Venice AI Crosses $1B Threshold on a Privacy-First AI Bet

Venice AI closed a $65 million Series A at a $1 billion valuation on Wednesday, making the privacy-first platform founded by crypto entrepreneur Erik Voorhees a unicorn. Dragonfly led the round, with Coinbase Ventures, F-Prime, North Island Ventures, Morgan Creek and other crypto-native investors participating. It is Venice’s first outside capital since launching in 2024.

The wager is that users will pay a premium for AI conversations that never reach a permanent server-side log. Voorhees framed the raise in constitutional terms on X the same morning. In the four weeks before the announcement, Anthropic was forced offline on a U.S. export order, OpenAI was sued for allegedly piping ChatGPT queries to Meta and Google, and a Canadian regulator found that the same company had broken privacy law while training ChatGPT. None of those events targeted Venice. Each one handed the privacy thesis a fresh data point, and the privacy-first AI category moved from niche pitch-deck slide to priced unicorn round.

The Round That Priced the Wager

Voorhees announced the round Wednesday morning. He said the company holds more than 30 million of the roughly 80 million VVV tokens in circulation and has not sold any to date, opting instead to dilute equity at the company level. Series A investors received 8.98% of Venice and a vesting grant of 1.5 million VVV, plus an eight-year option to buy another 5 million VVV. Exercising that option in full would require an additional $66.5 million in payments to Venice, bringing the total potential proceeds to $131.5 million.

This capital will be used to uphold the First and Fourth Amendments to the Constitution as they relate to mankind’s interaction with AI.

Erik Voorhees, founder and CEO of Venice AI, in a post on X on July 1, 2026.

Dragonfly led, with Coinbase Ventures, North Island Ventures, Archetype, Morgan Creek Digital, Liquid 2 Ventures and Seattle-based Founders’ Co-op joining, per The Defiant and GeekWire. Coinbase CEO Brian Armstrong, who joined the cap table, called Venice “important infrastructure for freedom” in a post on X. The deal closed two years after Voorhees launched Venice in mid-2024 with Seattle-based president and CTO Jesse Proudman. The two met as classmates at the University of Puget Sound in Tacoma. Proudman previously founded cloud-computing company Blue Box, which IBM acquired in 2015, and crypto trading startup Strix Leviathan, acquired by Parataxis in early 2025. He started moonlighting on Venice in 2024 before leaving to join full time.

For Voorhees, the unicorn moment closes a sequence he framed as the reverse of the typical crypto project’s sequencing. Most token launches raise outside capital before any token trades; Venice let VVV trade openly for about 18 months before bringing in equity investors. The company also reports it became profitable in the first quarter of 2026 and sits, by Voorhees’ description, at the largest intersection of AI and cryptoeconomics by revenue. VVV’s reaction was brisk across sources: Cointelegraph reported a 6% move on Wednesday, The Defiant tracked an 11.6% rise over 24 hours to $13.80, and Crypto Briefing logged a roughly 20% intraday spike. The Defiant also reported that the warrant component, exercisable over eight years, would add just under 6,000 VVV to daily float if fully exercised, about 0.2% of current daily trading volume.

Inside the Proxy

Venice’s privacy claim rests on what it doesn’t keep, not on what it encrypts. The platform routes every request through a proxy that strips user-identifying data before forwarding it to a third-party model provider, then streams the response back to the user’s device without persisting the conversation on Venice servers. Conversation history lives in the user’s own browser, encrypted at the device level, so the same account on a second device shows an empty chat list. For models from OpenAI, Anthropic, xAI and Google, the proxy obscures the user’s IP address, account and session data; other models on the platform offer higher levels of privacy. Venice’s own privacy architecture blog post lays out the design choice in plain terms: “The only way to achieve reasonable user privacy is to avoid collecting this information in the first place.” That sentence comes from a July 18, 2024 blog post that doubled as Venice’s launch manifesto.

The platform offers access to more than 200 AI models covering text, images, audio and video, per Crypto Briefing. It processes roughly 1.7 million API calls a day and integrates with NEAR AI for verifiable private inference, the technical mechanism that lets Venice make a credible claim about what it doesn’t do with user data. About 8% of transactions on the platform run through crypto payments; the rest go through Stripe and similar card processors. The architecture underwrites both the privacy marketing and the VVV token economy: users stake VVV to mint DIEM credits, which they redeem for compute.

$70M ARR and a Token That Actually Burns

Venice says it has reached 3.5 million users as of the funding announcement. The company turned profitable in the first quarter of 2026 and now runs an annualized revenue rate above $70 million. That puts the valuation at roughly 14 times ARR. About 45 employees support that revenue, up from roughly 15 a year ago, with six based in Seattle. Most of the team works remote, and Venice does not currently operate an office.

The numbers put Venice inside the same consumer conversation as the apps it positions against. Proudman told GeekWire the company wants Venice “thought of in the consumer landscape on the same terms as a ChatGPT or an Anthropic,” with users opening their phones and seeing Venice’s app alongside the mainstream assistants. Reaching 3.5 million users from a privacy-first position is the result the round’s lead investors priced.

The revenue mix shows how the VVV token has been woven into the business. Users stake VVV to generate DIEM credits that can be redeemed for inference on the platform, the staking mechanics The Defiant described as treating AI compute like bandwidth. About 8% of transactions are paid in crypto; the bulk still flows through Stripe for credit card customers. News.Bitcoin.com reported Venice has burned roughly 42% of VVV’s circulating supply since the token launched, a structural deflationary mechanism that runs in parallel with the equity round.

VVV trades openly on exchanges; inside the platform it is staked by users to mint DIEM credits for inference. The Defiant reported that the warrant component of the Series A, exercisable over eight years, would add just under 6,000 VVV to daily float if fully exercised, about 0.2% of current daily trading volume. Tokens granted or optioned to investors are locked for a year and unlock linearly over three additional years, meaning the supply release would not reach the market until roughly two years after Venice’s token launch. About 8% of Venice’s transactions are paid in crypto; the rest go through Stripe, a figure News.Bitcoin.com said suggests most of the platform’s recent growth has come from users outside the crypto-native audience. News.Bitcoin.com reported Voorhees said the structure aligns incentives across the company, investors and the token-holding community without flooding the market with fresh supply. Voorhees framed the sequencing as the reverse of how most crypto projects go public: most tokens raise outside capital before they trade; Venice let VVV trade openly for 18 months before bringing in equity.

Venice AI by the numbers:

• Annualized revenue: above $70 million
• Active users: 3.5 million (as of funding announcement)
• Employees: ~45
• Daily API calls: ~1.7 million
• Crypto payment share: ~8% of transactions

Why Privacy Stopped Being Niche This Quarter

The round landed in a month that gave the privacy pitch more ammunition than any campaign could buy. Three separate events in the four weeks before the announcement hit the major AI platforms where their data practices are most exposed. None of them involved Venice, and each one made the privacy thesis easier to sell.

On June 12, the U.S. Department of Commerce imposed export controls on Anthropic’s most advanced models, forcing the company to cut foreign access to Claude Fable 5 and Mythos 5 for an 18-day stretch that ended June 30, according to the federal order and its June 30 lifting. In May, a proposed class action filed in the U.S. District Court for the Southern District of California accused OpenAI of embedding Meta Pixel and Google Analytics on ChatGPT.com and transmitting user queries to both companies, as detailed in the tracking-pixel allegations. Earlier in May, Canada’s privacy watchdogs ruled that OpenAI had broken Canadian privacy law while training ChatGPT, the first formal Canadian finding of its kind against a generative AI company. A parallel suit against Perplexity, filed in San Francisco federal court on April 1, made the same allegation against a different chatbot vendor. Each event hit a different mainstream AI vendor in a different jurisdiction; Venice was not named in any of them.

• June 12, 2026: U.S. Commerce imposed export controls on Anthropic’s Claude Fable 5 and Claude Mythos 5
• May 2026: Class action filed in California accusing OpenAI of sharing ChatGPT queries with Meta and Google
• May 2026: Canadian privacy watchdogs ruled OpenAI broke the law training ChatGPT
• April 1, 2026: Class action filed in San Francisco making similar allegations against Perplexity

Control over intelligence is the defining fight of the coming decade. Whoever owns the AI delivery stack owns a direct window into your interior life. They log all your chats, train on them, and will hand them over when asked.

Haseeb Qureshi, managing partner at Dragonfly, in a statement on July 1, 2026.

Qureshi made the macro case in a statement the same morning Venice announced the round. The framing maps directly onto the events of the prior month: a regulator forcing a top model offline because of cyber risk, a lawsuit alleging that prompts reached ad-tech vendors, a Canadian ruling that OpenAI had trained on data without consent. Voorhees, asked on X what the capital would fund, answered in constitutional language. The two statements positioned the round as a bet on who holds the delivery stack when AI moves from optional helper to default interface.

What Privacy-First Still Has to Defend

The privacy-first pitch comes with a misuse problem Venice has to manage in public. GeekWire flagged that stripping content filters from commercial models makes Venice more useful for some users but also reopens the questions that lead mainstream services to build guardrails in the first place. Proudman told the outlet Venice includes some safeguards to prevent abuse and illegal activity, while casting the surveillance of users’ thoughts as the larger risk.

Proudman told GeekWire that user data on mainstream chatbots is only as safe as the company holding it: a breach, a disgruntled employee, a government subpoena or a change in government policy could expose it all. Venice’s answer is to keep its prompts away from any central store, and to put the GPU hardware behind the inference inside its own data centers. Crypto payments account for only about 8% of transactions on the platform, and the 3.5 million users Venice reports are not gated behind holding the VVV token. The category now has a priced round, a defended architecture and a regulatory environment that now includes federal export controls on frontier AI models. Voorhees and Qureshi framed the round as a wager on who controls the AI delivery stack over the next decade.

Frequently Asked Questions

What does Venice AI actually do differently from ChatGPT?

Venice routes every prompt through a proxy that strips user-identifying data before forwarding it to a third-party model and stores no conversation history on its own servers. Chat history lives in the user’s local browser, encrypted at the device level, so the same account on a second device shows an empty chat list. For models from OpenAI, Anthropic, xAI and Google, the proxy obscures the user’s IP address, account and session data; other models on Venice offer higher levels of privacy.

Who led Venice AI’s $65 million Series A?

Crypto-focused venture firm Dragonfly led the round, with Coinbase Ventures, North Island Ventures, Archetype, Morgan Creek Digital, Liquid 2 Ventures and Seattle-based Founders’ Co-op participating. Coinbase CEO Brian Armstrong called Venice “important infrastructure for freedom” in a post on X the morning of the announcement. It is Venice’s first outside equity raise since launching in 2024.

What will Venice do with the $65 million?

Voorhees said the capital will fund owned data center infrastructure and GPU purchases, the first time Venice will own its compute, ending its dependence on rented capacity. The remainder is earmarked for customer growth, new markets, hiring and acquisitions Voorhees described as “additive businesses.” Series A investors received 8.98% of the company plus a vesting grant of 1.5 million VVV and an eight-year warrant for an additional 5 million VVV; exercising the full warrant would bring total proceeds to $131.5 million.

What is the VVV token used for inside Venice?

VVV is Venice’s staking and access token. Users stake VVV to mint DIEM credits that can be redeemed for inference on the platform, and News.Bitcoin.com reported the company has burned roughly 42% of the circulating supply. About 8% of Venice’s transactions are paid in crypto; the rest flow through Stripe. Venice holds more than 30 million of the roughly 80 million VVV in circulation and has not sold any tokens to date.

How does Venice’s $1B round relate to the OpenAI lawsuits and the U.S. export controls on Anthropic?

The Series A announcement came weeks after the U.S. Commerce Department imposed and then lifted export controls on Anthropic’s Claude Fable 5 and Claude Mythos 5, and after separate class actions in California accused OpenAI and Perplexity of sharing ChatGPT queries with ad-tech vendors. Canada also ruled in May that OpenAI had broken Canadian privacy law while training ChatGPT. Dragonfly’s Haseeb Qureshi tied those events directly to Venice’s thesis in his statement on the round, arguing that whoever owns the AI delivery stack “owns a direct window into your interior life.” Venice’s pitch is that users will pay to keep that window closed.