On Monday, August 17, 2026, Erik Voorhees said that Venice had crossed a $100 million annualized revenue run rate, roughly six weeks after the company raised a $65 million Series A at a $1 billion valuation on July 1. VVV, the token attached to the platform, was trading at $13.83 with a $654.57 million market capitalization when we pulled CoinGeckoat 07:29 UTC on Wednesday, August 19, up 15.87% over seven days and 21.22% over thirty.
An AI token with a disclosed revenue figure is a rare object, because most of the sector, from infrastructure plays down to AI agent tokens, sells a thesis rather than a P&L. But a run rate is not revenue, the token's supply expands every single block, and the emissions schedule meant to fix that steps down twice between the announcement and November.
What Voorhees Actually Announced
Venice is a privacy-oriented generative AI platform on Base that routes users to open-source models without retaining prompts on company servers, and VVV is the Base-native token gating access to its API tier. We covered the platform in "What Is Venice Token (VVV) and Why the Decentralized AI Token Is Trending," and Voorhees, who founded ShapeShift, got his own profile from us earlier in 2026. This piece assumes both and goes at the number instead.
The figure was posted to X and picked up the same day by Crypto Briefing, which put the prior mark at roughly $70 million around the time of the Series A. Venice's own Series A post from July 1, 2026 states no revenue figure at all, so the $70 million baseline comes from coverage of the round rather than from the company, and the growth rate everyone is quoting rests on that softer number.
What Venice did publish is usage, and that is the firmer anchor. The Dragonfly-led round came with 3.5 million registered users, 1.3 trillion tokens processed per month, and 2 million API calls per day from developers, peaking above 2.1 million.
A Run Rate Is Not Revenue and the Gap Matters
A run rate takes a recent, short window of income and multiplies it out to twelve months, so a $100 million annualized run rate means the most recent period was earning at a pace of about $8.3 million a month. It does not mean Venice has collected $100 million, and it carries no information about churn, refunds, credit expiry, or how much of that spend came from a handful of large API customers.
This is where most of the later disappointment comes from. A subscription business that grows from $70 million to $100 million on the run-rate line has genuinely grown, while a business that booked one large enterprise contract in a strong month prints the identical figure and is not the same company at all.
Three things would make the number durable. A second consecutive quarter at or above the same pace, a split between subscription revenue and API credit purchases, and some indication of net revenue retention. Venice publishes none of those, so the $100 million remains a founder's statement rather than an audited disclosure, which is less a knock on the founder than a description of the category.
The Emissions Step Down Twice Before November
Venice published the schedule itself in a tokenomics update dated July 17, 2026 and amended August 5. Annual VVV issuance falls from 3 million to 2.5 million on September 1, 2026, then to 2 million on October 1, 2026.
|
Date
|
Annual VVV issuance
|
Value at $13.83
|
|
Before September 1, 2026
|
3,000,000
|
$41.49 million
|
|
From September 1, 2026
|
2,500,000
|
$34.58 million
|
|
From October 1, 2026
|
2,000,000
|
$27.66 million
|
Both cuts are still ahead, which makes them the cleanest scheduled catalyst this token has. Against the 114.65 million tokens that exist on-chain, 2 million a year is dilution of 1.74%, while against CoinGecko's stated circulating supply of 47.35 million the same 2 million is 4.22%. The real rate depends on which supply figure you believe, and that turns out to be a live question. Our guide to token inflation covers why the denominator decides the answer.
What the Buyback Actually Offsets
Venice routes $5 of every $100 of credits purchased on the platform into buying and burning VVV, a mechanism live since mid-July 2026. Take the announced run rate at face value and assume, generously, that every dollar of it is credit spend. That caps the buyback at $5 million a year.
Emissions at the October 1 rate are worth $27.66 million a year at our pull price. So even after both cuts, new issuance runs at roughly 5.5 times the largest buyback the stated revenue could possibly fund. Before the cuts, at the 3 million rate, the multiple is about 8.3 times.
Nobody has to be lying about the business for that to hold. Revenue can grow fast while the token stays structurally net inflationary, because the burn is a percentage of a revenue line while emissions are a fixed token quantity revalued at market price. The burn catches up only if revenue multiplies several times over or the price falls far enough to make emissions cheap in dollar terms, and holders do not want the second one.
DIEM, Venice's compute bond, puts a second claim on the same revenue. It is minted by locking staked VVV, each staked DIEM entitles the holder to $1 per day of API credits, and the supply target rises in four steps from 38,000 to 40,000, reaching 39,500 on August 31, 2026 and 40,000 on September 14, 2026. At a full 40,000 staked that is $14.6 million a year in inference credits issued in kind. Hyperliquid crossed $1 billion in revenue before its own buyback started to visibly bite, which is the scale this kind of mechanism needs.
The Supply Number Depends on Where You Look
We called
totalSupply() on the VVV contract at 0xacfe6019ed1a7dc6f7b508c02d1b04ec88cc21bf across two independent Base RPC endpoints, and Basescan confirms it as a standard ERC-20 with 18 decimals, name "Venice Token" and symbol "VVV". Base is an Ethereum Layer 2, so every one of these reads is repeatable by anyone with a browser.At block 50,167,780, timestamped 07:35:07 UTC on Wednesday, August 19, 2026, total supply read 114,652,250.49 VVV. CoinGecko's total supply field for the same asset reads 80,814,881, and its own page settles the disagreement against itself. Its fully diluted valuation of $1.585 billion divided by its own price implies a supply near 114.65 million, matching the contract, while its stated 80.8 million multiplied by the same price gives $1.12 billion, nowhere near the FDV it publishes. The fields contradict each other, so we used the contract.
The reads also show something a static number hides. Between block 50,167,618 at 07:29:43 UTC and block 50,167,780 at 07:35:07 UTC, total supply rose by 29.87 VVV, an annualized pace of roughly 2.9 million tokens and a close match for the stated rate. VVV is not handed out in scheduled tranches. It mints continuously, block by block, and it was minting while this article was being written.
The Control Powers That Are Still on the Contract
Mint authority has not been renounced. That is not a hidden defect, since continuous emissions require exactly that power, but it does mean the schedule above is a policy rather than a hard cap enforced by immutable code. The contract carries no maximum supply, and the 2 million a year rate holds because Venice says it holds.
The
owner() address is 0x321b7ff75154472b18edb199033ff4d116f340ff, itself a contract rather than a person's wallet, and it held 35,085,987.76 VVV when we read it at 07:35 UTC on Wednesday, August 19. That is 30.60% of on-chain total supply sitting at one address, on a token whose mint rights that same address controls. Whatever governance sits behind the proxy, the concentration is real and it is the number worth watching.Nothing worse than that showed up. Automated contract screening on Base returned zero buy tax, zero sell tax, no modifiable slippage, no blacklist, no transfer pause and no honeypot flag, and a direct
paused() call reverts because the function does not exist. Claims that the deployer retains fee-setting or sell-disabling powers did not survive contact with the contract, and the deployer address holds no VVV at all. The mint power and the 30.6% concentration are the material risks, and the rest of the scare list is not there.Frequently Asked Questions
Does Venice AI revenue go to VVV holders?
Not directly, because holders receive no dividend and no revenue share. The link runs through the 5% of credit purchases used to buy and burn VVV on the open market, and through DIEM, which converts locked VVV into a daily API credit entitlement.
What happens to VVV on September 1 and October 1, 2026?
Annual issuance drops to 2.5 million tokens on September 1 and to 2 million on October 1, per Venice's published tokenomics update. Nothing is distributed or released on those dates, so anyone expecting a supply event will be looking for something that does not happen. The change is to the rate of continuous minting, and it shows up gradually rather than in a single candle.
Why do CoinGecko and the blockchain disagree about VVV supply?
Data aggregators maintain supply fields manually or through periodic feeds, and those go stale when a token mints continuously. The contract is the source of truth, and in this case CoinGecko's own fully diluted valuation is calculated off the on-chain figure while its total supply field is not.
Is a $100 million run rate large for a crypto AI project?
For this sector it is unusually large, because most crypto AI tokens report no revenue at all and are valued on projected demand, which is the same foundation under much of the broader AI investment case. The caveat is that a run rate is an annualized snapshot of a short period, so the comparison only means something once it repeats.
Bottom Line
The revenue claim is the strongest thing this token has, and it stays a founder statement until a second period confirms it. The two dates that change the arithmetic are September 1 and October 1, 2026, when annual issuance falls to 2.5 million and then 2 million VVV, and even at the lower rate new supply is worth roughly five times the largest buyback a $100 million revenue line could fund. Three things are worth watching over the following two quarters. A second disclosure at or above the same pace, a burn figure published by the company rather than inferred by outsiders, and any movement out of the 30.6% owner balance. Get the repeat number with the owner address sitting still and the emissions cuts start to matter, but if the $100 million turns out to be a one-off, the schedule is a slower version of the same dilution.
This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency trading involves substantial risk. Always conduct your own research before making trading decisions.
