Private credit on-chain offers higher yields than tokenized Treasuries (8–14% vs 4–5%) in exchange for credit risk — the borrower defaulting. The 2024–2026 market has matured into a functional segment of DeFi, though not without losses.
Centrifuge: the leader
Centrifuge tokenises real-world credit pools — trade receivables, invoice financing, real-estate-backed loans, asset-backed securities. Each pool has its own structure, issuer, and risk profile. MakerDAO relies on Centrifuge for a portion of its RWA collateral. ~$3B in active deals by 2026.
Maple Finance: institutional lending
Maple originally lent to crypto-native funds (and lost meaningfully in 2022 during the FTX/Alameda fallout). The relaunch (Maple 2.0) emphasises stronger underwriting and a mix of crypto-native and real-world borrowers. Yields 8–12% on stablecoin lending.
Goldfinch: emerging market credit
Goldfinch lends to vetted emerging-market financial institutions, who in turn lend to small businesses. The model — DeFi lenders backstopping traditional credit operations — is the cleanest "real economy" link in DeFi, though scale remains modest.
Yield vs risk
- 8–12% — Senior tranches on diversified credit pools (Centrifuge, Maple). Manageable risk for sophisticated lenders.
- 12–20% — Junior or first-loss tranches. Higher upside, real possibility of principal loss.
- Specialty pools — Yields vary; some Centrifuge pools have lost capital in past defaults.
What went wrong in 2022 — and what changed
Maple's 2022 losses came from under-collateralised loans to crypto funds that failed during the cycle. The relaunch emphasised secured lending, real-world borrowers, and tighter underwriting. Centrifuge's pool defaults during the same period taught the protocol to require deeper subordination. The 2026 market is meaningfully more conservative than the 2022 version.
When private credit fits a portfolio
Investors with patience (positions usually lock for months to years), willingness to do underwriting work, and tolerance for occasional principal loss. Not appropriate for short-term liquidity needs or for users uncomfortable with credit analysis.
See our RWA tokenization guide.
