The archive · AI & Models · Product decision · 2024–2026
Venice AI's uncensored-privacy bet hits $70M ARR and a $1B Series A
Erik Voorhees' Venice AI reached 3M users and profitability before its first VC round: a $65M Series A at a $1B valuation.
Venice AI
What the business is
A privacy-first generative AI platform that hosts uncensored open-source models on its own infrastructure, routes queries to closed models anonymously, and monetizes through subscriptions, API credits and crypto tokens.
Starting capital：$65M Series A led by Dragonfly with Coinbase Ventures, North Island Ventures, F-Prime, Archetype, Liquid2 Ventures and others; its first outside capital.
How it started
Crypto founder Erik Voorhees - creator of ShapeShift and Satoshi Dice - launched Venice AI in mid-2024 on a conviction carried over from Bitcoin: a neutral protocol should work the same for everyone, so an AI service should not watch, restrict or lecture its users. He funded the company without venture capital and treated the product as a mass-market alternative to ChatGPT.
What happened
Growth was driven by closing the feature gap with ChatGPT plus a token flywheel: users buy and stake VVV tokens to mint DIEM, which generates $1 per day in AI credits, while Venice uses revenue to buy back and burn tokens. About 8% of users pay with crypto; subscriptions and API usage carry the revenue. By April 2026 the platform reported 3M+ users and roughly 1.7M daily API calls, and it became profitable in Q1 2026 with over $70M in annualized run-rate revenue - all before taking outside capital.
How it ended up
Scaling: on July 1, 2026 Venice announced its first external round, a $65M Series A at a $1B equity valuation led by Dragonfly, with plans to build its first owned data center and buy GPUs to cut leased-compute costs, raise gross margins and expand into a mass-market consumer app and agent API.
Background
Venice AI is a privacy-first generative AI platform founded in mid-2024 by crypto entrepreneur Erik Voorhees, creator of ShapeShift and Satoshi Dice. Its pitch: users should get unrestricted access to powerful AI models without a lab watching, storing or moderating their prompts. The service hosts uncensored open-source models on its own infrastructure and routes queries to closed models from OpenAI and Anthropic anonymously, encrypting requests client-side and keeping no prompts on its own systems.
The operating bet was explicitly ideological and borrowed from Bitcoin: treat the platform as a neutral tool that works the same way for everyone, and let users choose their own level of moderation. To build distribution without venture money, Venice launched a token ecosystem: users buy and stake VVV to mint DIEM, which generates $1 per day in AI credits, and the company buys back and burns tokens from revenue. Only about 8% of users pay with crypto; subscriptions and API usage generate most of the money.
By April 2026 Venice reported 3M+ active users and roughly 1.7M daily API calls, with 850,000+ unique website visitors. The company says it became profitable in Q1 2026 and reached over $70 million in annualized run-rate revenue, helped by features converging on ChatGPT so that privacy became a tiebreaker rather than the only reason to switch.
On July 1, 2026 Venice announced its first outside capital: a $65 million Series A at a $1 billion equity valuation, led by Dragonfly with Coinbase Ventures, North Island Ventures, F-Prime, Archetype, Liquid2 Ventures and others. Voorhees said the money would fund Venice's first owned data center and GPU purchases to cut leased-compute costs, lift gross margins and grow toward a consumer app for hundreds of millions of people plus an API for AI agents.
What has to be true
- Tokens came before venture capital: staking VVV to mint DIEM credits gave users a reason to try the platform while Venice was still building, creating distribution without diluting the company.
- Privacy was the wedge big labs could not copy quickly: no stored prompts, client-side encryption and routing through an external proxy addressed the trust gap left by data-collecting chatbots.
- The uncensored positioning captured users who wanted unrestricted answers, and offering 200+ models with different moderation levels let each user pick their own tradeoff.
- Profitability before equity changed the negotiation: entering a Series A with 3M users and $70M+ run-rate revenue let Venice pick its first investors and price the round at $1B.
- Feature parity with ChatGPT amplified the pitch: as model quality closed the gap, privacy became the deciding difference instead of the only reason to switch.
What can be applied
Venice sold equity only after product and tokens proved revenue, so it negotiated from profitability: users and cash came first, capital followed to scale rather than subsidize adoption.
Aftermath
As of September 2, 2026, Venice AI is scaling a profitable business: the Dragonfly-led $65M Series A at a $1B valuation closed in early July 2026, its first outside capital. It had 3M+ users, about 1.7M daily API calls and over $70M in annualized run-rate revenue, profitable since Q1 2026. The round funds Venice's first owned data center and GPU fleet to cut leased-compute costs and raise gross margins, plus market expansion and acquisitions. VVV had risen over 700% in 2026 and Venice, its largest holder, had sold none.
Sources
- Venice AI becomes a unicorn with $65M Series A as its privacy-first AI platform takes off
- Venice AI raises $65 million Series A at $1 billion equity valuation led by Dragonfly
spotted an error? The archive wants to know.
Your turn
You just read one. Describe what you are building, and see who is betting on the same thing.
Free account · 3 free questions · no card