Venice AI reaches $1 billion valuation with $65M Series A on a bet that enterprises want AI without surveillance

Summary

Venice AI, a Wyoming-based startup whose platform runs open-weight AI models locally on user hardware and logs no prompts, raised US$65 million in a Series A led by Dragonfly, with Coinbase Ventures and North Island Ventures participating. The round values the company at US$1 billion and is its first external funding; Venice was already generating over US$70 million in annual recurring revenue, unusual for a startup raising its first outside round. Co-founders Erik Voorhees (a crypto entrepreneur) and Jesse Proudman built the platform explicitly around the premise that surveillance-free AI is a compliance requirement rather than a niche preference.

The split

The US venture community frames Venice's raise as evidence of a structural privacy-AI market emerging alongside the cloud-AI incumbents. European commentary is more pointed: GDPR advisers in Brussels note that most major cloud-AI platforms fail the "data residency" test for sensitive industrial data, and Venice's local-inference model is the only current approach compatible with strict European data-transfer rules. Chinese and Russian regulatory environments, which have their own data-sovereignty mandates, represent a separate potential market for the architecture, though Venice has not disclosed plans to operate there.

By the numbers

• US$65 million, Series A round size
• US$1 billion, post-money valuation (unicorn)
• US$70 million+, Venice's reported ARR at time of funding
• 19 days, duration of Anthropic's Fable 5 export ban in June 2026 (the compliance event credited with accelerating Venice's growth)
• 0, server-side prompt logs in Venice's stated architecture

Why it matters

The Fable 5 export ban demonstrated in June that even the largest AI providers can become unavailable overnight due to a government order, creating a regulatory gap for enterprise buyers who had written specific models into contracts. Venice's raise suggests investors believe this compliance anxiety is large enough to support a parallel market for AI that removes the third-party-access surface entirely. The round is also a marker for the "local inference" thesis, which competes with the hyperscaler model on control rather than capability.

What to watch

• Venice's product roadmap after July 7, when Fable 5 returns to full access limits for Pro users
• Whether EU enterprise buyers begin standardising on local-inference architecture following the Fable 5 precedent
• Competing fundraising rounds from other privacy-first AI platforms (Mistral, Together AI)
• Regulatory clarity from the US Commerce Department on what triggers export-control review for AI models

---