Privacy-first Venice AI hits unicorn status on $65M Series A

The News

Venice AI, the privacy-focused platform led by crypto veteran Erik Voorhees, has raised a $65 million Series A that vaults it to a $1 billion valuation, joining the small club of profitable AI unicorns. The round was led by crypto-focused venture firm Dragonfly, with Coinbase Ventures and North Island Ventures also participating.

Unusually for an AI company burning cash to chase scale, Venice says it is already in the black. The firm reports annualised run-rate revenues of over $70 million, drawn from more than 3 million active users and upwards of 850,000 unique website visitors. Its systems field an average of 1.7 million API calls a day.

The product itself bundles access to more than 200 models spanning text, image, audio and video generation. Its distinguishing feature is architecture rather than any single model: user inputs are encrypted on the device, routed through an external proxy, and, the company says, never stored on Venice's own servers. Paid subscribers get end-to-end encryption on top.

Why It Matters

Profitability is the headline that should make rivals uncomfortable. The prevailing model-lab economics, where firms raise billions and defer margins indefinitely, has trained investors to treat losses as a badge of ambition. Venice inverts that. It is a reminder that a thin aggregation layer sitting atop other people's models can capture real revenue without owning the frontier.

The positioning is also a bet that privacy is a product, not a feature. Voorhees, who previously founded the crypto exchange ShapeShift, frames Venice as "a neutral tool or a neutral platform," likening it to how Bitcoin functions as a neutral protocol. That philosophy runs through the crypto plumbing: a VVV token launched in January 2026 and a DIEM credit token added in August 2025, where users stake VVV to mint DIEM and generate a dollar of daily AI credits. Tellingly, only 8 percent of users actually pay in crypto, which suggests the token is more ideology than checkout button.

The last time a privacy-first challenger tried to unseat incumbents on principle rather than performance, it was the search engines promising not to track you. They carved out a loyal niche but never dislodged the default. Venice is making the same wager in AI, and its early revenue suggests the niche is larger than sceptics assumed.

Indian Angle

For Indian users, Venice arrives at an awkward intersection of two very different regulatory moods. On privacy, the direction of travel is friendly. The Digital Personal Data Protection Act, whose rules are being operationalised through 2026, pushes enterprises towards data minimisation and tighter control over where personal inputs travel. A no-logs, client-side-encrypted architecture is, on paper, a neat fit for compliance-nervous Indian firms in banking, health and legal services that want generative tools without handing raw customer data to a foreign cloud.

But Venice's crypto core is precisely where the Indian pitch stalls. India taxes virtual digital assets at a flat 30 percent with a 1 percent tax deducted at source on transfers, a regime designed to discourage exactly the sort of stake-and-mint token flows that power VVV and DIEM. An Indian developer wanting Venice's privacy benefits would be steered firmly towards the fiat subscription, stripping out the token economics that make the platform distinctive. The 8 percent crypto-payment figure would likely be far lower here.

There is a competitive read too. Home-grown model builders such as Sarvam and Krutrim are chasing sovereignty and local-language depth, not privacy-by-encryption. Venice's rise shows there is a second axis of differentiation, confidentiality, that Indian startups have largely left open. For enterprises weighing MeitY's evolving stance on AI accountability, an auditable no-retention design may prove more marketable in India than raw benchmark scores.

FAQ

What exactly does Venice AI sell?

Access to more than 200 third-party AI models for text, image, audio and video generation, wrapped in a privacy layer. Inputs are encrypted on the user's device and routed through a proxy, with the company saying nothing is stored on its servers. Paid tiers add end-to-end encryption.

How can it be profitable when most AI firms are not?

Venice is an aggregation layer rather than a model lab, so it avoids the enormous cost of training frontier models. It reports over $70 million in annualised run-rate revenue from more than 3 million active users, and plans to buy its own GPUs to widen margins further.

Do I have to use crypto to use it?

No. Venice offers standard fiat subscriptions, and only about 8 percent of its users pay with crypto. Its VVV and DIEM tokens are optional, letting users stake to generate daily AI credits, but they are not required to access the platform.

Why is the crypto element tricky for Indian users?

India levies a flat 30 percent tax on virtual digital assets plus a 1 percent tax deducted at source on transfers. That makes the stake-and-mint token mechanics costly and cumbersome for Indian users, who would generally be better served paying by conventional subscription.

This story was reported by TechCrunch. Read the full original coverage at TechCrunch.