment needed to get back to your target LTV
- BTC top-up needed to get back to your target LTV
Then follow a simple decision tree:
- If BTC rises and your liquidation price stays comfortably below your policy floor, do nothing first; let the position heal.
- If BTC rises enough that your LTV falls well below the cap, decide whether a new loop still fits product limits, fees, and venue rules.
- If BTC falls and liquidation price approaches your redline, repay or top up before the protocol reaches 86%.
- If you are relying on Coinbase loan protection, make sure the top-up asset is already sitting in balance; otherwise the feature may partially fail or not fire at all. citeturn8view0
The central conclusion is straightforward. The recursive BTC-collateral loop is mathematically elegant and can be balance-sheet efficient if BTC outgrows debt costs. But on the current reviewed Coinbase documentation, it is not a frictionless in-app “infinite loop”: it is constrained by product limits, fees, interest, liquidation math, protocol risk, and—in a very practical sense—by Coinbase’s stated prohibition on using loan proceeds for trading on Coinbase. The strategy only works if you survive the path, not merely if you are right about the destination. citeturn18view1turn18view0turn8view1turn9view2turn20view0