A single key, a $700M token: inside Venice’s uncapped mint

Venice Token (VVV) is a 20-line Solmate ERC-20 whose entire supply authority resolves to one line of code:

**function mint(address to, uint256 amount) external onlyOwner.** No hard cap. No timelock. No multisig visible on-chain. No public audit. This is the architecture currently supporting a ~$400M market cap and ~$700M fully diluted valuation. The finding matters because institutional capital cannot diligence what a single key can redefine overnight — not because Erik Voorhees, Teana Baker-Taylor, or the Venice team have shown any indication of malicious intent, but because the structural risk is indistinguishable, to an allocator's risk model, from one where they might. Venice launched on Base in January 2025, merged an AI-infrastructure thesis with genuine privacy product-market fit, and rode a 320% year-on-year rally. The contract architecture that underpins the rally, however, sits in a risk category of one among the nine decentralized-AI tokens surveyed for this report. What follows is the on-chain record, the peer comparison, and the historical precedents that explain why that outlier status should be priced. The contract, verbatim

The contract, verbatim

The Venice token contract lives at 0xacfE6019Ed1A7Dc6f7B508C02d1b04ec88cC21bf on Base. It is verified on BaseScan with a green "Exact Match" badge, compiled with solc v0.8.26+commit.8a97fa7a, 200-run optimization, EVM target paris, MIT license, Base source distributed across three files: Venice.sol, ERC20.sol, and Owned.sol, with dependencies on solmate 6.8.0. It is not a proxy — the bytecode at the address is the implementation, not a forwarder, and there is no upgrade path.

The entirety of the project-specific Solidity is reproduced here exactly as it appears on BaseScan: solidity

// SPDX-License-Identifier: MIT
pragma solidity =0.8.26;

import {ERC20} from "solmate/src/tokens/ERC20.sol";
import {Owned} from "solmate/src/auth/Owned.sol";

contract Venice is ERC20, Owned {
    constructor(address treasury)
        ERC20("Venice Token", "VVV", 18)
        Owned(msg.sender)
    {
        _mint(treasury, 100_000_000 * 10**decimals);
    }

    function mint(address to, uint256 amount) external onlyOwner {
        _mint(to, amount);
    }
}

The onlyOwner modifier resolves, via solmate's Owned, to a one-line check: require(msg.sender == owner, "UNAUTHORIZED"). There is no cap parameter, no MAX_SUPPLY constant, no require(totalSupply() + amount <= …) invariant, no TimelockController, no Governor, no per-block or per-year emission limit, no pause, no blacklist, no fee switch, no two-step ownership transfer. The complete set of external state-changing functions on the contract is five: the standard ERC-20 triad (approve, transfer, transferFrom), EIP-2612 permit, and the two privileged functions mint(address,uint256) and transferOwnership(address).

There is also no public burn() function. Solmate's underlying ERC20.sol exposes only an internal _burn, which the Venice contract does not wrap. This is a material, underappreciated detail — covered below.

BaseScan’s audit field, quoted verbatim, reads: “No Contract Security Audit Submitted — Submit Audit Here.” No entry on CertiK Skynet, OpenZeppelin’s audit index, Trail of Bits’ publications GitHub, Zellic’s public portfolio, Spearbit’s portfolio, Quantstamp, Halborn, PeckShield, Hacken, or Sherlock references Venice, VVV, or the contract address. Venice’s own launch blog includes the word “Audit” as a bullet point with no linked report, no firm named, no date, and no scope. An Uphold asset page describes “successful audits by well-regarded third-party entities”; UpholdUphold this claim is not substantiated by any primary source and is most plausibly boilerplate. A Skywork.ai review from September 2025 is blunt: “We did not find a third-party privacy audit… as of September 2025, I found insufficient data on any independent, formal security/privacy audit of Venice’s claims.”

Who holds the key

The contract creator — labeled by BaseScan as “Venice: Deployer” — is the address 0xc9c88391e50eeadb43647fac514fa26f8dfd7e7f, and the genesis 100 million VVV were minted to the constructor-argument treasury at 0x2D8CB8DC596daD0e1E34E2042E7ae6Df93B11524. Neither address carries a Safe/Gnosis multisig label on BaseScan. Major protocols that custody mint authority in a multisig almost invariably get that fact labeled on the explorer and disclosed in their tokenomics documentation. Venice's public materials do not mention any multisig, any timelock, or any governance contract controlling VVV. The default inference — pending direct on-chain reconfirmation by readers — is that the mint key is an externally owned account, a single private key, holding unlimited authority to issue VVV.

The key has been exercised. Total supply on BaseScan grew from 100,000,000 VVV at genesis to 113,043,350.57 VVV by early April 2026 — a ~13M expansion consistent with Venice’s publicly stated emissions. The owner is live, and mint() has been called.

The burn that is not a burn

Venice markets VVV as a “long-term deflationary capital asset” and publishes monthly “buy and burn” updates. On-chain, the 0x0000…0000 null address holds 33,539,739.93 VVV — the mechanical destination of Venice's buy-and-burn program, constituting roughly 33% of the original genesis supply. The problem, visible only by reading the contract, is that these tokens were sent to the null address via ordinary transfer() rather than consumed via the internal _burn() — because no public burn function exists. They sit as stranded balances that still count toward totalSupply(). Third-party aggregators are inconsistent about this: CoinGecko computes a derived "Est. Total Supply" of 79.16M by subtracting the null-address balance, while the contract itself still reports ~113M outstanding. The "42% of supply burned" figure circulated in favorable media (Phemex, OneKey, BingX, CMC AI) is arithmetically inflated — it divides the null-address balance by the reduced derived supply, not by the 100M genesis against which the claim should be made.

This is a discrepancy between the contract and the marketing, not a fraud. But it is the kind of discrepancy an audit would flag in its first page.

What third parties already disclose

CoinGecko displays a GoPlus-sourced warning on every currency subpage of the VVV listing. The exact text, reproduced identically across the VVV/USD, VVV/EUR, VVV/ETH and every other fiat conversion page:

"According to GoPlus, the contract creator can make changes to the token contract such as disabling sells, changing fees, minting, transferring tokens etc. Exercise caution."

CoinGecko also displays Max Supply as the literal infinity symbol (∞). The claim about disabling sells and changing fees is technically inaccurate for this specific contract — there are no fee toggles in the source — but the mint, transfer, and ownership elements are accurately characterized.

CryptoRank states verbatim: “The current circulating supply of Venice Token is 43.28 Million tokens, and the maximum supply of VVV is unlimited.” CryptoRank declines to display a market cap for VVV — an editorial policy consequence of the unlimited-supply designation. CoinMarketCap leaves the “Max. supply” field blank. GeckoTerminal’s pool-level assessment scores the VVV/WETH Aerodrome pool at 92.66 out of 100 and correctly notes “no proxy was found in the contract” and “no indication of a honeypot” — the contract is clean on those specific attack vectors, which is a different question from mint authority.

The gap between what Venice markets (“deflationary capital asset”) and what third-party risk engines display (“max supply ∞”, “contract creator can mint”) is the gap in which the structural thesis lives.

The tokenomics claims versus the record

The author’s starting framework cited an emissions schedule of “10M → 8M → 6M per year.” That figure requires correction. Venice’s original January 27, 2025 launch blog and The Block’s reporting from the same day both stated 14M VVV per year — a 14% inflation rate. The current Venice Help Centre now retroactively describes emissions as “10M VVV per year at token launch, reduced to 8M, then to 6M in February 2026.” This is a material, undisclosed rewriting of the original tokenomics, executed unilaterally — because there is no governance layer that could execute it any other way. The fact that the direction of revision was tighter, not looser, does not change the structural point: the emission schedule is a social commitment, not an on-chain invariant, and the owner key is the sole authority that can redefine it.

The allocation itself — 50% airdrop, 35% Venice.ai (which Messari splits 25% company / 10% team), 10% incentive fund, 5% liquidity — verifies across Venice’s blog, The Block, and Messari’s token-unlocks page (which shows monthly team vests of ~326,100 VVV on the 27th). The December 10, 2025 first buy-and-burn execution is verifiable on Messari’s project timeline, as is the DIEM launch on August 20, 2025. DIEM, importantly, does not burn VVV. The mechanism locks staked VVV (sVVV) to mint DIEM, where each DIEM yields $1/day of Venice API credit indefinitely; burning DIEM unlocks the underlying sVVV. The only VVV destruction path remains the revenue-funded buy-and-burn whose size depends on Venice’s unaudited revenue reporting.

Where VVV sits among its peers

A matrix of nine decentralized-AI tokens surfaces the outlier clearly.

Among the nine peers, two (TAO, AKT) have no owner mint function at all. Worldcoin’s contract explicitly overrides renounceOwnership() to revert World — the owner cannot walk away, but neither can the owner mint for fifteen contract-enforced years. Render, UNI, and COMP either burned the upgrade authority, bound mint to a timelock and a cap, or pre-minted the entire supply. Fetch.ai, io.net, Akash, and Ocean Protocol retain mint authority but bind it to an algorithmically defined or governance-gated schedule. Only Grass approaches VVV's risk profile — no public audit, foundation-held mint — but Grass's supply cap is enforced at the Solana mint-account level (supply-changing events are publicly observable) and the token inherits Solana's standard SPL primitives rather than custom Solidity.

VVV is the only token in the cohort combining (1) unbounded onlyOwner mint, (2) custom Solidity code, (3) no published audit, (4) no on-chain cap, (5) no timelock, and (6) no multisig disclosed on-chain. That is a risk category of one.

What the historical record says happens next

Markets do price owner-privileged architecture — but only after it is surfaced in a credible, high-visibility forum. The catalysts and the magnitudes are worth cataloging, because they establish a base rate.

On April 10, 2026 — two weeks before this report — Covenant AI founder Sam Dare publicly exited the Bittensor ecosystem via The Block, alleging that Jacob Steeves maintains unilateral control over TAO’s triumvirate multisig. TAO dropped 15% in two hours, from $338 to $285. PANews framed the move precisely: “a repricing of Bittensor’s ‘governance risk discount’ by institutional funds.” TAO is a fair-launched, hard-capped, algorithmically-emitted token — and yet an accusation of centralized governance, with no contract-level exposure, was enough to wipe $500M of market cap in a session.

SushiSwap, September 2020. Chef Nomi’s single-key admin of the MasterChef contract enabled a $14M dev-pool swap that halved SUSHI in 24 hours and bled 35% off Uniswap TVL. Only community mobilization — forcing a transfer to a 6-of-9 Gnosis Safe signed by Sam Bankman-Fried, Robert Leshner, and others within a week — prevented indefinite repetition. The structural lesson is not that Nomi was uniquely bad; it is that a single key, even under a nominally aligned operator, is one decision away from a 50% drawdown, and refactoring under crisis takes a week of coordination that institutional capital cannot reliably underwrite in advance.

HAWK, December 2024. Haliey Welch’s token peaked at a ~$491M market cap within minutes of launch and fell 90% within three hours. Seventeen-and-a-half percent of supply was sniped by a single wallet for $993K in WSOL. A class-action followed in the Eastern District of New York naming overHere and the founders — not Welch — in December 2024. LIBRA, February 2025. Javier Milei’s endorsement produced a $4.56B peak market cap within 40 minutes and a 94% drop in 11 hours; $251M was erased across 114,000 wallets. Bubblemaps and Lookonchain traced 84% of supply to three wallets. The New York Times, as of April 7, 2026, has linked phone records showing seven calls between Milei and intermediary Mauricio Novelli on launch night, with a draft $5M payment agreement recovered from Novelli’s phone.

The academic literature is consistent with the market record. The RPHunter paper (arXiv:2506.18398) identified 4,801 rug-pull tokens among 9,528 ERC-20 contracts deployed in May 2024 — the same 50% base rate CertiK publishes. Its top-ranked backdoor pattern is “Hidden Mint/Burn, [which] allows for the creation and destruction of tokens arbitrarily. The Tokeer paper from Tsinghua’s Wingtecher group documents a specific NOVA case where an owner with onlyOwner mint minted and dumped, extracting $105,811; it classifies this as a Category-1 backdoor.

None of this implies Venice will repeat these outcomes. None of the listed precedents involved founders of Voorhees’s or Baker-Taylor’s stature. But the market prices structural risk based on what a contract can do, not what a team is likely to do, because the former is verifiable on-chain while the latter requires trust that does not scale to institutional diligence.

What Venice has said — and not said

A thorough review of Venice’s blog, its token landing pages, its Help Centre, Erik Voorhees’s X account, Teana Baker-Taylor’s public statements, and relevant podcast appearances produced zero direct statements on any of the structural questions: no plan to renounce ownership, no plan to implement a timelock, no plan to migrate mint to a multisig, no plan to encode a hard cap, no roadmap to DAO governance, no acknowledgment of centralization concerns. These absences are themselves the record.

The closest thing to an affirmative statement of the architecture is Venice’s own launch-blog bullet: “No governance.” The Block’s January 27, 2025 coverage reinforces the position: “the token comes with no governance privileges.” The Block A third-party summary on CoinStats attributes a specific Venice rationale: “governance tokens often create misaligned incentives where token holders vote for decisions that benefit them financially but harm the platform long-term.” The position is intellectually coherent. It is also a position that concentrates every tokenomic decision in a single private key.

Teana Baker-Taylor, a former VP of Policy & Regulatory Strategy EMEA at Circle, told DL News in May 2024: “We don’t have any regulatory concerns. We’re just a front end to provide access to the average person to open-source models.” This quote predates the VVV launch and does not address the token contract architecture.

One non-Venice source, pp.one, claims VVV “contracts are audited by Trust Security” with “multisig treasury.” The claim appears nowhere on Venice.ai, in Venice documentation, on BaseScan, or in Trust Security’s public portfolio. It should not be relied on.

The insider-trading event the architecture made possible

On January 28–29, 2025, within minutes of VVV’s launch on Aerodrome, two Aerodrome Finance contributors allegedly converted approximately $50,000 into $1M before any public announcement. Aerodrome’s official statement: “The timing of a small percentage of the trading activity around the $VVV launch was flagged by internal monitoring in less than 30 minutes — triggering an internal investigation. This investigation resulted in the suspension of two contributors within three hours of launch.” VVV dropped roughly 50% on the news. The event was not enabled by VVV’s contract architecture per se — it was enabled by privileged knowledge of launch timing — but it illustrates the adjacent principle: at a $400M+ valuation, any privileged information or privileged key is a first-order asset whose custody and disclosure regime matters more than the narrative around it.

The legal backdrop Voorhees has actually navigated

Erik Voorhees’s regulatory record is narrow and specific. In June 2014, the SEC settled with him personally over unregistered securities sales (SatoshiDICE and FeedZeBirds shares) for a $50,000 fine plus $15,843.98 disgorgement; he neither admitted nor denied findings. In March 2024, the SEC entered an administrative order against ShapeShift AG (Exchange Act Release 34–99676) with a $275,000 civil penalty for acting as an unregistered dealer between July 2017 and February 2020. Voorhees was not personally named in the 2024 order. Commissioners Peirce and Uyeda dissented, calling the action part of a “serial drama of the Commission’s poorly conceived crypto policy.” Voorhees’s response was not a substantive rebuttal but a tweet quoting Federalist Paper №51. In 2021, Voorhees led ShapeShift’s corporate dissolution and FOX token airdrop — 340M FOX to 1.1M addresses — under his declared principle: “Open, immutable finance requires open, immutable organizations.” That principle is the one a reader should hold alongside the Venice contract when evaluating direction of travel.

Applied to VVV, the four-prong Howey analysis is uncomfortable even under the friendlier post-April 2025 Paul Atkins SEC. The common-enterprise prong resolves cleanly: all VVV holders’ outcomes depend on Venice.ai’s revenue, its buy-and-burn execution, its emission decisions, and its GPU operation. The expectation-of-profits prong is directly reinforced by Venice’s own marketing copy, which uses the exact phrase “deflationary capital asset” on the VVV landing page — the precise framing the SEC cited against LBRY Inc. in the New Hampshire summary-judgment order. The efforts-of-others prong is harder to rebut when contract ownership is literal: owner = msg.sender, with every emission, every burn, every DIEM parameter decided unilaterally by Venice management. Airdrop recipients likely fall outside the investment-contract framework under the March 17, 2026 SEC/CFTC joint interpretive release (Release Nos. 33-11412 / 34-105020), which treats consideration-free airdrops as failing the "investment of money" prong. Secondary-market buyers on Coinbase and Aerodrome enjoy no such carve-out. The CLARITY Act, which would move "digital commodity" jurisdiction from the SEC to the CFTC, had roughly 50/50 passage odds as of April 21, 2026 (per Galaxy Research tracking via CoinDesk) and ties classification to a decentralization test that VVV's present architecture would not plausibly satisfy.

A risk matrix for the author’s readers

What changes the picture, and what doesn't

A credible, timely, targeted set of remediation steps would close most of the structural gap without requiring Venice to abandon its "no governance" thesis. The minimum-viable reforms, ordered by implementation difficulty: (1) transfer ownership to a Gnosis Safe with a 3-of-5 or 4-of-7 threshold and publish the signers, which takes one transaction; (2) wrap the mint function in a timelock with a minimum 48-hour delay, which is standard OpenZeppelin tooling and was the architecture Compound shipped in 2020; (3) encode an annual emission cap and a lifetime supply cap in a successor contract and migrate, which is the path Render took in 2023 when it moved from Ethereum to Solana and was rewarded with a +3.8% market reaction on top of a 10x rally. Worldcoin's 2024 extension of its team unlock from three to five years produced a +27% single-day move — tangible evidence that markets reward supply-side tightening.

None of these steps requires Venice to become a DAO. None of them contradicts Voorhees's stated philosophical preference against governance-token misalignment. All of them would move VVV out of the risk category of one it currently occupies.

Conclusion

The strongest version of the Venice bull case is that Erik Voorhees is among the most ideologically committed decentralization advocates in crypto, that Teana Baker-Taylor's regulatory background suggests disciplined operational judgment, and that Venice.ai has built a genuinely useful privacy-preserving AI product with real users and measurable revenue feeding a real buy-and-burn. All of that is true. None of it is encoded in the contract. The contract encodes one line — require(msg.sender == owner, "UNAUTHORIZED") — standing between an active, uncapped, unaudited mint function and 113 million tokens of supply. That line resolves to one address. That address has no Safe label. That architecture is the most permissive among every major decentralized-AI peer this report examined.

The historical record on centralized-mint tokens is not that they all fail; it is that when they are priced on their contract architecture, rather than on their narrative, they reprice sharply and suddenly. Bittensor lost 15% in two hours on nothing more than a governance allegation; SushiSwap lost 50% in a day; HAWK and LIBRA lost 90%+ in hours. Institutional due-diligence teams, faced with a $700M FDV token whose BaseScan page reads "No Contract Security Audit Submitted" and whose risk aggregator reads "Max Supply ∞," do not arrive at a buy recommendation — they arrive at a deferral. The structural thesis is not that Venice will fail. It is that Venice's current contract makes it impossible for a serious capital allocator to conclude it will not, and that the spread between the FDV and that conclusion is the price of a single private key's continued good behavior. That is a price, not an accusation, and it is the price this report suggests is not yet in the tape.