Maple Finance (SYRUP): institutional crypto loans behind syrupUSDC, and what makes them risky

Maple Finance lends stablecoins to crypto institutions against collateral, and the people who hold syrupUSDC or syrupUSDT fund those loans. This week’s attention is on XRP. On 31 October 2026 CryptoSlate reported that about 225 million XRP, which it valued at $280.2 million, backs about $151.8 million of Maple loans maturing on 4 December, 12 December and 15 January 2027, under a headline saying those dates threaten liquidations.

The loans exist: Maple’s own syrupUSDC and syrupUSDT disclosure pages list all three. The headline goes further than the facts. A maturity is the day a loan must be repaid. It is not in itself a liquidation trigger, and Maple’s default procedure says what is.

The same gap between the top line and the mechanism shows up elsewhere. Maple reported $4.7 billion of assets under management and $1.7 billion of loans outstanding at the end of the third quarter of 2026, yet its own pages show lenders’ deposits of $1.87 billion, credit decisions are made off-chain by Maple’s team, withdrawals can take up to 30 days, and the SYRUP token carries no described legal claim on revenue. We rate the project 3 out of 5 for risk.

What Maple Finance is: a credit desk with an on-chain wrapper

Sidney Powell, the chief executive, and Joe Flanagan, the executive chairman, are the co-founders listed on Maple’s about page, and the company dates itself to 2019. Its documentation describes yield coming primarily from fixed-rate, overcollateralised loans to institutional borrowers, supported by DeFi liquidity provision and futures basis trading.

There are two product lines. Maple Institutional offers permissioned pools for institutions. The syrup line opens the same kind of credit to a wider audience through three vault tokens: syrupUSDC, syrupUSDT and syrupUSDG. Deposit a stablecoin, receive a share, and the share’s value rises as borrowers pay interest.

These tokens are not stablecoins, even though they sit close to the yield-bearing stablecoins many readers already know. There is no peg to defend. The risk is different: a write-down of the share’s value after a credit loss, or a wait in the withdrawal queue. Maple reported 30-day returns of 5.0% on syrupUSDC and 4.4% on syrupUSDT in its 15 October update, adding that yields are variable and not guaranteed.

How a syrupUSDC deposit becomes a loan

A lender goes through Maple’s router into a pool, and a permission contract controls which wallets may take part. The money is then lent by Maple Direct, which the syrup documentation calls Maple’s lending arm and the party responsible for underwriting, structuring and managing the loans. Borrowers are crypto-native institutions that post liquid digital assets as collateral.

This is the point that separates Maple from most lending protocols. On Aave or Morpho, a smart contract decides who may borrow, by checking collateral. On Maple, people decide. The contracts record the loan, but the choice of borrower, the terms and the collateral levels come from a credit team working off-chain. Maple does not name its borrowers on the pages we read, so outside lenders cannot assess them one by one.

What lenders can see is the book. On the syrupUSDC page, read at about 00:45 UTC on 31 October 2026, 24 loans totalled $898.72 million against $1.53 billion of collateral, a ratio of 169.7%. The syrupUSDT page showed seven loans totalling $215.05 million against $309.94 million, or 144.1%, and the syrupUSDG page eight loans of $176.7 million against $278.34 million, or 157.5%. Loans come in two forms: fixed term, with a maturity date, and open term, with a next call date.

The averages hide a range. Across all 24 syrupUSDC loans, collateral levels ran from 119.2% to 618.5%, and four 25 million USDC loans sat at the bottom of that range. On syrupUSDT, one 50 million USDT loan was shown with $50 million of USDC as collateral, a level of 100.0%. The page does not explain that line. The words “Fully overcollateralized” on the same page sit in the Proof of Reserves panel, which it describes as verifying that reserves back deposits, and the pool as a whole shows 144.1%. A dollar stablecoin against a dollar stablecoin is a different risk from bitcoin, but the figure leaves no cushion on that loan.

Bitcoin dominates. Adding the three pages, BTC and wrapped bitcoin made up about 77% of roughly $2.11 billion of collateral, XRP about 13%, ETH about 7% and USDC about 2%. Those shares are our calculation from the asset tables. Lending is not the only use of the money: it took 96.7% of syrupUSDC deposits, 72.2% of syrupUSDT, where a further 22.8% sat in a Treasuries category, and 55.6% of syrupUSDG, which had 29.3% in DeFi venues including Morpho and Aave v4.

Getting out: the queue

A withdrawal request joins a first-in, first-out queue and is paid as liquidity becomes available. Maple’s risk page says most are processed in under 24 hours but “could take up to 30 days”, and that requests keep earning interest until then. The alternative is selling on a decentralised exchange such as Uniswap or Balancer, with slippage at the seller’s expense.

Maple’s list of assumptions is blunter than its marketing. It says “Pool Delegates are trusted actors”, and that the withdrawal queue gives the Pool Delegate and the Operational Admin discretion: they can leave requests in the queue unprocessed. In July 2025 LlamaRisk, a risk reviewer for Aave, warned that redemptions would be queued in a stress scenario, leaving decentralised exchanges as the only immediate exit. Chaos Labs, another reviewer, put syrupUSDC pool depth at about $20 million each on Fluid and Uniswap in January 2026. Those figures are dated, but small next to a $930 million deposit base.

What actually triggers a liquidation on Maple

Each loan sets three loan-to-value levels at origination: initial, margin call and liquidation. Maple’s Loan Default and Impairment Procedure, an undated document it hosts publicly, says these are tested continuously against market data. It then lists the conditions that authorise liquidation:

• A margin call’s cure period ends with the loan’s LTV still above the margin call level.
• The loan’s LTV breaches the liquidation level. The lender docs add that this applies even if a margin call is running.
• An Event of Default occurs under the master loan agreement. The listed events include non-payment of principal, interest or late fees, uncured covenant breaches, insolvency, failure to transfer collateral and a material default on other loans.

The cure window depends on the document. The procedure says a 12 to 24 hour timer starts, the lender docs say 24 hours, and the borrower docs say the term sheet defines it. A borrower can add collateral or pay down principal. Maple prefers to sell through OTC desks, falling back on centralised or decentralised exchanges, and borrowers have no say over the order, manner or price of a sale. Our guide to how DeFi liquidations work covers the bot-driven version; here Maple does the selling.

Price comes first, then, not the calendar. A maturity matters in only one way: if the borrower does not repay on the due date, non-payment of principal is an Event of Default, and that can authorise a sale of the collateral once any grace period in the contract has run. A borrower who repays, or agrees a refinancing, triggers nothing.

If a sale falls short, the order is fixed by the procedure. Proceeds pay down principal, interest and fees. Any remainder is booked on-chain as an impairment, which needs an Investment Committee approval and a 3-of-5 multisig. Maple then pursues the borrower, since the loans are recourse. Only when recovery has failed does it mark an on-chain default, and the loss reaches lenders through a permanently lower exchange rate. The procedure adds that a legal default does not immediately touch the on-chain exchange rate, so a share price is not a live signal of trouble on a loan.

Timing can cost a lender money. Under Maple’s legal terms on impairments, someone who withdraws during an impairment takes a permanent loss and gives up any claim on later recoveries, which go pro rata to those who stayed. The same page describes impairment as a tool used before a technical default, when Maple doubts a borrower can repay, whereas the procedure says it follows liquidation. The two documents do not describe the same sequence.

The procedure claims “hundreds of margin calls, one partial liquidation and no lender capital lost” on over $5 billion of originations, without giving a period. A May 2026 post claims zero losses on over $15 billion of overcollateralised loans. Neither sentence mentions the 2022 losses described below, which came under Maple’s earlier model.

The $280 million of XRP collateral and three maturity dates

CryptoSlate’s token counts match Maple’s disclosures: the syrupUSDC book holds 142,766,086.05 XRP and the syrupUSDT book 82,314,668.8 XRP. No XRP appears on the syrupUSDG page. DefiLlama’s XRP Ledger line for Maple read $281.2 million at 22:19 UTC on 30 October, the same collateral from another angle. Dollar values move with the price: the table uses Maple’s pages at about 00:45 UTC on 31 October, when the two XRP positions added up to about $283 million.

The headline needs reading with care. The $280 million is the value of the collateral, not of the loans, whose principal is about $151.8 million. “Locked” only means pledged, and the XRP goes back under the release terms of each loan. Liquidation would need the conditions listed above, not the arrival of a date.

Our own arithmetic, not a forecast: the thinnest of the three loans stands at 180.1%, so XRP would have to fall about 44% from the level on Maple’s page before the collateral only equalled the principal (1 minus 100 divided by 180.1). A margin call would come well before that, but the margin call and liquidation levels of these loans are not shown on the public pages.

CryptoSlate itself notes limits that its headline does not carry. Three positions need not mean three borrowers, because Maple’s Borrower Hub supports several legal entities under one organisation. A maturity also does not guarantee release of the collateral, since renewals can extend its use. Whether these loans will be repaid, refinanced or renewed is not public, and we found no Maple statement on them. The exposure that matters is concentration: by our calculation XRP is about 15% of collateral on the two books that hold it, and a sharp fall would test all three loans at once.

Who controls Maple: Cayman foundation, Guernsey trust, three multisigs

The structure below comes from Maple’s Blockworks Token Transparency filing of 18 June 2026. It is self-reported and marked partial, with four gaps, and Blockworks reviews completeness rather than verifying the answers. A Maple post of 8 July 2026 describes the same outline.

According to the filing, the Maple DAO Foundation, a memberless Cayman foundation, wholly owns Maple DAO (BVI) Ltd, which issued SYRUP, and Maple International Operations SPC, the Cayman segregated portfolio company behind the loans. Maple Labs Pty Ltd, in Australia, develops the software under a services agreement. The intellectual property sits in the Maple Purpose Trust in Guernsey. The Foundation holds a licence, and the July post calls it the trust’s sole beneficiary.

Each syrup product has its own segregated portfolio within that SPC, which the docs say gives full legal separation between products and from the institutional pools.

On-chain power sits with three multisigs, per the filing. A 4-of-7 DAO multisig acts through a GovernorTimelock with a three-day delay. A 3-of-5 Operational Admin handles borrower approvals and fees. A 3-of-6 Security Admin can pause the protocol. The timelock is a real constraint, with one exception worth knowing. The docs say the Security Admin can call upgrades of PoolManager, LoanManager and WithdrawalManager contracts with no scheduled call, and of strategies with no timelock, although registering an upgrade path on a factory is a timelocked Governor action.

Access has limits too. Maple’s jurisdiction page bars residents, citizens and tax residents of 34 listed territories from syrupUSDC, syrupUSDT and syrupUSDG, along with entities based there. The United States is on the list, so US residents are excluded, and so is Australia, home of Maple Labs. Maple’s KYC page says it checks all lenders and borrowers and screens lender wallets. The EU is not on the restricted list. That is not a licence, and the Maple documents cited here claim none.

Is Maple Finance a $4.7 billion or a $2.55 billion protocol?

Both numbers exist because they count different things, and neither is the money lenders put in. DefiLlama showed total value locked of about $2.55 billion at 22:19 UTC on 30 October 2026. Its method adds pool cash, cash deployed in DeFi strategies on Ethereum and the loan collateral Maple reports, each asset placed on the chain where it is held. That explains the split: Bitcoin $1.93 billion, Ethereum $336 million, XRP Ledger $281 million. Loan principal is shown apart, as $1.82 billion borrowed.

AUM is Maple’s own measure. Neither the Q3 update nor its transparency page defines it, but the page’s figures point to an answer. Read on 31 October, it showed $4.41 billion of AUM against $1.87 billion of deposits. On syrupUSDT, $298.2 million of deposits plus $309.94 million of collateral gives exactly the $608.14 million of AUM shown, and syrupUSDG works the same way. So by our reading, displayed AUM is deposits plus the borrowers’ collateral, which Maple holds as security and which can move with the bitcoin price. Maple does not state that formula. The $4.41 billion is also below the $4.7 billion of the Q3 update and the $4.8 billion of the October memo.

The trend is less flattering than the annual growth. Maple said AUM rose 18% over a year and loans 12%, while DeFi lending shrank 29% this year. But the Q3 update also says loans fell 9% from $1.8 billion in Q2, and its July report had put Q2 loans at $1.9 billion, an all-time high. Maple does not explain the difference. DefiLlama’s TVL went from about $2.96 billion on 1 October to about $2.55 billion, a fall of about 14% with no stated reason. Collateral counts at market value, so prices alone can move that line.

Revenue is modest beside those balances. Maple reported $4.4 million for the third quarter, an annual run rate of $17.6 million, and a net interest margin of 92 basis points. Two 12-month views differ. DefiLlama, read on 31 October, showed about $112 million of fees and about $14.6 million of revenue, about 13% by our calculation. Maple’s financials page showed $127.4 million gross and $22.2 million of protocol revenue, about 17%. The filing says Maple typically charges an average of 1.5% to 2% of loans outstanding.

The SYRUP token: votes and buybacks, not a share of revenue

SYRUP replaced MPL at 100 to 1. Maple’s self-reported filing dates the migration’s launch to 15 November 2024, a date the docs do not give, and the docs say the converter closed for good on 21 May 2025. Conversion was set to create about 1.15 billion SYRUP. The filing says a 5% yearly emission was approved from November 2023 to November 2026, and the docs expected 1,267,875,000 tokens by September 2026. Maple’s SYRUP page showed 1.14 billion circulating and 1.27 billion total on 31 October, with 126.09 million held by Maple. The filing gave 1,191,759,306 circulating on 5 June 2026, and neither source explains the difference.

The vesting facts also rest on the filing alone: it reports original investor and team vesting as complete, about 5% of circulating supply still locked or unvested for employees, and new core team grants on a 12-month cliff and 24 months of linear vesting. It lists Keyrock, STS and Pulsar as market makers on 12-month loan and call option terms, committing about 2.5% between them.

A holder gets a vote, on Snapshot, over a seven-day window. Staking rewards ended with MIP-019, and stSYRUP can be unstaked at any time. Maple’s own explainer says all revenue flows to the Foundation, which has no members or shareholders. It describes no legal claim for token holders on that revenue or on Maple’s assets.

Value is meant to reach holders through buybacks. Since MIP-021 went live on 13 July 2026, the share of revenue used depends on monthly revenue: 10% below $1.5 million, 20% from $1.5 million to $2 million, 30% above. September crossed $1.5 million, and the Syrup Strategic Fund bought 1.37 million SYRUP, against 0.68 million in August and 2.9 million for the quarter. By our calculation that quarter’s total is about 0.25% of the 1.14 billion circulating, and the roughly $598,000 Maple’s page shows for it is about 14% of Q3 revenue.

Bought tokens stay in the DAO treasury, neither burned nor distributed, according to the filing. Maple says buybacks “shrink the circulating supply” but also that they are “not a dividend, distribution, profit share, or guaranteed return”.

Audits, reserves and the exploit record

The audit list in Maple’s docs is long and the reports open from its public GitHub repositories: Trail of Bits, Spearbit and Three Sigma for the 2022 rebuild, then 0xMacro, Three Sigma and Cantina through 2023 and 2024. The September 2025 GovernorTimelock upgrade went to Sherlock and 0xMacro, the November 2025 Withdrawal Manager upgrade to Spearbit and Sherlock, and the cross-chain receiver to Dedaub and Sigma Prime. The docs also list a November 2026 shared strategies release. Its two reports sit on Google Drive links from that page rather than in the repository: Sherlock lists no high or medium issues and four low or informational ones, and Cantina, dated 15 October 2026, lists five low and four informational. The release itself is not confirmed live.

Audits cover code, not credit: none tells a lender whether a borrower will repay. Maple also runs an Immunefi bug bounty.

For collateral, Proof of Reserves launched on 7 May 2026 for syrupUSDC and syrupUSDT, with The Network Firm. The launch post marked the yes-or-no answer on full collateralisation as coming soon, and the pages now carry the “Fully overcollateralized” label with a last check on 30 October at 17:00 UTC. The post does not state an assurance level, so it should not be read as a full audit.

The filing says no exploit had hit Maple’s contracts or protocol funds as of 4 June 2026, and DefiLlama’s hacks list had no Maple entry when read on 31 October 2026. The losses below came from credit, not code.

Past losses: the 2022 Orthogonal default and the Core dispute

Orthogonal Trading, December 2022

On 3 December 2022 Orthogonal Trading told Maple and pool delegate M11 Credit it could not repay. ForkLog reported eight loans worth $36 million, about 30% of Maple’s active loans, with about $31 million in the M11 USDC pool and $5 million in a WETH pool. Decrypt reported that Maple and M11 accused the firm of misrepresenting its position and said large losses stemmed from funds on FTX.

The sources we could open give no final loss. Maple deployed its V2 contracts on 14 December 2022 and later moved to the overcollateralised model it uses now, with credit run in-house rather than by outside delegates.

Core Foundation and lstBTC, 2025 to 2026

Core Foundation and Maple built lstBTC, a bitcoin yield product, in early 2025. In November 2025 DL News and Decrypt reported that a Cayman Islands court had granted Core an injunction blocking Maple from launching a rival bitcoin yield product. Press accounts of the order say the judge found a “serious issue to be tried”, a threshold test and not a finding of wrongdoing.

Core alleged misuse of confidential information and breach of a 24-month exclusivity clause, and said it did not understand why Maple maintained it was “unable to return the Bitcoin to their lenders at this time”. Those were allegations. Maple denied wrongdoing, said Core’s actions were “directly against lender interests” and told DL News: “The dispute is strictly limited to the pilot programme conducted in partnership with Core Foundation for BTC Yield.”

On 22 May 2026 the two sides announced a full and final settlement, with mutual releases and no admission of liability or wrongdoing by any party: “The financial and other terms of the settlement are confidential.” The announcement does not say whether or how lenders in the pilot were repaid, and we found no public source that does. It does say Maple will launch syrupBTC “in the ordinary course”; we could not confirm that it has.

Aave, Robinhood and Kraken: where syrup tokens travel

Distribution is Maple’s growth story, and it spreads the exposure. Aave‘s public market data, read at about 00:45 UTC on 31 October 2026, lists syrupUSDT on Ethereum Core (about 80.6 million supplied, cap 100 million) and on Plasma (about 40.7 million, cap 50 million), and syrupUSDC on Base (about 46,000, cap 10 million). Chaos Labs dates the Plasma listing to late October 2025. On Base, a proposal of 30 January 2026 sought to correct a liquidation protocol fee mistakenly set to 0%. syrupUSDC does not appear on Ethereum Core in that data, although a Snapshot vote on listing it closed in July 2026 with no votes against.

Aave’s reviewers backed the listings with conditions. In its July 2025 review of syrupUSDC, LlamaRisk wrote: “we strongly recommend that supply caps for syrupUSDC on Aave be sized conservatively”, tying them to market depth, and noted that credit underwriting had been centralised under Maple Direct.

syrupUSDG launched on 1 July 2026 on Ethereum and Robinhood Chain, built on the Paxos-issued USDG. It is collateral in a Morpho vault curated by Steakhouse Financial behind Robinhood Earn, which Maple says has passed $136 million. That puts Maple’s credit book one step behind the app.

With Kraken, Maple closed a warehouse facility on 25 June 2026, giving senior financing through a bankruptcy-remote vehicle to BTC and ETH-backed loans from Kraken’s OTC desk. The size was not disclosed. Sid Powell said this infrastructure “has never existed onchain, until now”.

Maple against other lending protocols

By TVL, Maple ranked fifth among lending protocols on DefiLlama, as read on 31 October 2026, behind Aave V3, Morpho Blue, SparkLend and JustLend V1. Its line had moved up slightly from the 30 October reading used above. The models differ: see our guide to crypto lending and borrowing for the public-pool version.

Maple’s pitch is yield: it says syrupUSDC and syrupUSDT averaged 126 basis points above the peers it tracks over 30 days, and that it holds 17% of tracked DeFi loans. Both are company figures with its own benchmark. In exchange, lenders accept what public pools do not ask of them: unnamed borrowers, human credit judgment and a queue.

What to watch next

The next AMA is on 11 November 2026, and the docs point to a November contract release, no day given. The filing’s emission period also ends in November. The XRP loans fall due on 4 and 12 December and on 15 January 2027: the Maple pages will show whether each is repaid, refinanced or renewed. The next quarterly call is due in January 2027, no day given. The company also plans fintech receivables and lending against rated securities, which would move the book away from the liquid crypto collateral its current risk case rests on.