The State of Crypto Leverage – Q2 2026: An Orderly, Measured Decline

Introduction

Q2 was the first quarter since Q4 2022 in which onchain lending declined across every category (CeFi, DeFi, and the crypto-collateralized portion of collateral debt position stablecoins), as the market’s deleveraging trend continued. The notable difference between current conditions and those of the previous bear cycle is outstanding loans are falling in a steady, stepwise decline rather than in outright collapse. In Q2 2022, the crypto-backed lending sector caved by more than 55% before experiencing additional 9% and 29% declines in Q3 and Q4 2022, respectively. Compare that to the recent deleveraging cycle where the market has seen three consecutive quarters of just 10%, 5%, and 17% declines. This measured pace points to a much healthier deleveraging cycle, driven by gradual risk reduction rather than forced liquidations or counterparty failures, in our view. Should lending activity continue to contract in the coming quarters, we'd expect it to follow this same stepwise pattern rather than the sharp, cascading losses that defined the 2022 unwind.

On the corporate treasury front, we saw some deleveraging driven primarily by a May 2026 debt repurchase of $1.5 billion by Strategy. This took the amount of debt used to supplement digital asset treasury strategies down to $16.1 billion, which is the approximate level of debt taken on by these companies in July 2025.

In the futures market, open interest (OI) ended the quarter effectively flat, declining 3.08% quarter-over-quarter to $103.2 billion. The modest headline decline masked more pronounced moves beneath the surface with BTC open interest slipping 6.24% to $45.04 billion and ETH open interest falling more sharply, by 26.31%, to $21.99 billion. Together, BTC and ETH accounted for 65% of total futures OI at quarter end. Notably, this relative stability in the futures market did not persist as OI climbed back to roughly $114 billion by the end of July, with both BTC (~$48 billion) and ETH ($25.74 billion) open interest rebounding from their Q2 lows.

Key Takeaways

• All told, crypto-collateralized lending contracted by $11.33 billion (-16.78%) in Q2 2026 to $56.16 billion. This is 40.13% lower than the Q3 2025 high of $78.69 billion.

• The dollar-denominated value of outstanding loans on DeFi lending apps fell for the third consecutive quarter in Q2, contracting $7.79 billion (-27.61%) to $20.43 billion.

• Galaxy Research is tracking $16.1 billion in debt outstanding used to directly buy or supplement the treasury strategies of firms using debt to hold digital assets.

• Futures open interest (OI), including perpetual futures (perps), declined 3.08% QoQ to $103.2 billion.

Crypto-Collateralized Lending

The market map below highlights some of the major past and present players in the CeFi and DeFi crypto lending markets. Some of the largest CeFi lenders by loan book size crumbled in 2022 and 2023 as crypto asset prices tanked and liquidity dried up. These lenders are flagged with red caution dots in the map below.

CeFi

The table below compares the CeFi crypto lenders in our market analysis. Some of the companies offer multiple services to investors. Coinbase, for example, primarily operates as an exchange but also extends credit to investors through over-the-counter cryptocurrency loans and margin financing. The analysis shows only the size of companies’ crypto-collateralized loan books, however.

As of June 30, Galaxy Research tracked $22.98 billion of open CeFi borrows. This represents quarter-over-quarter (QoQ) contraction of 9.62%, or $2.45 billion, and $16.14 billion (+235.94%) growth since the bear market trough of $6.8 billion in Q4 2023. Still, CeFi borrows outstanding are 37.16% below their Q1 2022 all-time high of $36.58 billion.

The second quarter saw CeFi books compress in aggregate. However, it was mostly driven by a reduction in Tether’s outstanding secured loans. Galaxy, Coinbase, Ledn, Arch, Sygnum, and Milo all saw their books grow during the second quarter.

Tether remains the dominant lender in our analysis, commanding a 58.54% share (down 371 basis points from last quarter) of the CeFi lending market. Add in Maple (8.91% market share, up 52 basis points from the last quarter), and Nexo (7.51% market share, up 49 basis points), and the top three tracked CeFi lenders control 74.96% of the market (down 270 basis points).

When comparing market shares, it’s important to note the distinctions between CeFi lenders. Some lenders only offer certain types of loans (e.g., BTC-collateralized only, altcoin-collateralized products, or cash loans disbursed in fiat rather than stablecoins), only service certain types of clients (e.g., institutional vs. retail), and only operate in certain jurisdictions. The combination of these factors allows some lenders to scale more easily than others.

The table below details the sources of Galaxy Research’s data about each CeFi lender and the logic we used to calculate the size of their books. While DeFi and onchain CeFi lending figures are retrievable from onchain data, which is transparent and easily accessible, retrieving CeFi data is tricky. This is due to inconsistencies in how CeFi lenders account for their outstanding loans and how often they make the information public, as well as the general difficulty of obtaining this information.

Note: the values supplied by private third-party lenders have not been formally vetted by Galaxy Research.

CeFi and DeFi Lending

The dollar-denominated value of outstanding loans on DeFi lending apps fell for the third consecutive quarter in Q2, contracting by $7.79 billion (27.61%) to $20.43 billion. Combining DeFi apps with CeFi lending venues, there were $43.41 billion of outstanding crypto-collateralized borrows at quarter end. This represents a reduction of $10.24 billion (19.08%) QoQ, predominantly driven by compression of onchain borrows. Notably, this is the first quarter since Q3 2023 when CeFi loans outstanding eclipsed that of DeFi lending apps.

Note: There is potential for double-counting between total CeFi loan book size and DeFi borrows. This is because some CeFi entities rely on DeFi applications to lend to offchain clients. For example, a hypothetical CeFi lender may pledge its idle BTC to borrow USDC onchain, then lend that USDC to a borrower offchain. In this scenario, the CeFi lender’s onchain borrow will be present in the DeFi open borrows and in the lender’s financial statements as an outstanding loan to its client. The lack of disclosures or onchain attribution makes filtering for this dynamic difficult.

As a result of the quarter-over-quarter decline in outstanding borrows on DeFi lending applications outpacing that of CeFi venues, DeFi’s lead over CeFi lending disappeared in Q2. At the end of Q2 2026, DeFi lending app share fell to 47.05%, down 555 basis points QoQ and down from 52.6% at the end of Q1 2026.

The third leg of the stool, the crypto-collateralized portions of collateral debt position (CDP) stablecoin supply, decreased by $1.09 billion (7.86%) QoQ. Again, there is potential for double-counting between total CeFi loan book size and CDP stablecoin supply, because some CeFi entities might rely on minting CDP stablecoins with crypto collateral to fund loans to offchain clients.

All told, crypto-collateralized lending contracted by $11.33 billion (16.78%) in Q2 2026 to $56.16 billion. This is 40.13% lower than the Q3 2025 high of $78.69 billion.

At the end of Q2 2026, DeFi lending applications represented 36.37% (-544 basis points from Q1 2026 share) of the crypto-collateralized lending market, CeFi venues captured 40.93% (+324 basis points) of the market, and the crypto-collateralized portion of CDP stablecoin supplies held 22.7% (+220 basis points). Combining DeFi lending apps and CDP stablecoins, onchain lending venues held a 59.07% dominance (-324 basis points from Q1 2026) over the market.

Additional Views of DeFi Lending

Outstanding borrows on DeFi lending applications have significantly shrunk since reaching an all-time high of $47.13 billion on Sept. 19, 2025. Sitting at $21.94 billion as of July 21, 2026, onchain lending has collapsed by $25.19 billion, or 53.45%.

The current drawdown in DeFi borrows has continued to intensify since the end of Q1 2026 but recently started to abate slightly.

Stablecoins

The weighted average stablecoin borrow rate increased over the quarter, rising 27 basis points between March 31, and June 30 using the seven-day moving average. After quarter-end, stablecoin rates continued to climb to 3.88%.

This figure is calculated by blending the costs of borrowing from lending protocols and CDP stablecoin mint fees, weighted by outstanding borrows.

The following breaks out the costs of borrowing stablecoins through lending applications and minting CDP stablecoins with crypto collateral. The two rates track each other closely, although CDP stablecoin mint rates are typically less volatile because they are manually set periodically and do not move in lockstep with the market. Both rates have used the Fed Funds Rate as a floor for the last 21-plus months.

Benchmark over-the-counter (OTC) interest rates for USDC fluctuated between 4.25% and 5% throughout the quarter. As of quarter-end they stood at 4.25% and remained there through Aug. 3.

The chart below tracks the same rates as above, but for USDT lending. Like USDC OTC rates, those of USDT fluctuated between 4.25% and 5%.

Bitcoin

The chart below shows the weighted borrow rate for wrapped bitcoin (WBTC) on lending apps across several applications and chains. The cost of borrowing WBTC onchain is often low because wrapped bitcoin tokens are primarily used for collateral in onchain markets and are not in high demand for borrowing. In contrast to stablecoins, the cost of borrowing BTC onchain remains stable, because users borrow and repay it less frequently. The rate to borrow BTC onchain fluctuated between 0.44% and 0.5% during the second quarter.

The historical divergence between onchain and offchain (over-the-counter) borrow rates for BTC persisted throughout the second quarter. In the OTC market, BTC borrowing demand is driven primarily by two factors: 1) the need to short BTC and 2) the use of BTC as collateral for stablecoin and cash loans. The former is a source of demand not commonly found in onchain lending markets, hence the spread between onchain and over-the-counter BTC borrow costs.

OTC rates for BTC remained flat throughout the quarter at 1%.

ETH and stETH

The chart below shows the weighted borrow rate for ETH and stETH (staked ether on the Lido protocol) across several lending applications and chains. The cost to borrow ETH has historically been higher than for stETH because the former is in greater demand. Users repeatedly borrow ETH to fuel looping strategies to gain leveraged exposure to the Ethereum network staking APY, pledging stETH, the token they receive for staking ETH through Lido, as collateral. As a consequence, the cost of borrowing ETH fluctuates within 50 basis points of the Ethereum network staking APY on average under normal conditions. This strategy becomes uneconomical when the cost to borrow exceeds the staking yield, so it is uncommon for the borrow APR to clear the staking APY for extended periods of time.

As with WBTC, the cost of borrowing stETH is often low because the asset is primarily used as collateral.

Pledging liquid staking tokens (LSTs) or liquid restaking tokens (LRTs) (both of which earn yield) as collateral, users borrow ETH at low, often negative, net interest rates. This cost efficiency fuels a looping strategy where users repeatedly deposit LSTs and LRTs as collateral to borrow unstaked ETH, stake it, and then recycle the fresh LSTs and LRTs to borrow even more ETH, thereby amplifying their exposure to the ETH staking APY. This strategy only works so long as the borrow cost for ETH is below the staking APY achieved on the LSTs and LRTs. Users have mostly been able to conduct this strategy without a hitch outside of a few notable periods.

ETH Over-the-Counter Rates

As with bitcoin, borrowing ETH through onchain lending apps has been historically cheaper than borrowing it over the counter. This is driven by two factors. First, as with BTC, there is borrowing demand from short sellers through offchain venues that is not as common onchain. Also, the Ethereum staking APY serves as a floor rate for offchain borrowing because there is little incentive for suppliers to deposit assets at offchain venues, or for offchain venues to lend assets out, at rates less than staking would earn. By contrast, onchain, the staking APY is often the ceiling rate for lending ETH.

Aave Book View

The following is a detailed, filtered view of the Aave protocol’s book on its V3 Core instance (the largest onchain lending market). The filters applied:

• Minimum debt ($100): We excluded positions below this threshold from the main aggregates. That removes dust but skews summaries toward larger loans.

• Health factor (HF) reporting cap (HF <= 50): We also excluded positions with snapshot health factors above 50 from the headline cohorts; these heavily overcollateralized loans tends to be small and scarcely relevant to a risk analysis. Among loans that fall below the cap, debt-weighted HF averages and HF percentile summaries count only loans with snapshot health factors greater than or equal to 1 and less than or equal to the cap of 50. (Loans with health factors below 1 are omitted from these HF readouts; any loan where the health factor equals 1 is included.) The higher the health factor, the safer the loan; a health factor below 1 means the position is eligible for liquidation. The health factor is (total collateral value * weighted average liquidation threshold) / total borrow value.

• Debt-to-equity (D/E): We computed this ratio only for loans where collateral exceeds debt (positive equity). We omitted positions without positive equity from D/E distributions and from the debt-weighted average D/E. That rule is separate from the minimum-debt filter: a loan can clear the $100 threshold and still be excluded from D/E-only aggregates when equity is not positive.

As of a snapshot taken on Aug. 7, 2026, there were 19,073 loans open net of the above filters. While the number of “efficiency mode,” or e-mode, Aave loans only made up 8.91% of total open positions, the outstanding debt was split roughly 50/50 between e-mode and vanilla loans. (In e-mode loans, the borrowed assets and the collateral are correlated in price, e.g. ETH and WETH.) This split has reduced since Galaxy Research last covered Aave’s book when it was closer to 60/40 on April 22, 2026. The difference is due to a reduction in outstanding e-mode debt.

This table below summarizes debt-weighted risk metrics for the filtered book, split into all positions, e-mode, and vanilla. E-mode borrowers carry much higher leverage, on average. Their debt-weighted LTV is about 90% with a debt-weighted health factor near 1.06 and debt-to-equity around 10.7, meaning a small collateral shock can move many of these loans toward stress. Non-e-mode loans look far more cushioned, with a debt-weighted LTV near 49%, HF near 1.79, and D/E near 1.07 (compensating for the risk of uncorrelated borrow and collateral assets, e.g. using cbBTC to borrow USDC).

Debt-weighted average is calculated as D/E = Σi (Di × (Di ÷ (Ci − Di))) ÷ Σi Di over loans with Ci > Di, where Di is debt and Ci is collateral for individual position i.

The next table ranks all enabled collateral lines across Aave V3 Core as a cumulative market by how much USD value sits in each symbol across the cohort. ETH-linked collateral dominates: WETH, wrapped Etherfi restaked ETH (weETH), and wrapped Lido stETH (wstETH) each account for a large share of the total (roughly 24%, 16%, and 14%, respectively, for a combined share of 54.6% of enabled collateral), with WBTC also material (~14%). So, a handful of tickers carry most of the book’s posted collateral. Below that tail there are smaller but still meaningful slices in stablecoins and other yield tokens.

The following table ranks borrowed assets by symbol in USD and as a share of the cohort’s total borrowing. WETH dominates liabilities at a little over 37%, which is typical when we consider the meaningful presence of leveraged looping strategies where ETH-correlated assets serve as collateral. Stablecoin borrows are also large: USDT and USDC together make up half of the total (~28% and ~22%, respectively), and everything below that is comparatively small.

Since our last analysis of the Aave V3 Core book, WETH’s share of outstanding liabilities has decreased meaningfully from 51.1% of the total. This coincides with the reduction in e-mode loans noted above.

E-mode View

Within e-mode loans, enabled collateral is heavily skewed toward ETH staking/restaking wrappers: weETH (not to be confused with WETH) alone is ~42% of the category, and combined with rsETH and wstETH, this asset type covers roughly 66.2% of e-mode posted collateral. As a result, e-mode risk is less “diversified collateral” and more a concentrated bet on Ethereum’s staking basis.

For e-mode borrows, the book is overwhelmingly WETH‑denominated. WETH alone is ~73% of e‑mode debt, which is exactly what you would expect when users are looping ETH-correlated collateral against ETH debt. Stablecoins still show up meaningfully. USDT, USDe, and USDC together are on the order of low‑teens percentages of e‑mode borrows.

The following table ranks e-mode positions that touch each collateral ticker (with debt-weighted risk and, where applicable, implied loop counts on the ≥99% single-asset collateral subcohort). As a result, it reads like a “which asset is being looped hardest?” scoreboard, unlike the neutral market-cap tables above. Liquid restaking/staking ETH wrappers cluster at the top with high debt-weighted LTVs, D/E often in the high single digits to low teens, and health factors not far above 1. This is consistent with tight, repeated ETH-beta loops.

The implied number of loops is calculated on a per-loan basis using the following formula: Per-loan N_i = ln((1 − (D/E)_i(1 − L_i)) / L_i) / ln(L_i), using each loan's own current LTV L_i and debt/equity, over the ≥99% single-symbol-concentration subcohort. The reported value is the debt-weighted average of N_i. Loans with L_i ∉ (0,1), a non-positive log argument (leverage above the L_i/(1−L_i) loop ceiling), or a non-finite result are dropped. The values are only printed for liquid-staking-ETH, liquid-restaking-ETH, yield-bearing-stable, or Pendle PT symbols when at least one loan survives.

Corporate Debt Strategies

We are now tracking $16.1 billion in debt outstanding used to directly buy or supplement the treasury strategies of firms using debt to hold digital assets. Due to constraints in Bloomberg’s tracking of Strategy preferred stock, the timing of the increases in outstanding STRC share issued by the company looks off in the time series. However, the total debt issued is representative of outstanding liabilities.

The outstanding debt issued by digital asset treasury companies (DATs) declined by $1.5 billion over the quarter as Strategy completed a $1.5 billion debt repurchase in May.

The following details the effective quarterly interest due on DAT-issued debt. Note that Strategy’s STRC dividends are payable only when declared by the board out of legally available funds, though any unpaid dividends accumulate and must be satisfied before distributions to junior securities. As a result, STRC dividend payments may be made unevenly and at non-fixed intervals.

Total crypto-related debt outstanding, including the debt taken on by DATs, declined 15.08% quarter-over-quarter. After reaching an all-time high in Q3 2025, total outstanding debt through onchain and offchain channels ended Q2 at $73.2 billion and saw its third consecutive quarter of decline.

Futures Market

Futures open interest (OI), including perpetual futures (perps), declined 3.08% QoQ to $103.2 billion. In July, open interest trended higher, ending the month around $114 billion.

It’s important to note that the entirety of the futures open interest figure does not constitute an absolute amount of leverage. This is due to the fact that some portion of the open interest figure can be offset by long spot positions, giving traders delta-neutral exposure to the underlying asset, and the total leverage ratio of the market is not directly observable from open interest alone.

BTC futures open interest was rangebound between $44 billion and $62 billion through the second quarter. Starting the quarter at $48.04 billion, BTC-based open interest declined 6.24% to 45.04 billion by June 30. Since then, BTC open interest trended slightly higher back to $48 billion by the beginning of August.

ETH open interest experienced a larger decline than that of BTC in Q2. Opening the quarter at $29.84 billion in OI, ETH held just $21.99 billion in open interest as of June 30 (representing a 26.31% decline). Since the end of the quarter, however, ETH OI rebounded to $25.74 billion.

Taken together, BTC-based and ETH-based open interest totaled 65% of the futures market at $67.07 billion at the end of Q2.

Conclusion

In our view, Q2 2026 provided further confirmation that leverage in the crypto market is being worked off gradually after the sharp decline in the futures market on Oct. 10, 2025. Lending markets are taking the stairs down, not the elevator, with three consecutive quarters of contraction in measured steps rather than a massive single quarter decline like the one that defined the 2022 bear market. Corporate treasury debt and futures open interest told a similar story of controlled retracement rather than forced deleveraging, even as early July data hinted that OI and DeFi borrows may already be finding a floor.

Should this trend continue, the market looks better positioned to absorb further contraction without the cascading liquidations and counterparty failures that scarred the last cycle. For now, the descent continues one step at a time.