Connecting AI Agents and Humans through lending.
Joined April 2025
- Tweets3.1K
- Following550
- Followers744
- Likes10.3K
Pinned Tweet
Introducing @agio_network:
The future of decentralized peer-to-peer lending.
Quick thought experiment:
You have $500k to lend.
Option A: Deposit in Kamino/Aave. Earn whatever the algorithm decides. Pray utilization stays high.
Option B: Set your own rate. Choose your own collateral. Match with a specific borrower. Keep the full spread.
Why would anyone with serious capital choose Option A?
Genuinely asking. Reply with your take 👇
UNPOPULAR OPINION:
LP pools are the riskiest way to earn yield in DeFi.
Impermanent loss is not "impermanent."
It's a slow bleed on your capital that most people don't notice until it's too late.
Lending gives you yield without betting against your own position.
Same capital. Less risk. More control.
LP pools vs. Lending
Here's what nobody tells you:
LP pool: you deposit 2 tokens. If one pumps, the pool rebalances and you end up with MORE of the losing token. You earned fees but lost on price. That's impermanent loss.
Lending: you deposit 1 token. It earns interest. When you withdraw, you get back your token + yield. No rebalancing. No IL. No surprises.
One is a yield strategy.
The other is a gamble disguised as one.
You provided $50k in liquidity to a SOL/USDC pool.
SOL pumped 40%.
Your position? Worth less than if you'd just held both tokens in your wallet.
You earned $300 in fees and lost $2,000 to impermanent loss.
This happens every single day across DeFi.
Lending protocols don't have this problem.
Your collateral stays your collateral.
Your yield is real yield.
Why lending is more programmable than LPing:
→ Custom risk parameters: set your own LTV, choose your collateral mix, define liquidation thresholds
→ Conditional logic: automated deleveraging, auto-repay on price triggers, scheduled interest payments
→ Composability: lending positions plug into other protocols — use your receipt token as collateral elsewhere
→ Isolation: one bad token doesn't drain your entire position
LP pools give you one dial: deposit more or withdraw.
Lending gives you an entire control panel.
LP pooling is like opening a currency exchange booth at the airport.
You profit when volume is high and prices are stable. But when the market moves? You're stuck buying what nobody wants and selling what everyone needs.
Lending is like being the bank. You set the terms. You earn interest. You get your money back.
Which business would you rather run?
A study of LP positions on major Solana DEXs showed that over 50% of LPs would have been better off just holding their tokens.
Fees earned < impermanent loss.
Meanwhile, lending protocols pay predictable interest based on utilization rates. No IL. No directional exposure. No praying the pair stays correlated.
DeFi lending isn't exciting. It's not degen. It's not sexy.
It's just better risk-adjusted yield for most people.
Lending Pools are like savings accounts at a bank:
You deposit. The bank decides what rate you earn.
You have no negotiating power.
The rate changes whenever they want.
P2P lending is like being the bank yourself:
You decide who gets a loan. You set the interest rate. You choose the collateral requirements.
You keep the full spread.
DeFi was supposed to disintermediate banks.
Lending pools just recreated them on-chain.
P2P actually delivers the promise.
The programmability gap between pools and P2P:
Lending pools (Kamino, Aave):
→ Fixed interest rate curves (you can't customize)
→ Protocol-defined collateral factors (you can't override)
→ Liquidation thresholds set globally (same for everyone)
→ One-size-fits-all risk parameters
P2P lending (Agio):
→ Custom rates per offer (you price your own risk)
→ Choose which collateral YOU accept (not what the protocol decides)
→ Negotiate duration, LTV, and terms per deal
→ Programmable conditions: auto-roll, rate triggers, partial fills
Pools give you a slot machine.
P2P gives you a spreadsheet.
Which one do you want managing your capital?
Risk profile comparison:
LP Pool:
→ Impermanent loss (guaranteed in volatile markets)
→ Smart contract risk on the DEX
→ Rug risk if one token in the pair goes to zero
→ Your loss is CORRELATED with market volatility
Lending:
→ Liquidation risk (but YOU control the LTV)
→ Smart contract risk on the protocol
→ Bad debt risk (mitigated by isolated markets)
→ Your risk is MANAGED by parameters you set
One gives you risk you can't control.
The other gives you risk you configure.
That's the difference.
How lending pools actually work (and why it matters)?
Kamino/Aave: Everyone deposits into one big pool. Borrowers take from the pool. Interest rates are set by an algorithm based on utilization.
→ High utilization = high rates for everyone
→ Low utilization = low rates for everyone
→ You have ZERO say in what rate you earn or pay
P2P lending: You make an offer.
Someone accepts it. The terms are yours.
→ You set the rate
→ You choose the collateral you accept
→ You pick the duration
One is a vending machine.
The other is a negotiation table.
Lending pools like Aave and Kamino
were a breakthrough.
But they have a fundamental design flaw:
- Lenders earn the AVERAGE rate.
- Borrowers pay the AVERAGE rate.
Nobody gets the best deal.
P2P lending fixes this. You set your rate.
You choose your counterparty. You keep the spread.
You deposited $100k USDC into a lending pool.
The dashboard shows 8% APY. Great.
Then utilization drops. The rate falls to 2.5%.
Your yield just got cut 70% overnight.
You had no warning, no control, and no say.
Meanwhile, a whale borrows $50M from the same pool and your rate barely moves because the algorithm smooths everything out.
In P2P lending, your rate is your rate.
Locked. Agreed. Non-negotiable.
No algorithm surprises. No utilization games.
Programmable lending:
→ You define your risk tolerance
→ Smart contracts enforce your parameters
→ Yield is predictable
Honest question for DeFi users:
Have you ever actually
calculated your LP returns
after impermanent loss?
Not the APY shown on the dashboard.
The REAL return.
Deposited $10k, earned $800 in fees,
lost $1,200 to IL, net: -$400.
Most LPs are paying the market
to provide liquidity without knowing it.
Lending doesn't lie to you.
The rate is the rate.
The yield is the yield.
Can you find @agio_network logo in the image below? 👇
Solana hosted the largest crypto hackathon ever
Colosseum Frontier, broken down 👇
📊 Relevant data:
• 2,857 submissions
• 19,040 registered participants
• 165 countries
• 45 judges
• +81% vs the previous Colosseum hackathon
• +166% vs the first Colosseum hackathon
🌍 Top 5 countries by registrations:
• India: 6,656
• Indonesia: 1,467
• Nigeria: 1,410
• Brazil: 1,088
• US: 962
👥 Participants breakdown:
• New users: 16,034
• Returning users: 3,006
- Students: 1,563
- Non-students: 17,477
After mapping every project logo, I wanted to look at where builders are actually spending their time 👇
🧩 Category breakdown:
• Consumer/Social: 607
• AI: 540
• DeFi: 515
• Infra: 279
• Payments/Fintech: 255
• Gaming: 148
• Security/Privacy: 147
• Verticals: 137
• RWA: 98
• Data/Governance: 83
• Wallets: 48
Consumer/Social, AI and DeFi alone account for 1,662 projects
That is more than 58% of all Frontier submissions
The top categories are all very close to users, user behavior and actual product usage
🏆 Winners will be announced around June 23, but the full project list already says a lot
With this many projects across so many verticals, there is a real chance that more than one of them will matter in the next phase of Solana
Good luck to all the builders.
The judges have a very hard job ahead.
Great breakdown by @xBartega on why isolated markets beat shared pools for lending.
The core insight is right:
risk should stay where it happens.
That's exactly why we built Agio as P2P.
@Kamino isolates risk into 26 markets.
@agio_network isolates risk into each individual loan.
Every loan on Agio is its own market, own collateral, own rate, own duration, agreed directly between lender and borrower. No curator, no shared pool, no inherited exposure.
If one loan goes bad, no other position on the platform is affected. Not because we manage risk well, but because the architecture doesn't allow it.
Same principle. Taken to its logical end.
Where they differ:
Pool → rates set by algorithm
Isolated market → rates shaped by curators
P2P → rates negotiated between two parties
Both models have tradeoffs:
- Pools give you instant liquidity.
- P2P gives you total control over your risk.
Different tools for different occasions.
Congratulations to the 95 participants of the @cloak_ag track at the FRONTIER global hackathon @solana, organized by @colosseum. Only 5 won an award, but all submitted projects are already winners, demonstrating the importance of onchain privacy.
@agio_network received an honorable mention.
The winners of the Cloak Track are here.
95 submissions. Builders from everywhere exploring what happens when privacy stops being a feature and becomes infrastructure.
The hardest part wasn’t finding good projects. It was choosing between them.
When there was a tie, we optimized for range: one winner per use case, to show how broad private payments can actually become.
🏆 1st: TENDR (@tendrdotbid )
Private RFP procurement where bids stay sealed until close, eliminating information leakage and preventing either side from cheating.
🥈 2nd: BIDO (@usebido )
Ad monetization for AI agents, where intent is privately detected and matched before the final purchase decision.
🥉 3rd: AEGIS (@aegis_ledger )
A privacy execution layer on top of Squads v4, routing multisig approvals through Cloak’s ZK infrastructure.
4th: TUMAR
A non-custodial family investment vault for the Kazakh diaspora, combining private payments with real-world asset exposure.
5th: TIRAI (@tiraiFrontier )
Private bounty payouts for security researchers, breaking the onchain link between deposits and withdrawals.
Honorable mentions: NORI, @DefundsFinance , @agio_network , @buildcoldstar , @usekixa , and GRAYBOX.
Massive respect to every team that shipped.
And to everyone who didn’t win this time: keep building. Cloak will continue supporting teams pushing privacy forward.
This is what the future of private payments looks like when builders take the problem seriously.
We are improving the P2P loan health of @agio_network.
- repayment is enforced
- borrowers/lenders can create offers with < 80% LTV
- bigger APY and long term loans need more LTV
- borrowers pay a 1% origination fee & 5% liquidation fee
We are tagged on the "Frontier Monkes"
by @zinnresearch
Many excellent projects are being built
by members of @MonkeDAO.
The Colosseum Frontier Hackathon had over 19,000 registrations 🤯
It is no surprise that the @MonkeDAO builder community showed up strong with dozens of submissions covering everything from RWAs and DeFi to agentic infrastructure and beyond.
Here are their submissions⬇️
