Onchain yields can fall even as Treasury returns rise. Patrick Heusser, Head of Risk at Sentora, says borrowing demand drives onchain lending rates, while token rewards and repeated borrowing can inflate quoted returns. Investors need to identify who pays the yield, what remains when incentives end and whether they can exit when liquidity tightens.

In this written interview with Korea IT Times, Heusser explains how to assess those risks and why he expects institutions to focus first on liquidity management and products for their customers.

 

The following Q&A is with Patrick Heusser, Head of Risk at Sentora.

Q1: Following the Fed’s September 15–16, 2026 meeting, when it raised its target range by 25 basis points to 3.75%–4.00%, how should investors expect a higher policy rate to affect stablecoin lending rates, onchain borrowing demand and the relative attractiveness of onchain yield compared with US Treasuries?

Patrick Heusser: “The Fed's September hike raised the return on Treasuries, which is the bar that onchain yield has to clear. It shouldn't, however, pull up the rate borrowers pay for stablecoins onchain.

On Aave, the lending rate rises and falls with how much of each pool is lent out. That makes it a function of borrowing demand, with no direct tie to the Fed's target. We track that rate daily through CDOR, the benchmark we built with CoinDesk Indices. CoinDesk Indices publishes CDOR daily using Aave V3’s USDC and USDT pools, with historical data going back to May 1, 2023. Our comparison of the day-to-day changes in CDOR and SOFR, the overnight rate that follows Fed policy, has shown almost no relationship between those changes.

The Fed's influence arrives slowly, as investors shift dollars between tokenized Treasury funds and lending pools toward whichever pays more. The risk for lenders is interconnected with the nature of borrowers. Most are traders using stablecoin loans to buy more bitcoin or fund leveraged trades, and their appetite tracks onchain markets.

If further hikes cool that appetite and borrowers repay their loans, onchain lending rates can drop in the same months Treasuries pay more.

Stablecoin issuers hold their reserves in cash and short-term Treasuries, so each hike raises their income, but Section 4(a)(11) of the GENIUS Act prohibits issuers from paying holders interest or yield solely for holding, using or retaining a payment stablecoin once the provision takes effect. That issuer prohibition does not directly cover exchanges, wallets and fintechs building separate products on top of stablecoins. The return that attracts new users to stablecoins reaches them through products and programs built on top of the stablecoin. Each hike makes it more valuable to be the bank, fintech or wallet offering them.”

Q2: Which components of onchain yield represent genuine credit, liquidity or basis risk, and which commonly depend on temporary token incentives or recursive leverage?

Patrick Heusser: “The first thing we do with any yield is ask where it comes from, because two products can show the same APY and carry different risk underneath. The part that pays for real risk is the lending itself. The money sits behind borrowers whose collateral can fall short in a sharp sell-off, and in pools that can lock up when too many lenders try to withdraw at once.

Basis trades belong in the same bucket, because the yield comes from what leveraged traders pay in the futures market, and that can shrink or turn negative when sentiment flips. The yield that disappears is whatever someone else is paying for. Token rewards and points last as long as the budget behind them lasts, and, according to Sentora’s own tracking, deposits on one chain fell 98.7% once its points program ended.

We run programs for stablecoin issuers ourselves, each built to hit a specific goal like deepening liquidity on a venue. We tell clients to treat that as distribution spend and take it out before comparing yields.

Looping is the other interesting caveat when investors borrow against a yield-bearing token, buy more of it, and repeat until a thin margin looks like a high yield. It works until borrowing costs more than the token earns, which is what happened with USDe on Aave in September 2026. Following LlamaRisk’s recommendation to raise the USDe base borrow rate from 5% to 6%, borrowing USDe cost more than staked USDe paid, so each additional round of the loop lost money.

Strip out the rewards and the loops. If what's left doesn't pay you more than Treasuries for the risk you're taking, the yield is coming out of someone else's budget.”

Q3: When digital-asset prices and borrowing demand decline, why do onchain lending rates sometimes fall even as traditional interest rates rise? Which strategies or structures tend to come under pressure first?

Patrick Heusser: “Onchain lending rates are set by how much of a pool is borrowed, rather than by the Fed setting interest rates. In a sell-off, borrowers repay or get liquidated while lenders mostly stay put, so utilisation falls, and the rate falls with it, even in a week when the Fed is hiking. Speculative borrowing leaves first, but borrowing against bitcoin is stickier, so rates drop but rarely collapse.

Basis products come under pressure first as funding compresses. Loops built on them follow, because the collateral earns less while the debt still costs something. Incentive-heavy vaults whose rewards are paid in a falling token come next. This means investors should focus on stress testing the yield, rather than the price.”

Q4: Which underwriting, collateral, liquidity and counterparty controls from traditional financial markets can be applied effectively to permissionless or semi-permissionless onchain infrastructure?

Patrick Heusser: “A permissionless pool can’t vet who is borrowing, so the lending rests on the collateral, with the loan-to-value ratio set by how much of that collateral could be sold in a stress. We view T+1 redemption as the bare minimum for any form of tokenized collateral being posted. That is Sentora’s own risk-management position, not a regulatory requirement. Margin management translates directly. Our xStocks strategy repays debt automatically when the stock sells off, cutting leverage before a liquidation event can do so.

On the other hand, semi-permissioned venues bring back counterparty controls, something which the institutions we are now engaging with expect from a sophisticated lending market. Because of this, we know almost every deployer in our vaults, and we support regulating curators for that exact reason.”

Q5: What metrics should an institutional investor examine before allocating to an onchain yield strategy, and what warning signs suggest that the quoted return may not be sustainable?

Patrick Heusser: "CDOR is where we start. It's the benchmark we built with CoinDesk Indices to track overnight borrow rates for USDC and USDT on Aave, so it tells you what lending dollars onchain pays before any extra risk is added. The gap between a vault's yield and CDOR is what you're being paid for that extra risk, and the work is finding out which risk it is.

That means knowing how much of the return comes from borrowers, how much from basis, and how much from token rewards or leverage inside the strategy. It also means looking at the market the strategy actually sits in, which is what our risk analytics engine, Risk Radar, was built for.

It tracks available liquidity, historical slippage, market depth and depositor concentration, among other factors, and those are what tell you whether you can exit at the size you entered. A pool with a few large depositors can be drained fast if one of them leaves.

The warning signs follow from that - a yield sitting well above CDOR with no source anyone can point to, a return paid mostly in the platform's own token, or growth that arrived on the back of a points campaign. Any curator should be able to show you where the yield comes from and what the exit looks like before you allocate."

Q6: Over the next 12 months, do you expect regulated institutions to treat onchain yield primarily as a credit product, a liquidity-management tool or an alternative to traditional fixed-income strategies?

Patrick Heusser: “For most of them, it will be liquidity management, because that is the only one of the three that requires no new mandate. In our view, a treasurer can hold a tokenized money market fund under the same policy that already governs cash. That is Sentora’s assessment, not a regulatory determination that applies to every institution. Lending onchain is a different exercise. It goes through the credit committee, and it gets underwritten the way any collateralised lending book would be, with collateral policy, counterparty limits and reporting that holds up in an audit. A few firms are working through that now, and the rest will follow once the first ones have proved it.

Another caveat to consider is that most regulated institutions aren't looking to allocate their own balance sheet here. Their customers are already earning yield onchain, and the question that follows is how those balances are retained and where they are reinvested. That is the work Sentora does, building and running the strategies that sit behind the products these platforms offer their own customers.”

About Patrick Heusser

Patrick Heusser is Head of Risk at Sentora, with more than 20 years of experience in global interest-rate markets as an investment banker, trader and portfolio manager. His expertise spans risk management, quantitative analytics and portfolio design across traditional finance and digital assets.

Read the Korean version