Since the start of the year VVV, Venice's token, has risen more than sixteenfold. Bitcoin is slightly down over the same period. Venice belongs to Erik Voorhees, the founder of ShapeShift, and sells AI without data collection. In July it raised 65 million dollars in a round led by Dragonfly.
The buying case for the token goes like this. If you use Venice's AI through its interface, you can pay with DIEM. And DIEM is made from locked VVV. More usage, more locked VVV.
If that is why you buy VVV, you should know one number: what does it cost today to make a new DIEM, and what does one cost on the market?
What Venice does: AI without storage, as a subscription and as an API
Venice is an AI service like ChatGPT, online since May 2024. You chat in the browser or call the models through an interface (API) from your own programs. It mainly offers open models such as DeepSeek, Qwen or Llama, plus pass-through access to GPT, Claude and Gemini.
The promise is privacy. According to Venice, your chat history stays on your device and prompts are not stored. There is a free tier with 10 text prompts a day and subscriptions from 18 to 200 dollars a month. The API is billed per million tokens.
The product itself has little to do with crypto. Venice runs no blockchain of its own and no decentralized compute network. Crypto only sits in payments and in the token: VVV and DIEM live on Base, Coinbase's blockchain.
Venice is an AI company with a token, not a crypto network with an AI.
What is technically different, and where the limits are
Venice offers four levels of privacy:
- Anonymous: Venice forwards your request to GPT, Claude or Gemini. The provider sees the prompt, but not who you are.
- Private: open models on rented GPUs. That nothing is stored is agreed with the providers by contract.
- TEE: the computation runs in a sealed area of the GPU that even the operator cannot read. This can be proven technically (remote attestation).
- E2EE: like TEE, but already encrypted on your device.
TEE and E2EE have been available since 18 March 2026, only in the Pro subscription and only for selected models, built with Phala and NEAR AI.
The limits: the "Private" level rests on contracts with GPU providers that Venice does not name. No independent audit of the no-storage claim has been published. And according to its own privacy page, Venice processes metadata such as account ID, timestamps and IP addresses.
The privacy is partly technology and partly promise. Which part applies depends on your plan.
Who uses Venice: big numbers that contradict each other
On-chain you can only measure who holds and stakes VVV and DIEM. User numbers and revenue exist only as claims:
| Claim | Who, when | Assessment |
|---|---|---|
| 3.5m registered users, 850,000 to 1m visitors a month | Venice blog, Voorhees, 1 Jul 2026 | reported |
| "more than 3 million monthly active users" | Crypto Briefing, 17 Aug 2026 | contradicts Venice's own visitor numbers |
| 2m API calls a day | Venice blog, 1 Jul 2026 | reported |
| "$70+ million" annual revenue, profitable since Q1 2026 | The Block, 1 Jul 2026 | reported, no financial statements |
| $100m annual revenue | Crypto Briefing, 17 Aug 2026, attributed to Voorhees | original post not found |
As of 7 October 2026.
So Venice is a real product with many users. How many of them pay, and how much runs through the API, is not published.
What Venice earns, Venice says. What the token gets from it, the chain shows.
The core: DIEM, an annuity paid in compute
This is how DIEM works:
- You stake VVV and receive sVVV.
- You lock sVVV and mint DIEM with it.
- You stake the DIEM. In return you get 1 dollar of Venice API credit every day, indefinitely. At midnight UTC whatever you did not use expires.
- To get out, you burn the DIEM and get your VVV back.
So a DIEM is an annuity paid out in compute. You can also simply buy one on the market.
How much VVV you have to lock for a new DIEM is set by a price curve: the more DIEM exist, the more expensive the next one. The first ones cost 90 sVVV. And today:
| 7 October 2026 | |
|---|---|
| DIEM on the market | about $1,870 |
| minting a new DIEM | 438 VVV, about $11,500 |
| ratio | minting costs about six times as much |
| DIEM in total | 36,667 (end of April: 38,774) |
| credit from all staked DIEM | $27,857 a day |
Base on-chain and CoinGecko, 7 October 2026. VVV at $26.35.
Buying VVV to mint one ties up 11,500 dollars for something the market sells for 1,870. That does not pay. A little minting still happens, though, by people whose VVV is staked anyway. Locked sVVV keeps 80 percent of the staking yield, so minting costs them only a fifth of their yield. From 30 September to 7 October, that added a net 82 DIEM.
Over half a year the supply is still shrinking, by just over 2,100 DIEM since the end of April. People who minted cheaply earlier buy DIEM back on the market, burn it and free up their VVV. On 17 July Venice set a target of 40,000 DIEM by 14 September. It was not reached.
For the buying case this means: anyone who uses Venice more today and wants DIEM for it buys an existing one. No new VVV get locked.
Whoever mints a DIEM today locks VVV they already had. No new VVV gets bought in the process.
Is DIEM at least worth it for you as a user?
Is DIEM worth it as a user?
A DIEM gives you 365 dollars of credit a year if you use the full dollar every day. At a price of 1,870 dollars, it pays for itself after a little over five years, not counting moves in the DIEM price itself.
Three catches: the credit expires every day. It applies at the prices Venice sets. And DIEM itself swings hard: in November 2025 it cost 81 dollars, in September 2026 up to 2,566 dollars.
The sums change if your VVV is already staked. Then you mint yourself and give up a fifth of the yield on about 11,500 dollars, roughly 140 dollars a year. In return you get up to 365 dollars of credit. Your VVV stays locked until you hand the DIEM back.
For comparison, the Pro subscription costs 18 dollars a month. If you only chat, you don't need a DIEM. It is interesting for someone who uses the API evenly, every day, for years.
A DIEM is five years of compute paid in advance, not free access.
The objection: "Then Venice will just lower the curve"
Venice can do that. And that is exactly the point.
The DIEM price curve is set in the staking contract, and that contract is controlled by a Safe multisig that needs 4 of 6 keys to sign. The same Safe is Venice's treasury. It can swap out the contract, change the curve and set how many new VVV are issued. The VVV contract itself has no cap on new tokens.
So far Venice has used this power to slow supply. Issuance fell from 14 million to 2 million VVV a year in 14 months:
| Period | New VVV per year |
|---|---|
| until July 2025 | 14m |
| September and October 2025 | 10m |
| November 2025 to January 2026 | 8m |
| February to April 2026 | 6m |
| May, June 2026 | 5m, 4m |
| July, August 2026 | 3m |
| September 2026 | 2.5m |
| since October 2026 | 2m |
Monthly growth of total supply, Base on-chain.
For bitcoin or Bittensor, the issuance schedule is in the code. For VVV it is a series of decisions. So far they have all gone in one direction, but they can go the other way too.
VVV's supply is a promise by Venice, not a protocol.
What the token gets from revenue
The company Venice earns from the product, and the company has shareholders, since July including Dragonfly with almost 9 percent. The token gets a share that Venice decides: according to the Venice blog of 17 July, 5 dollars of every 100 dollars of credits purchased go to buying back and burning VVV.
On-chain it looks like this:
| VVV | in dollars | |
|---|---|---|
| burned in September 2026 | 36,418 | about $0.89m |
| newly issued in September 2026 | 206,430 | about $5.1m |
| burned since December 2025 in total | 348,304 | about $3.3m |
Monthly average prices. Base on-chain.
Ongoing burns only started in December 2025. The big item of 33.5 million burned VVV dates from March 2025: unclaimed airdrop tokens and a team buyback. In September, about six times as many VVV were issued as burned. Net supply grows by about 1.9 percent a year.
On top come options: according to The Block, Dragonfly holds warrants on 5 million VVV with an eight-year term, about 6 percent of today's supply. The price is not published.
Buying VVV does not buy you a share of Venice. It buys a burn rate that Venice sets itself.
Tokenomics: who holds VVV
| Item | VVV | Share |
|---|---|---|
| originally minted | 100m | |
| issued since | about 15m | |
| burned | 33.9m | |
| supply today (net) | 81.1m | 100 % |
| staked | 34.2m | 42 % |
| of which locked for DIEM | 8.3m | 10 % |
| Venice treasury | 20.7m | 26 % |
| rest: exchanges, pools, wallets | about 26m | 32 % |
Base on-chain, 7 October 2026.
After the cuts, the staking yield is about 6 percent a year; at the start of 2025 it was 70 to 100 percent on paper. It comes entirely from new VVV. Unstaking takes seven days.
Ownership is concentrated: according to GeckoTerminal, the ten largest addresses hold 88 percent of all VVV, including the burn address, the staking contract and the treasury.
A quarter of VVV sits with Venice itself, and the same keys decide how much more gets added.
The price against bitcoin
| VVV | BTC | |
|---|---|---|
| 2025 (from 28 January) | −82.3 % | −13.5 % |
| 2026 to 6 October | +1,554 % | −2.4 % |
| since trading start (28 Jan 2025) | +192.8 % | −15.6 % |
| up / down months (Feb 2025 to Sep 2026) | 11 / 9 | |
| months better than BTC | 11 of 20 | |
| lowest close | $0.95 (1 Dec 2025) | |
| highest close | $31.96 (21 Sep 2026) | |
| worst drawdown | −89.9 % (January to December 2025) |
KuCoin VVV-USDT and Binance BTCUSDT, daily closes through 6 Oct 2026. The first trading day, 27 Jan 2025, ran only on Coinbase, up to $23.51.
VVV has had two completely different years. In 2025 the token fell about 90 percent. In 2026 it rose 34-fold from the low. That period includes the issuance cuts, the encrypted modes, the Upbit listing on 12 May (May +116 percent), the funding round and the new burn rules. Which of these carried the price cannot be measured.
The platform has had VVV in a parabolic phase since 8 September. It began when price reached twice its 200-day average after rising more than 50 percent in 30 days; today it sits 79 percent above that average. Of 517 such phases on Binance pairs, 67 percent were not in profit at the first exit. The cases cluster in 2021 and 2024, though, with only 74 independent time windows. It is a state, not a sell signal.
VVV does not move with bitcoin. It moves around its own decisions.
What was announced and what can be checked
| Claim | Who, when | What the data says | Verdict |
|---|---|---|---|
| "Venice does not store prompts" | Venice, privacy page | no audit; "Private" rests on contracts; IP addresses are processed | not verifiable |
| Encrypted modes with technical proof | Venice, Phala, NEAR AI, 18/19 Mar 2026 | partners confirm; Pro plan only | partly |
| 32.6m airdrop tokens and 1m team buyback burned | Venice, The Block, 12 Mar 2025 | 33.54m VVV at the burn address at end of March 2025 | true |
| "33.7m VVV burned, about 42 %" | Venice blog, 1 Jul 2026 | amount correct; 42 % only relative to circulating VVV, of all ever minted it is 29 % | partly |
| Issuance falls to 3, 2.5 and 2m | Venice blog, 1 Jul and 5 Aug 2026 | implemented on-chain exactly so | true |
| Issuance "started at 10 million" | Venice FAQ | it was 14m until July 2025 | false |
| 40,000 DIEM by 14 Sep 2026 | Venice blog, 17 Jul 2026 | 36,667 on 7 October | did not happen |
| "$70+ million" or "$100 million" annual revenue | The Block, 1 Jul 2026; Crypto Briefing, 17 Aug 2026 | no financial statements published | not verifiable |
| "more than 3 million monthly active users" | Crypto Briefing, 17 Aug 2026 | Venice itself cites at most 1m visitors a month | contradictory |
| Series A $65m, warrants on 5m VVV | Venice blog, The Block, 1 Jul 2026 | both sources agree; terms open | true |
What Venice promises on-chain, it delivers. What it says about users, revenue and privacy cannot be checked.
What this means for you
If you use Venice or want to hold VVV:
- Weigh DIEM against your subscription. A DIEM only pays off if you use the full API credit every day for about five years. For chatting, the subscription is enough.
- Don't buy VVV because "usage locks up tokens". At today's price new DIEM are minted only by people whose VVV is already staked. More usage does not lead to new VVV purchases. That only changes if the DIEM price rises above the minting cost or Venice lowers the curve.
- Check issuance yourself. It is in the staking contract (
emissionRatePerSecond), currently 2 million VVV a year. If the number goes up, the multisig has decided differently.
What is not claimed here: that Venice is a bad product or that the price has to fall. By its own account Venice is growing fast, and its on-chain commitments have been kept so far. The open question is whether the gap between DIEM's minting cost and market price narrows, and if so, whether through a rising DIEM or a new curve.
FAQ
What is the difference between VVV, sVVV and DIEM? VVV is the token. sVVV is staked VVV and earns the staking yield. DIEM is minted from locked sVVV and, when staked, pays 1 dollar of API credit a day.
Can I use DIEM for chatting? DIEM applies to the API. For chatting in the browser there are the subscriptions.
Why is a DIEM cheaper on the market than its VVV backing? On average each DIEM is backed by 227 locked sVVV, about 6,000 dollars. The market prices DIEM by the credit it delivers, not by the locked VVV. That is why early minters buy DIEM back and free their VVV.
Does locked sVVV still earn staking yield? Yes. 80 percent of the reward goes to you, 20 percent to the Venice treasury.
Where is VVV traded? Among others on Coinbase, Kraken, OKX, KuCoin, Bybit and Upbit, not on Binance spot.
All VVV monthly returns in US dollars
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2025 | −28.8 | −35.8 | 61.4 | −32.3 | 4.7 | −8.0 | 2.0 | −29.2 | −22.7 | −32.1 | 58.0 | |
| 2026 | 43.5 | 131.5 | 20.8 | 30.3 | 116.1 | −33.8 | −3.7 | 42.6 | 64.8 |
KuCoin VVV-USDT, monthly return close to close in percent; January 2025 only 28–31 Jan, therefore blank.
Not investment advice, not a recommendation, not a forecast — historical patterns are no promise.
Study the Past — Improve your Future 🥋