airKUNAIn one sentence: instead of selling an asset to fund airKUNA, the investor keeps it — and development is financed out of the yield that asset earns while it sits as collateral. The loan that goes into airKUNA's treasury is small relative to the collateral, so the carry pays it off over time. The result: a stake that repays itself, with SOL exposure retained and an allocation in airKUNA.
UserHarvestPayout event with a debt_repaid field — so "self-repaying" is not a metaphor but a number written on the chain. We decoded every one of those events: between 14 Apr and 27 Jul 2026, 1,280.73 USDC of yield was harvested, of which 1,033.01 USDC went straight to repaying users' debt. The engine ran again today.
On-chain proof — anyone can check it
Unlike the usual "trust me" pitch, the entire mechanism is public on Solana. Here are the key artefacts anyone can open in a block explorer — freshly measured on 27 Jul 2026.
| Artefact | Value (abbreviated) | Status |
|---|---|---|
| Hobba program (smart contract) | Hobbakk1…aE2i8qG ↗ Solscan | confirmed |
| Security audit | Ackee Blockchain Security, 15 Jun 2026 ↗ report | published |
| Self-repaying mechanism | 1,033.01 USDC of debt repaid from yield (sum of the debt_repaid field across all UserHarvestPayout events, 14 Apr – 27 Jul 2026) out of 1,280.73 USDC harvested in total | measured |
| Engine cadence | 699 harvests in 32 days (26 Jun – 27 Jul), 130 since 23 Jul, the latest today — without a single day off | running today |
| Yield integrations | Jupiter Lend jupr81…e3Bdzi, Kamino, Perena | on-chain |
| Live proof (our wallet) | 1 SOL collateral → 1.14 USDC loan ↗ wallet | executed |
| The engine services our position too | OperatorRepay covered our position account, 12 Jul 2026 ↗ tx | confirmed |
| Protocol fee | 15% of harvested yield — 0% on principal, deposit or loan. Read from the program's config account (profit_fee_bps=1500) ↗ Solscan and confirmed in every harvest since 14 Jul 2026 (before that: 0%, then briefly 25% — see the note below) | we measured it ourselves |
| Partner (origin) revenue share | Went live on-chain on 23 Jul 2026. The first partner channel (origin 1) over 5 days: 17.32 USDC of yield → 2.60 USDC of fees (exactly 15.000%) → 1.56 USDC to the partner (exactly 60.000% of the fee) | live, measured |
| Total deposits (TVL, 27 Jul 2026) | ≈ $345,544 · 3,836 SOL + 0.804 cbBTC · 84 wallets / 96 positions · 3,436 tx ↗ dashboard | early, but growing |
set_profit_fee) — that is a real risk, not a detail, which is why we re-measure it at every update.How it works — five steps
The only leg that touches airKUNA is the deposit into the treasury; everything else is the investor's position in self-custody, on an audited third-party protocol.
Live proof — our own transaction
So that it does not stay theoretical, we opened the smallest possible real position from our own wallet — the first, symbolic stake in the airKUNA flywheel — and let Hobba work:
Position #1 · founder wallet
6bwzpkSKSXbjVMBYMSdazEytkaCZatibRdmExpjSgCyA · open on SolscanHobba program:
Hobbakk1LmW2DhE4nAKnCeU1iK7V7pKHTq3ihaE2i8qG
Repayment progress — what actually happened, measured on-chain
This is the part we could previously only project. We have now decoded every Anchor event Hobba's program has ever emitted for our wallet and read the position account state from mainnet. The result is precise — and it forces one honest correction on us.
It turns out our position carries two separate debts, and the previous version of this page conflated them into one. Separated:
What the engine borrowed against our collateral
On 8 Jul the engine borrowed 45.038455 USDC against our 1 SOL of collateral and deployed it into a Perena vault (OperatorBorrowed). By 12 Jul it had returned 7.634980 USDC (OperatorRepaid). The position account state today shows deposited_to_vault = 37.403475 USDC — 45.038455 − 7.634980, exact to the micro-USDC. The decoded events and the live chain state agree to the last digit.
The 1.14 USDC that went into airKUNA's treasury
Here the number has not moved at all — but not because the capital is idle. For our wallet there is no UserRepaid and no UserHarvestPayout, because the harvest has not yet triggered: yield accrues in the vault share price and is only booked once harvest_vault crystallises it. Our account holds 34.713899 shares — the capital is deployed and earning.
Why it has not triggered yet: across all 60 positions there is a visible threshold of ~1 USDC — the first harvest of every small position lands between 1.0008 and 1.0035 USDC. At ~0.0084 USDC per day, our accrual today is ≈ 0.18 USDC, so we expect the first harvest around 2 Nov 2026. Repayment is therefore lumpy, not smooth.
accruing, not yet bookedThe complete log of our position — every event, ever
| Time (UTC) | Event | Amount | Meaning |
|---|---|---|---|
| 8 Jul 14:39:23 | UserInitialized | — | Hobba opens our position account |
| 8 Jul 14:39:52 | UserDeposited | 1.000000000 SOL | collateral enters, stays ours |
| 8 Jul 14:39:52 | UserBorrowed | 1.140000 USDC | our loan → airKUNA treasury |
| 8 Jul 15:02:44 | OperatorBorrowed | 45.038455 USDC | engine leverages the collateral and deploys into a vault |
| 12 Jul 14:00:00 | OperatorRepaid | 7.634980 USDC | repayment from yield |
| 13 Jul – 27 Jul | nothing | — | 15 days without a single event on our position |
Method: every Program data: log of the Hobba program was decoded via Anchor discriminators (sha256("event:<Name>")[0:8]) against IDL 0.1.4, and the account state was read from api.mainnet-beta.solana.com. No number on this page was read off a user interface — all of them are computed from the chain and reproducible.
How airKUNA builds its treasury — our own channel
So far you have seen one position — ours. But the real treasury machine is not our position, it is the channel. The two money flows are easy to confuse, so we separate them honestly:
| Flow | What happens | Effect on the treasury |
|---|---|---|
| Our own position (1 SOL) | Yield repays our debt; in the end we hold the collateral debt-free. | Deleveraging — it does not fill the treasury directly; it is skin-in-the-game proof. |
| airKUNA as a channel (origin) | Every user who enters through airKUNA is permanently tied to our channel; a share of their yield comes to us. | builds the treasury automatically, on every harvest. |
The fee is transparent and the same for everyone: Hobba takes 15% of the yield (not of the principal — measured on-chain as profit_fee_bps=1500). The contract has an "origin" revenue share built in: when a user enters through a partner channel, that 15% is split.
How the 15% yield fee is split
9.0% + 6.0% = 60% of the fee to the partner (fee_share_bps=6000). Revenue lands automatically in the treasury's USDC account on every harvest — with no manual "claim" step.The first 5 days of that channel (23–27 Jul 2026), decoded from the chain: 15 harvests · 17.3222 USDC of yield · 2.5983 USDC of fees (15.000% — on the nose) · 1.5590 USDC to the partner (60.000% of the fee) · 11.5696 USDC went to repaying users' debt.
The numbers — state of the protocol (honestly)
State as of 27 Jul 2026, computed from Hobba's on-chain events (our own Dune queries, not a reading off the interface). We deliberately show both what supports the thesis and what is a risk:
| Metric | 23 Jul 2026 | 27 Jul 2026 | Reading |
|---|---|---|---|
| Total deposits (TVL) | $338,446 | $345,544 | +2.1% in 4 days · +33% since 8 Jul |
| SOL collateral | 3,753.8 SOL | 3,836.4 SOL | +82.5 SOL of new deposits in 4 days |
| cbBTC collateral | 0.7616 | 0.8040 | +0.042 cbBTC |
| SOL price | $76.90 | $76.41 | important the price fell — all TVL growth comes from deposits, not from price |
| Collateral mix | 85% / 15% | SOL 84.8% · cbBTC 15.2% | SOL dominates, as the thesis assumes |
| Positions (wallet × collateral) | 90 | 96 | early stage +6 in 4 days |
| Distinct wallets | 79 | 84 | our own event count; Hobba's dashboard counts by a different definition |
| Transactions (total) | 3,281 | 3,436 | +155 · the protocol runs every day, without a break since 26 Jun |
| Yield harvests | — | 130 since 23 Jul | 699 in the last 32 days; the latest today |
| Debt repaid from yield | — | 1,033.01 USDC | cumulative out of 1,280.73 USDC of harvested yield |
| Concentration | top 3 wallets ≈ 72% of TVL (8 Jul) | high a small number of large depositors | |
A note on user counts: Hobba's public dashboard showed "87 users" on 23 Jul, while our count of the same events gives 79 wallets / 90 positions on that date. This is a different counting rule, not a contradiction — we publish our own, consistently measured definition and state what it is.
Risks — stated plainly
The architecture reduces some risks and accepts others. It does not remove them. Below is every material risk, how it is mitigated and what exposure remains.
| Risk | Mitigation | Residual exposure |
|---|---|---|
| SOL falls → liquidation | A conservative LTV keeps the liquidation price far below spot; Sonnar can reduce leverage before the threshold. | moderate a fast, deep drop can outrun the automation; oracles lag. |
| Young protocol / closed testing | Published audit (Ackee), on-chain verifiable code, small amounts for now. | high impact a new codebase; a bug or outage can lock or damage funds. |
| Low liquidity / concentration | Early stage; keep position size small relative to TVL. | high exiting a large position can be slow or expensive. |
| Yield falls below the borrow rate | The debt is tiny relative to the collateral, so the rate can rise many times over before carry turns negative. | low it only bites if interest exceeds the yield on the entire collateral. |
| Stablecoin (USDC) | Fully reserved, an established issuer within the MiCA framework. | moderate issuer or reserve failure is possible, though unlikely for a blue chip. |
| Smart contract / integrations | Audit, well-established integrated protocols (Kamino, Jupiter Lend, Perena), multisig over the treasury. | high impact exploits do happen; a tail risk that can wipe out the position. |
What the investor gets
- Retained SOL exposure — the principal stays the investor's, in self-custody, with full upside (and downside) of SOL.
- An allocation in airKUNA — the loan into the treasury funds development; the size and instrument of any stake or allocation is set by a separate, formal agreement, not by this page.
- A stake that repays itself — the debt is serviced out of collateral yield, not by selling the principal at the bottom of the market.
- Full verifiability — every claim is on-chain; nothing has to be taken on trust.
How you can do it too — from €1
This is what the path from curiosity to your own self-repaying position looks like. First, honestly: Hobba is in closed testing — you cannot sign up on your own through the web. Access goes through a whitelist, and we connect you with the Hobba team. That makes entry slower, but it also means things move in small, controlled steps.
Get in touch
By email or Telegram. We agree the framework and connect you with the Hobba team for the whitelist (closed testing).
Prepare a wallet
Solflare (or another Solana wallet) and SOL you intend to hold anyway. The asset stays yours throughout.
Deposit and borrow
SOL goes into Hobba as collateral; you draw a small USDC loan — from ≈ €1 upwards, at a conservative LTV.
The loan funds airKUNA
The funds go into the treasury under a separate, formal agreement defining your stake or allocation.
Yield repays the debt
The engine harvests and repays on its own. When the debt reaches 0, the SOL is free, and your stake in airKUNA remains yours.