Bitcoin credit — the financing layer
How much Bitcoin is posted as loan collateral across the venues we track, how much is borrowed against it, and the leverage between the two — read daily from contract state, every figure a number a contract reported at a block recorded with it. Below that, what the loans themselves do: how much is borrowed and repaid, how long borrowings last, and how often they are rolled over. Those come from one venue rather than all of them, and the page says so where they are.
BTC-USD · contract state at 2026-08-05
Bitcoin collateral locked · bcl
131,104 BTC
$8.40bn
Collateral posted at venues that expose an aggregate figure. Morpho reports collateral per position rather than per market, so its markets are absent from this total — they carry debt below but no collateral here.
Bitcoin-collateralised debt · bcd
$2.04bn
measured directly
+ $492.8m
upper bound, venue-wide
The bound is not a smaller measurement — it is a market-wide total from venues that pool collateral, where the Bitcoin share cannot be isolated until loan-level reconstruction lands. The two must not be added.
Aggregate loan-to-collateral · bclv
56.27%
exact-attribution subset only
Computed only where a venue reports both an exact debt figure and its collateral — today that is Sky alone. Dividing a market-wide debt total by a Bitcoin-specific collateral total would not be a leverage ratio, so it is not done.
What is behind these numbers. 15 of 15 venues with quantity coverage reported on 2026-08-05 (100%). Of 320 market readings, 257 yield Bitcoin-attributed debt directly, 18 yield only a venue-wide bound, and 45 yield no debt figure at all — pooled-collateral venues where the question is per-account and cannot be answered before W3. A further 74 readings carry debt in assets we do not value (AERO, AUSD, BOLD, CANA, EURA, eUSD, and 26 more), and are excluded rather than estimated.
How the debt is valued. $2.54bn of it is denominated in stablecoins, valued at par to the currency each is pegged to — a documented convention, not a market reading. No tracked stablecoin has broken its peg for a sustained period. Par is applied without a price feed for EURC, EURCV, EURe, EUROP, jEUR, USD₮0, USDT0 — the convention holds for them but nothing is checking it.
Not in the figures above. Some borrowing against Bitcoin collateral is denominated in something other than dollars. It belongs to a different curve, so it is measured and reported but never added to the totals above: $68.8m in ETH (37,022 WETH), $23.4m in euro stablecoins, $3.0m in Bitcoin itself.
Which venues are these? Every venue we have identified that lends against Bitcoin collateral — on-chain markets, CeFi desks, derivative-implied rates and corporate debt — with what is observable at each and what is not. See the coverage matrix →
The loan-level layer — Morpho Blue
These six are one venue, not the market. These series are computed from Morpho Blue only, on Ethereum and Base. They are not market-wide. Aave v3 is mid-backfill and Compound v3 and Sky are not yet ingested, so any cross-venue reading — including the shape question this programme calls Checkpoint B — remains open.
Origination flow · bco
$263.9m
last 30 days
$9.14bn
since 2024-01-19
Gross new borrowing. Net flow is this minus repayment, never a single signed series — a day of heavy churn and a quiet day net to the same number and are not the same market.
Repayment flow · bcr
$127.2m
last 30 days
net $136.7m
over the same window
Gross repayment, including interest. The difference between the two flows is the net change in principal, and it reconciles against contract state to under one basis point.
Liquidation volume · blq
7,484 BTC
$568.4m
settling $532.3m of debt, since 2024-01-19
Collateral taken by liquidators, in Bitcoin because Bitcoin is what moved. Seizures of yield-bearing wrappers are excluded from the Bitcoin total — one unit of those is a claim on Bitcoin, not a Bitcoin.
Realised tenor distribution · rtd
85 days
median, survival-adjusted · p75 666 days
11 days
if unfinished loans are dropped — 42% of spells have not ended
How long a borrowing lasts, with positions that have not ended right-censored rather than dropped. The closed-only figure is shown beside it because dropping unfinished loans biases the answer short — and short is the answer that would suit a fixed-term venue.
Rollover rate · rrr
30.73%
within 7 days, same market
10.5h
median gap before re-borrowing
A repaid borrowing followed by a new one from the same account in the same market within seven days. Read the median gap with the rate: much of it closes within hours, which is position management as much as a decision to borrow again.
Liquidation incidence · lir
7.46%
across 61,428 closed borrowings
LTV 0-40%2.05%
LTV 40-55%7.89%
LTV 55-65%7.65%
LTV 65-75%8.66%
LTV 75-85%10.08%
LTV 85+%34.84%
Liquidated spells as a share of closed ones, by the loan-to-value at which the borrowing was opened. Cohorts are computed only where one unit of collateral is one Bitcoin by construction.
What is behind these numbers. 95.29% of borrow, repayment and liquidation events carry a valuation. 30,247 are denominated in euro, ether or Bitcoin itself and belong to other curves; a further 5,885 are in assets with no declared parity and are excluded rather than assumed to be dollars. 4,458 closed borrowings carry no origination LTV — mostly collateral held as a yield-bearing wrapper, where one unit is a claim on Bitcoin rather than a Bitcoin — and are reported as their own cohort rather than estimated into one.
Why two tenor figures. The longest borrowing still open (918 days) has already outlasted the longest one ever repaid (824 days). Any statistic built only from finished loans cannot see it — not because it is extreme, but because it has not finished. That is the whole reason the survival-adjusted figure and the closed-only figure are both published, and why the second is the one to distrust.
What did it cost? The same loan-level history carries the rate every borrower actually paid. Grouped by how long each position was actually held, it makes a term structure — and replayed week by week, it shows how that structure moved since January 2024. It is a realised holding period, not a maturity: nothing here is a rate anyone can lock. See the term structure →
Not yet published
The financing layer is specified to carry 11 series. 9 are above. The two below are not held up by data we lack about Morpho — both are statements ABOUT the spread across venues, so neither can be computed from one venue however complete that venue is. They are listed with what gates them, so that what this surface does not yet show is as visible as what it does.
Venue concentration · vci
NOT YET PUBLISHED — W4
HHI over outstanding debt. Needs debt attributed per venue rather than bounded — Morpho alone cannot produce a concentration measure, since concentration is a statement about the other venues.
Collateral mix share · cms
NOT YET PUBLISHED — W4
WBTC / cbBTC / tBTC / LBTC split by venue. Morpho's share is computable today; the series is the split ACROSS venues, so it waits for Aave, Compound and Sky.
Contract state is read once a day and cannot be reconstructed backwards, so this series begins on the day it was first taken rather than reaching into the past. Every reading records the block it was taken at, and re-running that read at that block returns the same answer. Bitcoin valued at $64,075 for this reading.