Private credit is one of the most discussed categories of tokenised real-world assets, and one of the most loosely described. A press release calls a fund "onchain" and a headline calls the result "DeFi lending", as if putting a loan on a blockchain changed what the loan is. It rarely does. What changes is who can hold a claim on it, how fast that claim settles, and how visible the loan's performance is to the holder. The credit risk, the borrower, and the legal documents that decide what happens if the borrower stops paying stay where they always were.
How big it is, and what the number counts
RWA.xyz tracks two figures for tokenised private credit. As of 11 September 2026 it counts $7.98 billion of "distributed" value and $35.70 billion of "represented" value, across 2,591 assets and about 196,500 holders. The distinction is about mobility, not size: in RWA.xyz's own framework, distributed assets can be moved to wallets outside the issuing platform and transferred between wallets, while represented assets cannot, so the ledger serves as a record rather than a means of distribution.
Neither figure should be read against private credit as a whole. The Financial Stability Board's Report on Vulnerabilities in Private Credit, published on 6 May 2026, estimates the sector's total size at between $1.5 trillion and $2 trillion. On that estimate, the distributed figure is roughly 0.4 to 0.5 per cent of the asset class it borrows its name from. The published summary of the FSB report does not mention tokenisation at all. The wider tokenisation market is harder still to size, as our piece on what institutions actually argued about tokenisation found. Private credit is at least more legible: a loan has a principal balance and a payment schedule.
Four things called "tokenised private credit"
Home equity lines on a registry: Figure
Figure Technology Solutions originates home equity lines of credit (HELOCs) through its loan origination system, both under its own brand and for partner lenders who use its platform. It records those loans on Provenance Blockchain through DART, which its prospectus dated 18 February 2026 describes as "our lien and eNote registry technology". The same filing reports more than $60 billion "in both real-world and digital asset transactions" from Figure's launch in late 2018 to 14 October 2025, and says that in the nine months to 30 September 2025, 85% of loans originated through its system, Figure-branded and partner-branded, were boarded onto DART.
The prospectus is plain about what the token does and does not add. "The blockchain tokenized real-world assets are 'digital twins' of the real-world asset representations," it says, and "the token itself has no additional rights or value other than those granted by the ownership of the real-world asset." The HELOC is still a lien governed by state law. The blockchain entry records who owns the claim to it; it is not a second claim.
A feeder into an existing fund: Apollo's ACRED
ACRED, the Apollo Diversified Credit Securitize Fund, is a different construction. Announced by Apollo and Securitize on 30 January 2025, it is a feeder fund offering tokenised access to Apollo Diversified Credit Fund, which the announcement describes as a non-traded, closed-end interval fund investing across corporate direct lending, asset-backed lending, and performing, dislocated and structured credit. Qualifying investors reach ACRED only through Securitize Markets, LLC, Securitize's broker-dealer subsidiary, and Securitize acts as the fund's digital transfer agent and fund administrator. The tokens launched on Aptos, Avalanche, Ethereum, Ink, Polygon and Solana, with Wormhole named as the interoperability partner for moving between them.
A token holder's claim runs to shares in a feeder fund, which holds an interest in a credit fund Apollo already managed for conventional investors. Nothing about how those loans are made changes because the feeder exists.
A senior structured tranche: Janus Henderson Anemoy AAA CLO Fund
The Janus Henderson Anemoy AAA CLO Fund, whose token is JAAA, sits a step further from any individual borrower: it holds AAA-rated tranches of collateralised loan obligations. Janus Henderson is sub-adviser, Anemoy is the fund vehicle, and Centrifuge provides the tokenisation technology. Crypto Briefing reported on 10 August 2026 that the fund launched in June 2025 with a $1 billion commitment from the Sky Ecosystem through Grove, and held roughly $689.9 million across 36 AAA-rated CLOs. The fund is promoted on the claim that no AAA-rated CLO tranche has defaulted or suffered a principal loss in more than three decades. That is a statement about the seniority of the asset class, made by the parties selling access to it, not a property the token adds.
Crypto-collateralised lending: Maple Finance
Maple Finance is the one entry here that has visibly changed shape. Its site, accessed on 11 September 2026, says it has originated $25.52 billion of loans since launching in 2019. In 2022 its pools were run by delegates who underwrote the loans, and, as reported in December 2022, borrowers did not have to post collateral that could be seized or liquidated on default; the protocol was popular with crypto trading firms and market makers. Its current products lend against crypto collateral instead. OAK Research, in a report updated on 16 June 2026, says Maple's loans are kept at collateralisation ratios above 150%, with a Blue Chip pool accepting only BTC and ETH and a High Yield pool accepting riskier collateral such as XRP, SOL or POL, and that the protocol reports zero losses since 2023.
That is a real reclassification. Overcollateralised lending against liquid assets is closer to secured margin lending than to private credit in the usual sense, where a lender underwrites a borrower's cash flow or an illiquid asset. It moves the main risk from the borrower's creditworthiness to the value of the collateral and the speed of liquidation. That is a different risk, not an absent one, and it answers a different question from Figure, Apollo or Janus Henderson.
Origination and servicing sit where they did before
In three of the four cases, tokenisation does not change who decides whether to lend. Figure's system and its partner lenders originate the HELOCs; DART is a registry. The loans inside Apollo Diversified Credit Fund are Apollo's, and Securitize's roles are transfer agency and fund administration for the wrapper. Janus Henderson selects the CLO holdings; Centrifuge supplies the technology. Only on Maple does the platform sit close to the credit decision, because in its first version, OAK Research notes, delegates selected borrowers and set loan terms on the protocol itself.
Our piece on taking RWA tokenisation from pilot to production counted eight systems that have to agree before any tokenised asset reaches production. For credit, servicing carries the most weight: a loan, unlike a Treasury bill, can stop performing partway through its life, and someone has to be contractually responsible for the workout. None of the four structures solves that with code. Each relies on what lending has always relied on: a servicer, covenants, and in the end, lawyers.
Where the yield comes from
The yield on any of these products is the return on the underlying credit, whether a HELOC rate, a direct-lending spread, a CLO coupon or what Maple's borrowers pay to draw against collateral, minus whatever the platform or manager charges. The blockchain does not create yield. At most it removes friction and settlement lag between the loan and the holder of a claim on it.
The subject was on the programme at the archived Paris 2026 edition. "Where Stablecoins Meet Credit: Rewiring Finance Through Tokenisation" ran on the Hecto Main Stage on 3 June 2026, moderated by Michael del Castillo, Founder of Media Luna Creations. The panellists were Anthony Bassili, President of Coinbase Asset Management; Chris Cox, Head of Investor Services at Citi; Francesco Filia, CEO of Fasanara; and Maha Al-Saadi, billed as Independent. Titles are as the programme billed them.
Five weeks before that session, on 30 April 2026, Coinbase Asset Management announced the Coinbase Stablecoin Credit Strategy, which CoinDesk reported targets institutional investors seeking yield from lending activity tied to digital assets, with a tokenised share class on Superstate's FundOS platform across Ethereum, Solana and Base. None of this is investment, credit or legal advice. It describes where the return originates, which is the question to ask of any tokenised credit product before asking which chain it settles on.
What happens on default
Credit risk does not change because the claim sits on a ledger, and the record bears that out. On 5 December 2022 CoinDesk reported that Orthogonal Trading had defaulted on $36 million borrowed from Maple Finance pools run by the delegate M11 Credit. According to M11, Orthogonal told it on 3 December that it had "incurred larger losses than previously disclosed" on funds held at FTX. $31 million sat in M11's USDC pool, about 80% of that pool's $38.5 million of active loans; a further $5 million in wrapped ether was about 18% of the $28 million outstanding in a second pool. Maple cut ties with Orthogonal, alleging that it had misrepresented its financial position. M11 and Maple then applied to the British Virgin Islands High Court, which placed Orthogonal Trading in provisional liquidation, and hired Kroll to help recover assets for depositors. The recovery ran through a court and an investigations firm, not a smart contract.
Goldfinch had two credit events in 2023 that show a different problem: who absorbs a loss is not always fixed in advance. Tugende, an asset-finance company, took a $5 million loan through Goldfinch in October 2021, stayed current on interest until May 2023 and defaulted from June 2023, TechCrunch reported. It had breached loan-to-value and tangible net worth covenants, and $1.9 million of a loan designated for Kenya had gone to its struggling Ugandan operations without the Goldfinch community's consent. A restructuring was projected to cut the write-down of the senior pool's net asset value from about 3.95% to less than 0.79%.
Separately, in October 2023 Warbler Labs disclosed that loans to REZI ($5 million) and POKT ($2 million), made from a $20 million facility to Stratos, would be written down to zero, a combined $7 million, while the rest of the facility's loans were performing. Warbler said it would backstop all losses for the loan's backers, excluding Warbler and Stratos. In both cases, how much of the loss fell on whom was settled after the default, through a restructuring or a sponsor's decision, not by code written before the loan was made.
None of the four programmes above belongs in the same category as Stream Finance, which said on 4 November 2025 that an external fund manager had lost about $93 million of assets tied to the project. Its xUSD token fell from $1.26 to as low as $0.16 within a day, and because markets on Euler, Morpho, Silo and Gearbox had accepted xUSD as collateral, The Defiant reported some $285 million of debt intertwined across curators, vaults and lending markets. Stream ran a synthetic-dollar yield strategy, not a private credit fund. It belongs here only as a reminder that a yield-bearing token accepted elsewhere as if it were cash is a separate risk from the credit underneath it.
What regulators have said, and what they have not
On 11 November 2025 IOSCO published its final report on the tokenisation of financial assets, prepared by its Fintech Task Force. It identifies new or heightened risks, such as interoperability challenges and the lack of credible settlement assets, and encourages regulators to apply its existing recommendations on the principle of "same activities, same risks, same regulatory outcomes". It is a report with recommendations, not a binding rule anywhere, and its use cases are tokenised money market funds and fixed income, not private credit. The FSB's private credit report six months later covers private credit without discussing tokenisation. Neither treats tokenised private credit as a category of its own.
For anyone assessing a tokenised credit product, the practical test is the one in Figure's own prospectus: the token carries no rights beyond the real-world claim. Read the fund's or pool's default and recovery provisions before its blockchain's confirmation times. From Orthogonal Trading's provisional liquidation to Warbler Labs' backstop, when a tokenised credit pool has taken a loss, who absorbed it was settled by fund documents, restructurings, sponsors and courts, not by the contract that issued the token.