Venice AI Hits Unicorn Valuation as Privacy Concerns Shape AI Risk

Venice AI, the privacy-focused AI platform founded by Erik Voorhees, has raised $65 million in Series A funding at a $1 billion valuation, bringing the company to “unicorn” status. The round—led by Dragonfly and backed by Coinbase Ventures, F-Prime, North Island Ventures, Morgan Creek and others—was announced on Wednesday, and represents Venice AI’s first outside capital raise since it launched in 2024.

The funding arrives as privacy concerns around mainstream AI services are drawing renewed attention. Earlier this month, Anthropic cut off foreign access to two of its latest models, and in the broader public debate over AI data handling, a class-action lawsuit recently accused OpenAI of sharing ChatGPT data with third parties. Against that backdrop, Venice AI positions itself as a layer between users and model providers, designed to reduce what third parties can see about user activity.

Key takeaways

• Venice AI reached unicorn status after closing a $65 million Series A round at a $1 billion valuation, led by Dragonfly.
• The platform claims 3.5 million users and routes traffic through a proxy that can obscure IP address and user/account/session data from model providers.
• Venice AI says the new capital will fund more of its own infrastructure, including owning GPUs via data center expansion rather than relying entirely on rental capacity.
• The announcement lands amid heightened scrutiny of AI data privacy, including legal claims involving tracking technologies and alleged sharing of user information.

Unicorn funding for a privacy-first AI delivery layer

Venice AI’s Series A funding was announced by the company in a blog post published Wednesday, with Erik Voorhees describing the company’s mission in constitutional terms in a separate X post. Voorhees said the funding will be used to uphold the First and Fourth Amendments “as they relate to mankind’s interaction with AI.” In the U.S. legal framework, the First Amendment protects core freedoms including speech, while the Fourth Amendment restricts unreasonable government searches and seizures.

While the fundraising headlines focus on valuation and total capital, the more meaningful detail for potential users is the product model: Venice AI’s platform is built to act as an intermediary between a user and over 200 AI models. According to the company, users can choose the level of privacy they want, with different models routed through different privacy protections.

How Venice AI says it protects user data

Venice AI claims it has 3.5 million users. For models associated with OpenAI, Anthropic, xAI and Google, Venice AI says its proxy obscures users’ IP address as well as account and session data. The company also claims “other models offer higher levels of privacy,” indicating that its approach is not one-size-fits-all and may vary depending on which model is being accessed.

The core premise is that owning (or controlling) the “delivery stack” matters: if the intermediary is the part that can see traffic patterns and data flows, then that component can potentially reduce exposure to outside entities that operate the underlying model endpoints. Dragonfly managing partner Haseeb Qureshi framed the strategic stakes in those terms, arguing that whoever runs the AI delivery layer can see more about users’ behavior and ultimately influences the conditions under which users get access to powerful systems.

Where the $65 million will go

Voorhees said the Series A funding will be used to continue building Venice AI’s data center infrastructure. A central element of that plan is ownership of the compute resources—specifically, owning GPUs that power the platform—rather than renting them at higher costs.

Beyond infrastructure, Voorhees said remaining capital will support growth initiatives including expanding the customer base, entering new markets, hiring talent, and acquiring what he described as “additive businesses.” The acquisition language suggests Venice AI may be looking to broaden capabilities around its platform, though no specific targets were named in the materials provided.

Privacy scrutiny pushes privacy-focused AI into focus

Venice AI’s funding timing underscores how quickly privacy questions have become a defining topic for AI adoption. Earlier coverage from Cointelegraph reported that a user who consults an AI for legal matters could face the risk of chat logs being used against them in court. The broader theme is that AI interactions can generate sensitive records—even if users are not providing personal data intentionally.

In parallel, researchers and industry figures have proposed technical approaches to limit exposure. For example, the Ethereum Foundation’s AI lead Davide Crapis and Ethereum co-founder Vitalik Buterin proposed using zero-knowledge proofs and other techniques to help ensure that a user’s interactions with large language models are kept private.

Legal concerns have also intensified. In May, a proposed class action was filed in California federal court accusing OpenAI of disclosing private ChatGPT user data to third parties including Google and Meta. The complaint alleged that Meta Pixel and Google Analytics were embedded into ChatGPT.com, so that when users send queries, duplicate data is allegedly sent to Meta and Google along with advertising cookies and personally identifiable information—information that could then be used for targeted advertising.

These developments highlight a tension for users: modern AI platforms often involve multiple layers of data collection, analytics, and third-party integration, which can be difficult to disentangle from “model inference” itself. Venice AI’s proxy concept is an attempt to restructure that data path by introducing a dedicated intermediary that can obscure certain identifiers from model providers.

The recent industry shifts also reinforce why an intermediary approach is gaining attention. Anthropic’s sudden reduction in foreign access to two of its latest AI models earlier this month served as another reminder that availability and access controls can change quickly—while privacy-focused architectures aim to give users more predictable control over how their data is handled.

What to watch next

With Venice AI scaling its infrastructure and expanding adoption, the key question for investors and users will be how effectively its proxy-based design delivers measurable privacy protections across a wide set of models and real-world integrations. Readers should watch for more transparency around which metadata is obscured under each privacy mode, and whether Venice AI’s compute buildout translates into faster, more consistent performance without sacrificing its stated privacy goals.