When Erik Voorhees set out to build Venice AI in 2024, the pitch was straightforward, if bold: artificial intelligence without the surveillance. No prompt logging. No training on user data. Just unfettered access to AI models with privacy baked in from the start.
On July 1, investors decided that vision was worth a billion dollars.
Venice closed a $65 million Series A led by Dragonfly, joining a roster that includes Coinbase Ventures, North Island Ventures, F-Prime, Archetype, Liquid2 Ventures, Morgan Creek, and Founders' Co-op. The round marks the company's first outside capital and an unusually swift climb to unicorn status—24 months from founding to ten-figure valuation. Even in an AI market accustomed to breakneck growth, that trajectory stands out.
What's perhaps more unusual is how Venice structured the deal. Rather than a conventional equity round, investors received 8.98% of the company alongside 1.5 million VVV tokens and warrants for another 5 million tokens exercisable over eight years. Fully exercised, those warrants could add roughly $66.5 million to the kitty, pushing total proceeds north of $131 million. The tokens and warrants come with a one-year lockup, followed by a three-year linear vest.
Voorhees, who disclosed the structure publicly shortly after the closing, noted that daily unlocks would represent about 0.2% of current trading volume—a detail presumably meant to reassure markets that dilution won't flood secondary trading. VVV, an ERC-20 token running on Base, launched in January 2025 and trades on Coinbase.
The Numbers Behind the Bet
Venice's investors are betting on more than just philosophical alignment around privacy. The company says it logged 3.5 million registered users and now processes 1.3 trillion tokens each month. Daily API calls hover around 2 million, occasionally spiking past 2.1 million, while monthly site traffic ranges between 1.3 million and 1.6 million visits. At peak load, the platform handles 300,000 inference requests per hour.
More to the point for venture backers: Venice claims it turned profitable in the first quarter of 2026 and has surpassed $70 million in annual recurring revenue. Those figures, shared in press interviews the day of the funding announcement, suggest the business model is working—at least for now.
The platform offers access to over 250 models spanning text, image, video, and audio through OpenAI-compatible endpoints. Privacy settings range from anonymous queries to fully end-to-end encrypted inference running inside trusted execution environments, courtesy of infrastructure partnerships with Phala Network and NEAR AI Cloud. Conversations stay local on user devices; Venice insists it never logs prompts.
Crypto Roots, AI Ambitions
Voorhees brings a cryptocurrency pedigree to the AI table. Before Venice, he founded ShapeShift, a crypto exchange, and Satoshi Dice, an early Bitcoin betting service. His co-founder, Jesse Proudman—who serves as president and CTO—is a Seattle tech veteran whose past ventures include Blue Box, acquired by IBM in 2015, and Strix Leviathan, a quantitative crypto trading firm.
That crypto background shows up in Venice's business model. Users can stake 100 VVV tokens to unlock Venice Pro access. Staked tokens generate DIEM, a non-expiring credit token that yields $1 in platform credits daily—a mechanism that ties usage to token economics in a way traditional SaaS platforms typically don't. Since late 2025, Venice has run monthly buy-and-burn programs; as of July 1, it had burned 33.7 million VVV, or roughly 42% of total supply. The funding announcement came alongside a reduction in token emissions, from 4 million to 3 million VVV annually.
What the Money's For
Venice's immediate priority is owning more of its compute infrastructure. The company plans to build its first proprietary data center, a move designed to reduce dependence on leased GPUs and improve gross margins—a familiar playbook for AI platforms that reach a certain scale. Capital will also fund international expansion, customer acquisition, potential acquisitions (described as "additive"), and hiring.
Consumer pricing tiers range from a free plan with limited daily inference to a $200-per-month Max tier offering 22,500 monthly credits with three-month rollover. The developer API business runs parallel to the consumer app, though the company hasn't broken out revenue contribution from each.
The Open Question
Venice's early traction suggests a market for privacy-centric AI exists. Whether that market can support a billion-dollar valuation at scale is the bet Dragonfly and its co-investors are now underwriting. The privacy angle differentiates Venice from incumbents like OpenAI and Anthropic, but it also constrains certain monetization levers—data used to improve models, for instance, or advertising-supported free tiers.
Then there's the token structure itself, which adds a layer of complexity most AI companies avoid. Voorhees has argued it aligns incentives and gives users ownership in the platform. Skeptics might counter it introduces volatility and regulatory uncertainty.
For now, Venice is moving fast—profitably, it claims—while building out infrastructure and expanding its user base. The next chapter will reveal whether "private, unrestricted" AI can sustain momentum beyond the early adopter crowd that got it to unicorn status in two years.
