ZCHF (Frankencoin) — Risk Report

Elevated risk · 4.0/10 Stablecoin · Swiss francs · Ethereum + 7 chains · No issuer · ~CHF 34M · verified 2026-08-21

This token is pegged to the Swiss franc, not the dollar. If you price it in USD you will see it swing 10% or more and conclude it is broken. That is the dollar moving, not ZCHF. Every peg figure below is measured ZCHF against CHF. For a dollar-denominated portfolio, ZCHF is a currency position first and a credit position second, and the currency call is not what this report scores.

Summary

Frankencoin is a borrowing protocol that issues Swiss francs on-chain. Anyone can lock up collateral — bitcoin, staked ether, gold, tokenized shares — and mint ZCHF against it, the same basic shape as MakerDAO or Curve’s crvUSD. What makes it genuinely unusual is that it uses no price oracle anywhere. There is no feed telling the protocol what your collateral is worth. Instead, the borrower declares their own liquidation price when they open a position, and anyone who thinks that price is a lie can challenge it by putting up the same collateral and forcing an auction. Price discovery and liquidation are the same event.

There is also no issuer. A published legal assessment by LEXR Law Switzerland concludes ZCHF is generated by a decentralised protocol “not centrally organised or controlled by any legal entity.” We checked that from the other direction: owner() and admin() both revert on the token contract. There is no admin key, no freeze function, no custodian and no company that can be subpoenaed. New minting rights arrive only through a 14-day public application that any 2% of governance-token holders can veto. That is the entire control surface, and it is unusually clean.

The problem is not the design — it is what is behind it and how much of it there is. About CHF 34.2M of ZCHF is outstanding — CHF 30.3M on Ethereum plus CHF 3.9M bridged, mostly to Base and Gnosis — against CHF 33.5M of borrower debt, and roughly 31% of that debt is collateralised by shares in two private Swiss companies. Those shares do trade, in small size, on an on-chain OTC venue — but the larger of them, the tokenized share register of an ICT company called Boss Info AG, is marked above the prices it actually changes hands at, and its issuer’s audited consolidated accounts show CHF 362,014 of total equity behind a mark implying an CHF 86M valuation. The first-loss capital standing behind the whole system is CHF 3.81M. Three borrowers account for 77% of all debt; the largest holder of Boss Info shares — 52.8% of the company — is the same person borrowing against them; and the share contract is one where a single ordinary key can destroy the exact collateral backing four million francs of debt. None of it is hidden — all of it is readable on-chain, which is more than most stablecoins can say — but it is a concentrated credit view on a handful of counterparties, not a cash-equivalent.

At a glance


Backing	Borrower collateral: bitcoin (cbBTC/WBTC) 37%, tokenized Swiss private shares 31%, staked ETH 16%, tokenized S&P 500 9%, gold 4%. Plus 2.3% issued against CHFAU, a BaFin-licensed euro-style franc stablecoin.
How to redeem	You can’t. There is no issuer and no redemption right. You exit by selling on a DEX or an exchange. The peg holds on arbitrage and borrower incentives only.
Liquidity	Thin. About CHF 400k clears at roughly 0.5% slippage on the main Uniswap pool, about CHF 500k across two venues at roughly 0.6% blended. One pool carries about 80% of it.
Yield	~1% on the protocol’s own savings module, which holds 39% of all ZCHF. Funded by borrower interest with roughly a 1% margin.
Admin & custody	No admin key on the token — verified, `owner()` reverts. Governance is a 14-day minter application vetoable by 2% of FPS holders, which costs about CHF 228k to exercise.
Regulated?	No, and deliberately so. Not e-money, not a deposit, no licence, no deposit insurance. The LEXR opinion classifies it as a payment token with no identifiable issuer.
Biggest risk	31% of the debt is backed by tokenized private-company shares — register entries one key can mint or burn, held in blocks worth 22% and 36% of the companies themselves. The larger is marked above the prices it trades at, its issuer’s audited consolidated equity is CHF 362,014, and its majority shareholder is the borrower. Only CHF 3.81M of shared first-loss capital sits behind the system.

1 · Stability — 6.5

The record is good and it is measured, not asserted. Against the Swiss franc over the trailing year ZCHF traded between 0.9827 and 1.0151 — a worst deviation of −1.73% — and closed below 0.99 on five days. That is a tighter band than crvUSD has held, on one-sixth the supply, and it has been maintained for three years with no redemption backstop at all. Nothing forces ZCHF back to par except arbitrage, the cost of borrowing, and the savings rate. Holding that band without a redeem button is the single most impressive fact about the asset.

The oracle-free design deserves credit too. Because there is no price feed, there is no feed to manipulate — an attack surface that has cost other lending protocols nine figures simply does not exist here. Liquidation runs as a two-stage auction: a challenger posts collateral, and for a set window anyone can buy that collateral at the borrower’s stated liquidation price, which proves the position sound and cancels the challenge. If nobody does, a Dutch auction opens at that price and declines to zero, and the first bidder takes the collateral. Losses hit the borrower’s own reserve contribution first, then FPS equity, then other borrowers’ reserves pro rata.

What keeps this off a 7 is that the auction starts from a price the borrower chose, not a market price, and borrowers set it as high as the rules allow. Measured across all 71 open positions, the book is collateralised 1.18x at those liquidation prices — and on roughly 78% of the debt the auction can fall less than 7% before proceeds stop covering the debt. The buffer is really the gap between the declared liquidation price and the market price, which is comfortable today. It is not a structural cushion.

2 · Backing — 4.0

Start with the genuine strength: this is one of the very few stablecoins whose backing can be reconstructed entirely from public blockchain calls, to the franc, with no attestation, auditor or issuer dashboard in the loop. We did exactly that. Reserves in the equity contract total CHF 12.44M against CHF 30.3M on mainnet, and the internal accounting reconciles exactly. No fiat-backed stablecoin in our coverage can be verified this way.

Now the collateral. Roughly 31% of borrower debt is secured on shares in two private Swiss companies — BOSS (Boss Info AG, an ICT and ERP integrator) and DQTS (draggable quitt.shares) — both tokenized through Aktionariat under Switzerland’s 2021 DLT act, where the token is the entry in the company’s share register. Boss Info tokenized 100% of its capital in 2021: 10.65 million shares, one share to one indivisible token. This is legitimate and well-documented, and an oracle-free auction is arguably a better way to liquidate an illiquid share than a price feed would be. But it means the collateral is an issuer-controlled register entry, not a bearer asset. We verified on-chain what that implies: the BOSS token is an upgradeable proxy whose admin is a single ordinary wallet, and a second ordinary wallet can mint new shares with no cap and burn existing ones out of any address. Simulated against the live position backing 4,000,000 ZCHF, the burn call succeeds for exactly that position’s 800,000-share balance. Under Swiss corporate law that control is required — an issuer must be able to record capital increases and cancellations — but Frankencoin’s contracts hold nothing that protects against it.

Concentration compounds it, and in a specific way. Three borrowers hold 77% of all debt. The largest holder of BOSS — 5.6 million shares, 52.8% of the company — is the same address that owns three of the five BOSS positions and carries 24.7% of all ZCHF debt; between his wallet and his posted collateral he controls 75% of Boss Info AG. Collateral value and borrower solvency are therefore one variable rather than two, and a forced sale would come precisely when the company’s controlling shareholder is distressed. The blocks are large relative to the companies too: Frankencoin’s contracts hold 22% of Boss Info AG and 36% of quitt.shares.

And the mark all of those figures rest on does not reconcile. The shares do trade — an on-chain OTC venue printed CHF 107,000 of BOSS across 14 trades in six weeks — so this is not a market that fails to exist. But the scale is the point: that is about CHF 2,500 a day, against a BOSS position of 2.4 million shares. At the observed rate it would take roughly seven years of flow just to raise the CHF 6.9M of debt those shares secure. DQTS looks deeper at first glance — CHF 2.6M over five months — until you notice 99% of it was a single day at a pre-agreed price of exactly 8.000; strip that placement and five months of trading is CHF 25,000. And BOSS printed at 6.90–8.20, clustering 7.30–7.55, against the 8.072 mark Frankencoin displays, leaving that position carried roughly CHF 1.6–2.2M above realised prices. And Boss Info AG’s audited consolidated accounts show CHF 362,014 of total equity, with 2025 net income of CHF 57,318 after a CHF 3.99M loss the year before — against a mark implying an CHF 86M market capitalisation. Frankencoin’s stake is carried at CHF 19.19M; the attributable consolidated book value is about CHF 80,700. Operating cash flow is healthy and a services business can legitimately trade well above book, so this is not a valuation call — but nothing in the audited record anchors a gap of that size. That, not illiquidity, is why this axis is 4.0.

Separately, 2.3% of supply is now issued against CHFAU, a BaFin-licensed franc stablecoin from the Deutsche Bank/DWS/Galaxy venture AllUnity — better credit than a private share, but it inserts a regulated, freezable counterparty into a system whose whole proposition is not having one, and its approved ceiling is already 10M, a third of supply, with no further vote required. In fairness to the issuers: both share positions carry the protocol’s top-tier 40% borrower reserve, DQTS is marked below its own prints, an atomic on-chain settlement venue exists at all, and Swiss company law limits how much of its own stock Boss Info could buy back even if it wanted to. The realistic exit for a position this size is a negotiated block sale or a corporate transaction, not a market. The score would move up if BOSS prints converge to its mark or the tokenized-equity share falls; down if the CHFAU ceiling is used, if that share passes ~45%, or if any issuer key is exercised against a live position.

3 · Liquidity & Exit — 5.5

These are measured executable quotes, not pool TVL. On the main Uniswap V3 ZCHF/USDT pool, 100k ZCHF costs about 18bp, 400k costs 51bp, and at roughly 450k the range is exhausted and slippage jumps past 7%. A Curve pool against crvUSD adds 50k at 39bp and 200k at 378bp. Blended, about CHF 500k clears same-day at roughly 61bp, with a further ~$200k/day of centralised volume and a small euro pair on Gnosis that is the only on-chain route to a European fiat pair.

Three things cap the score. One venue holds ~80% of that capacity on about $1.3M of TVL, so this is concentration, not depth. The depth is in tight concentrated ranges that a single LP can withdraw in one block — re-read five hours apart, the deep Curve rung moved from −523bp to −378bp on a 7% shift in pool balances. And only about 20% of ZCHF is genuinely floating: 39% sits in the savings module and 41% is protocol reserve, which is why the tradeable book is thin relative to the market cap. Quote before you size; do not carry a number from this report.

One dated item to watch: a governance application filed 2026-08-20 would let anyone open a Uniswap position posting only the dollar side, with the ZCHF side minted fresh — up to CHF 2.5M into that same dominant pool. Its veto window closes around 2026-09-03. If it goes live it could multiply the main venue’s depth, but that would be protocol-manufactured liquidity with an expiry date, not organic LP interest.

4 · Dependencies — 6.0

What ZCHF passes through to is its collateral book — 71 positions against CHF 33.52M of debt. ⚠️ The finding is that there is no dominant counterparty, which on this coverage is unusual enough to be worth stating as the headline: BTC family 37.5% (cbBTC 22.4%, WBTC 15.1%), Swiss private company shares 31.0% (BOSS 20.5%, DQTS 10.5%, issued via Aktionariat), ETH family 17.6% (LsETH 9.8%, wstETH 6.3%, WETH 1.5%), SPYon 8.7% (Ondo tokenized S&P 500), gold 4.3%, misc 0.9% — plus CHFAU at 2.3% of supply, an AllUnity fiat claim under a BaFin e-money licence. The largest single collateral is 22.4%, across about fourteen types and four asset classes. Set against the other assets migrated this week — apxUSD at 86.7% netted STRC, sUSDat at 99% STRC, USDat at 100% PYUSDx — this is a genuinely diversified dependency set, and axis 4 is the axis that should say so.

⚠️ Held below 7 by two things, and the first is a distinction this axis exists to make. About 39.7% of debt is collateralised by tokenized equity, and BOSS plus DQTS — 31.0% — are Swiss private shares with essentially no on-chain market. That book is diversified by issuer and concentrated by liquidity class, and reading the first as covering the second is precisely the error a dependency axis is supposed to catch. ⚠️ DQTS’s entire 10.5% is a single position at 1.34x. Second, the CHFAU bridge puts a regulated, freezable, centralised counterparty inside a system whose whole proposition is that it has none — 2.3% today, and the number to watch is the ceiling rather than the balance.

5 · Contract & Admin — 7.5

⚠️ Re-verified on-chain 2026-09-09: owner() and admin() both revert on the token contract 0xB58E61C3…21cB, while totalSupply() answers normally — so the reverts are the absence of those functions, not an unreachable contract. There is no admin key to compromise and none to subpoena. Three audit engagements sit behind it (ChainSecurity twice, Code4rena, BlockBite).

⚠️ But adminless is not change-less, and this is why the axis is 7.5 rather than higher. The system still changes — it changes by silence. A proposed new minter passes BY DEFAULT after 14 days unless it is vetoed, so the security model is not “nobody can change this” but “somebody must object in time.” ⚠️ And the veto threshold is 2% of FPS — about CHF 228k — which is the security model and the concentration risk in the same number. Cheap enough that a motivated holder can defend the system; cheap enough that a motivated holder can also hold it up. A governance surface with no keys still has a quorum, and this one’s quorum is small.

6 · Issuer — 5.5

There is no issuer, and that is both the score’s floor and its ceiling. What it buys is real: no bank, no custodian, no attestation to trust, no freeze function, no bankruptcy estate, and a backing analysis that is fully verifiable from public data. Three years live with no depeg incident and no reported bad-debt event. The control surface itself — what can be changed, by whom, and how fast — is scored on Contract & Admin above rather than here, so this axis prices the absence of an entity and not the absence of a key.

What it costs is equally real: no redemption right against anyone, no recourse, no insurance, no regulator, and nobody obliged to make a shortfall good. Loss absorption stops at whatever FPS holders have voluntarily put up, which is CHF 3.81M — and note that this is the shared layer, smaller than the CHF 12.44M sitting in the reserve contract, most of which is borrowers’ own money already earmarked against their own debt. Governance is also small enough to be cheap in both directions: the 2% veto threshold is about CHF 228k, so blocking a dangerous new minter is affordable, and so is blocking a good one. On a governance base this size, the veto is simultaneously the security model and the concentration risk.

Bottom line

Frankencoin is the credible on-chain Swiss franc, and its engineering is better than its score. Oracle-free, adminless, transparent to the franc, with a three-year peg record tighter than much larger competitors and no issuer to trust. If you want CHF exposure on-chain, this is the serious option and there is no close second.

What you are actually underwriting is narrower than the design suggests: CHF 3.81M of shared first-loss capital behind CHF 33.5M of debt, roughly a third of it secured on private Swiss company shares that are register entries under issuer key control, with 77% of the borrowing in three hands. It is a concentrated credit position on a small number of counterparties. Size it that way — and for a dollar portfolio, remember the franc call dominates everything in this report.

Not for: anyone who needs to redeem at par, needs deposit protection, wants to move more than about half a million francs quickly, or is treating this as a dollar cash equivalent.

Watch items

Whether the UniswapAmplifier minter goes live after ~2026-09-03 — it would reshape the main pool.
CHFAU-backed supply rising toward its approved 10M ceiling (a third of supply, no vote needed).
Tokenized private-share collateral rising above ~45% of debt.
Any change to the BOSS token’s owner or proxy admin, or a rise in its share count.
Boss Info OTC prints drifting further below the 8.07 mark its collateral is carried at.
A DQTS drag-along vote opening — a 75% vote can force conversion of the shares held inside the position.
FPS equity falling toward CHF 2M.
ZCHF closing below 0.98 against the franc — never against the dollar.

This assessment is based on publicly available documentation and on-chain analysis only. It does not draw on private disclosures from any protocol, issuer or tokenization agent named here, and the observations about token contract permissions describe capabilities readable on-chain, not any suggestion that they have been misused. If something here is wrong, write to info@tidresearch.com and we will correct it in the next revision and credit the source.

2026-09-09 — Dependencies 6.0 and Contract & Admin 7.5 added; no existing score moves. Both axes were genuinely absent on this page rather than mislabelled. ⚠️ Contract & Admin is a re-scope off the Issuer row, where the control surface was being priced — the owner()/admin() reverts and the audit stack are moved rather than copied, so nothing is counted twice. Re-verified on-chain: both calls revert while totalSupply() answers, so the reverts are real absences. Held at 7.5 because a new minter passes by default after 14 days unless vetoed at 2% of FPS (about CHF 228k) — adminless is not change-less, and that threshold is the security model and the concentration risk at once. Dependencies 6.0 on a collateral book with no dominant counterparty (largest single collateral 22.4% across about fourteen types), held below 7 because 31.0% is Swiss private shares with essentially no on-chain market — diversified by issuer, concentrated by liquidity class.

Revision history: 2026-08-21 — first publication at overall 4.5. 2026-08-21, revised — supply corrected to CHF 34.2M; backing 4.5 → 4.0 and overall 4.5 → 4.0 after tracing the tokenized-share collateral to its trading venues and reading the issuer’s audited accounts. Overall 4.0 (peg 6.5 / backing 4.0 / liquidity 5.5 / issuer 5.5).