On-chain private credit protocols reported $12.9 billion in total value locked as of mid-September 2026, with cumulative originations exceeding $33.7 billion since inception. These figures represent less than 0.65% of the traditional private credit market, which Morgan Stanley projects at $2 tril...
"Capital is rotating toward transparent, collateralized, cashflow-backed credit infrastructure." — Sidney Powell, CEO, Maple Finance
On-chain private credit protocols reported $12.9 billion in total value locked as of mid-September 2026, with cumulative originations exceeding $33.7 billion since inception. These figures represent less than 0.65% of the traditional private credit market, which Morgan Stanley projects at $2 trillion AUM for 2026. Despite the modest penetration, growth rates remain steep: Maple Finance closed H1 2026 at $4.6 billion AUM, up 81% year-over-year, while Centrifuge held $1.62 billion TVL backed by a Coinbase equity stake and a New York Life Investment Management tokenization partnership signed in June 2026.
The sector's economic reality is more nuanced than headline numbers suggest. According to an analysis by YieldDesk, 73% of tracked on-chain private credit loans are "representative tokenization" — blockchain record-keeping without freely transferable tokens. Only approximately $5 billion exists as distributable, investable assets accessible to DeFi participants. Meanwhile, the market is bifurcating: overcollateralized institutional lenders are scaling, while undercollateralized emerging-market platforms are contracting. Goldfinch Prime, which offered access to institutional credit funds from Apollo and Ares, shut down via unanimous tokenholder vote on June 23, 2026, citing insufficient adoption. The sector's survivorship curve is becoming legible.
The global private credit market reached approximately $2 trillion in AUM by 2026, according to Morgan Stanley, with projections targeting $3 trillion by 2028. The structural driver is regulatory: under Basel III, U.S. bank underwriting of middle-market loans fell from over 70% in 1994 to roughly 10% by 2020, creating a persistent funding gap that nonbank lenders now fill.
On-chain private credit occupies a fraction of this market. RWA.xyz's tokenized credit dashboard shows $5.12 billion in distributed value and $20.32 billion in represented value across tokenized credit assets, with more than 184,000 holders. The distinction matters: "distributed value" refers to tokens that can actually move between wallets and integrate with DeFi. "Represented value" includes institutional positions that exist as blockchain entries but remain locked in custodial or permissioned structures.
The chain distribution is concentrated. Ethereum hosts the primary deployments across all major protocols. Maple Finance's syrupUSDC product extends to Solana, Arbitrum, and Base. Centrifuge's SPXA index fund operates on Base. Clearpool deploys on Polygon. Despite the multi-chain expansion, Ethereum remains the settlement layer for the majority of on-chain private credit originations.
The on-chain private credit market is not a single product category. It comprises at least four distinct credit models, each serving different borrower profiles and carrying different risk characteristics.
Maple Finance — $4.6B AUM, $22B+ cumulative originations. Following $50 million in lender losses during the 2022 crypto credit crisis (primarily from the $36 million Orthogonal Trading default), Maple rebuilt around overcollateralized institutional lending. Current loans maintain collateralization ratios above 150% in BTC, ETH, or SOL. The protocol reported zero principal losses since 2023 and $13.7 million in year-to-date revenue through H1 2026, implying an average revenue yield of approximately 0.185% on originations. Its syrupUSDC product grew 1,826% year-over-year. In H1 2026, Maple launched syrupUSDG — the first new Syrup asset in two years — which reached $200 million AUM in eight days and became the first decentralized lending product integrated into Robinhood's Earn feature.
Centrifuge — $1.62B TVL. Centrifuge operates a tranched real-world asset pool structure inherited from its Tinlake architecture, with senior/junior tranche risk separation. Through its affiliated asset manager Anemoy, the protocol offers institutional tokenized funds sub-advised by Janus Henderson and Apollo. Products span U.S. Treasuries, AAA-rated CLOs, and diversified credit strategies. MakerDAO/Sky integration positions Centrifuge as a core RWA collateral source. COO Jürgen Blumberg projected RWA TVL would exceed $100 billion by end-2026.
Clearpool — $950M+ cumulative originations. Clearpool operates permissionless single-borrower pools for institutional counterparties. The protocol crossed $950 million in total loan originations in August 2026. Its strategic pivot toward PayFi — payment financing for fintech firms processing cross-border transfers and card transactions — produced cpUSD, a yield-bearing stablecoin backed by short-duration (1-7 day) payment-flow lending. Clearpool Prime, the regulated institutional credit arm, is being scaled for vetted borrowers raising liquidity on-chain.
Goldfinch — Winding down. Goldfinch pioneered undercollateralized lending to emerging-market fintech companies across Africa, Southeast Asia, and Latin America. The model produced $56 million in active legacy loans but accumulated three documented defaults totaling approximately $18 million in losses: Tugende (Kenya, $5 million), Stratos (U.S. credit fund, $7 million), and Lend East (Southeast Asia, $10.2 million with approximately $5.9 million unrecovered). On June 23, 2026, tokenholders voted unanimously to shut down Goldfinch Prime, which had offered exposure to institutional private credit funds from Apollo Global Management, Ares Management, and Golub Capital. The stated reason: "has not achieved the level of adoption needed to justify continued investment in new product development, marketing, or operational expansion." A new U.S. trust will pursue recoveries from the eight legacy borrower pools.
The aggregate on-chain private credit default record is small in absolute terms but instructive in pattern.
| Protocol | Year | Loss | Cause | |----------|------|------|-------| | Maple (Orthogonal Trading) | 2022 | $36M | Undercollateralized lending to fraudulent trading firm | | Maple (other 2022 losses) | 2022 | ~$14M | Contagion from crypto credit crisis | | Goldfinch (Tugende) | 2023 | $5M | Kenyan fintech borrower default | | Goldfinch (Stratos) | 2023 | $7M | U.S. credit fund default | | Goldfinch (Lend East) | 2024 | ~$5.9M | Southeast Asian lending platform, partial recovery |
Total documented losses: approximately $68 million. Notably, all losses occurred under undercollateralized or insufficiently audited lending models. Since Maple's pivot to overcollateralized lending in 2023, the protocol reports zero liquidation events through multiple market shocks, with margin calls cured within hours through real-time monitoring and automated margin systems.
The caveat: on-chain private credit has not been stress-tested through a traditional credit cycle downturn. The 2022 defaults were crypto-specific contagion events, not broad economic deterioration. With the Federal Reserve hiking rates to levels not seen since 2023, the sector's first real credit-cycle test may be approaching.
On September 16, 2026, Tare announced a $13.25 million seed round led by Blockchain Capital, with participation from Janus Henderson, Strobe Ventures, the Avalanche Foundation, and individual backers including Aave CEO Stani Kulechov and Tether co-founder Phil Potter. The round closed in March 2026.
Tare's approach differs from protocol-level lending. Co-founded by Kevin Miao (formerly Citigroup; founder of BlockTower Credit, a $2 billion institutional private credit fund) and Lucas Vogelsang (co-founder of Centrifuge), the company is building infrastructure on Avalanche to consolidate the fragmented back-office operations that underpin private credit: loan origination, record tracking, payment distribution, and accuracy verification.
"We deserve to have a system that works for us and doesn't extract from both sides," Miao stated, according to Fortune. The company plans to establish Tare Credit LLC, a lending arm licensed to operate across the United States.
The infrastructure-layer thesis represents a shift in where value accrues. Rather than competing for borrower-lender spreads at the protocol level, Tare targets the administrative cost burden that inflates borrowing rates and compresses investor returns in traditional private credit workflows.
On-chain private credit yields operate across a defined risk spectrum, with spreads over the risk-free rate (currently approximately 3.6-3.7% on T-bills) providing the compensation for credit, liquidity, and smart-contract risk.
| Product | Yield Range | Spread Over Risk-Free | Risk Profile | |---------|-------------|----------------------|--------------| | Maple syrupUSDC | 4.6-7% APY | 90-350 bps | Overcollateralized institutional, BTC/ETH/SOL backing | | Centrifuge Senior | 8-12% APY | 400-840 bps | Tranched RWA pools, senior priority | | Clearpool Prime | 8-15% APR | 400-1,140 bps | Single-borrower concentration, institutional vetting | | Clearpool cpUSD | Variable | Short-duration PayFi | 1-7 day settlement cycles, payment-flow backing |
The yield sources are structurally different from DeFi lending protocols like Aave or Compound, which derive returns from on-chain collateral utilization. On-chain private credit yields come from off-chain borrower cash flows: trading desk revenue (Maple), trade finance receivables (Centrifuge), and payment processing float (Clearpool). This distinction means credit risk assessment depends on off-chain due diligence rather than on-chain liquidation mechanics.
Maple's MIP-021 buyback mechanism channels 10% of monthly revenue to token buybacks when earnings fall below $1.5 million monthly, scaling to 30% above $2 million — a structure that ties protocol token value directly to credit origination revenue rather than speculative demand.
Liquidity mismatch. Only Maple's syrupUSDC offers meaningful secondary liquidity. Centrifuge pools require lockup periods. Goldfinch's legacy pools have redemption queues extending months. In a stress scenario, investors in illiquid on-chain credit positions face the same exit constraints as traditional private credit — with the added complication that DeFi composability creates expectations of instant redeemability.
Legal enforcement gap. On-chain defaults rely on off-chain recovery in varying jurisdictions. Goldfinch's Lend East default in Southeast Asia produced different recovery outcomes than its U.S.-based Stratos default. Cross-border legal enforcement remains the sector's unsolved structural problem. The new U.S. trust pursuing Goldfinch legacy pool recoveries will test this enforcement mechanism in practice.
Borrower concentration. Maple's borrowers are concentrated in the crypto sector (trading firms, market makers). Clearpool's single-borrower pools carry concentration risk by design. A sector-wide crypto drawdown would simultaneously stress Maple's collateral values and its borrowers' ability to service loans — a correlation risk that overcollateralization ratios may not fully offset.
Regulatory uncertainty. The CLARITY Act's failure on September 15, 2026 (49-50 cloture vote) leaves yield-bearing stablecoin products like Clearpool's cpUSD in regulatory limbo. The SEC's September 1, 2026 transfer agent proposal (Release No. 34-106246) does not directly address lending protocol structures. Whether on-chain credit instruments require securities registration remains an open question.
On-chain private credit in September 2026 presents a sector that has survived its first default cycle, consolidated around overcollateralized models, and begun attracting institutional infrastructure capital. The $12.9 billion TVL figure, while frequently cited, obscures a more modest $5 billion investable market — still small against a $2 trillion traditional private credit industry.
The economic question is whether on-chain infrastructure produces measurable cost savings in loan administration, settlement, and transparency that justify the additional smart-contract and regulatory risks. Maple's 18.5-basis-point revenue yield on $22 billion in cumulative originations suggests the economics can work at institutional scale. Goldfinch's shutdown suggests they cannot work without institutional-grade underwriting. The sector's next test is not growth — it is whether overcollateralized models can maintain their zero-loss record through a macroeconomic credit tightening that the Federal Reserve's September 2026 rate hike may be beginning to produce.