Lendi replaces institutional judgment with programmable guarantees. Borrowers prove eligibility without exposing financial data. Lenders deploy capital with automated default protection. Neither side trusts the other — both trust the protocol.
How it works: Informal workers prove their income using Fully Homomorphic Encryption. The lender receives a yes/no answer — the worker reveals nothing. Every loan is automatically insured via an encrypted protection pool. Lendi is the only system in LATAM that delivers strict guarantees to both borrowers and lenders simultaneously.
Traditional lending forces a choice: either borrowers expose everything, or lenders deploy capital unprotected. Lendi eliminates that trade-off.
The guarantee to borrowers: Prove eligibility without revealing your income, clients, or transaction history. The lender receives only a yes/no answer — computed on encrypted data, never decrypted.
The guarantee to lenders: Deploy capital with automated insurance against default. Every loan is backed by an encrypted protection pool. Risk assessment happens on ciphertexts — borrower privacy is never compromised.
The architecture: Neither side trusts the other. Both trust the protocol. The answer comes from math, and it's identical for every borrower, every time.
1. Exclusion by design 47 million workers in LATAM earn real income in stablecoins but have no payslip. Traditional lenders and P2P platforms require full financial exposure to underwrite them. Most workers refuse. They stay excluded.
2. The trust trade-off problem Traditional lending forces one side to absorb all the risk: borrowers expose everything, or lenders deploy capital unprotected. On-chain income data exists and is verifiable, but sharing it means revealing everything. No existing solution delivers guarantees to both sides simultaneously.
3. The informal economy is going on-chain Stablecoin adoption in LATAM grew 63% YoY. Workers are already paid in USDC. The data is there. The privacy layer is not.
1. Encrypted income accumulation
Workers register income from on-chain stablecoin transfers (Privara / direct USDC). Each amount is encrypted client-side and stored as euint64 on Arbitrum. Nobody — including the protocol — can read the number.
2. Privacy-preserving credit proof
When a lender requests verification, the contract runs FHE.gte(workerIncome, threshold) on two ciphertexts. Neither value is decrypted. The lender receives only an ebool: qualifies or not.
3. Two-sided trust infrastructure via encrypted risk pool Every loan is automatically backed by a ProtectionPool that insures lenders against default — using FHE to calculate premiums and payouts without exposing individual worker risk scores. This delivers the core guarantee: lenders deploy capital with protection, borrowers transact without exposure. Neither side has to trust the other — both trust the protocol. No competitor offers this symmetry.
4. Local AI financial advisor Workers can chat with an AI advisor that sees their decrypted income — but only in their device's RAM, via WebLLM. No server is ever called. Not ours, not OpenAI's.
Why now FHE on EVM became production-ready in 2025. Fhenix CoFHE is live on Arbitrum Sepolia. The window for first-mover advantage is open — and it closes as protocols ossify.
Why us
- We are building from Colombia — one of the top informal economies in LATAM, with 19M+ digital wallet users (Nequi alone)
- We have a co-build agreement with ReinieraOS (Privara ecosystem) for payment rails, escrow infrastructure, and VC introductions
- We are one of two teams in this buildathon addressing consumer use cases outside of DeFi/trading
| Size | Basis | |
|---|---|---|
| TAM | $12B | P2P + microlending market in LATAM (2025) |
| SAM | $800M | Colombia + Mexico informal worker lending |
| SOM (Year 1) | $2M | 5,000 loans, average $400/loan |
Realistic starting point: The $2M SOM assumes 5,000 loans in year one at $400 average loan size.
Go-to-market approach:
- Waves 1-4: Start with individual users and small microlending organizations (community credit circles, P2P lenders)
- Why small first: Validates PMF in days/weeks, not months. Avoids long enterprise sales cycles without warm intros.
- Wave 5+: Once we have proven traction (50-100+ successful loans), approach larger fintech partners (Nequi, Rappi) with real usage data as leverage
- Key insight: Large tech tends to come to you once you build momentum with smaller segments that scale fast
| Privacy | Works without payslip | On-chain | Lender protection | |
|---|---|---|---|---|
| Lendi | ✅ FHE | ✅ | ✅ | ✅ via pool |
| Kueski (MX) | ❌ | ✅ | ❌ | ❌ |
| Addi (CO) | ❌ | Partial | ❌ | ❌ |
| Juancho te Presta (CO) | ❌ | ✅ | ❌ | ❌ |
| Bloom Credit | ❌ | Partial | ❌ | ❌ |
| ConfidentialCredit (Fhenix) | ✅ FHE | ✅ | ✅ | ❌ |
Honest note: Kueski, Addi, and Juancho te Presta already serve informal workers. They have distribution, trust, and a working product. Our advantage is not that we serve informal workers — it's that we've eliminated the trust trade-off entirely.
Two-sided trust infrastructure is our moat. Most lending products solve for one side and ask the other to accept the trade-off. We deliver strict guarantees to both simultaneously. No competitor — traditional or crypto-native — offers insurance against default while keeping borrower income encrypted. This is a regulatory advantage (no data liability), a capital efficiency advantage (lenders deploy with protection), and a technical advantage (only possible with FHE). We are not faster or simpler than Addi. We are more defensible.
| Quarter | Focus | Target |
|---|---|---|
| Q2 2026 | Ship core FHE contracts + demo | Arbitrum Sepolia deployment |
| Q2 2026 | AI advisor + Privara integration | First 50-100 individual users (Colombia) |
| Q3 2026 | Direct-to-consumer validation | Small microlending orgs + P2P lenders |
| Q3 2026 | Prove traction, then approach large tech | If momentum is strong, explore fintech partnerships |
| Q4 2026 | Mainnet (Fhenix, scheduled autumn 2026) | First 1,000 loans |
Strategy: Start small and fast. Individual users and community lenders validate PMF in days, not months. Once we have real traction and usage data, larger fintech partners (Nequi, Rappi) become warm intros instead of cold outreach.
Note: Fhenix mainnet is not available until autumn 2026. We start on a public chain to validate product-market fit, then plug in privacy at the protocol level with Fhenix support.
Fhenix CoFHE · @cofhe/sdk · Privara + ReinieraOS · Arbitrum Sepolia · WebLLM · React
Wave 1 — Fhenix Privacy-by-Design Buildathon · March 2026