Venice founder Erik Voorhees warns against government control of AI

Erik Voorhees has a message for Washington: please don’t make yourself the bouncer at the door of artificial intelligence.

The Venice.ai founder posted on X on November 12, 2026, warning against putting politicians in charge of AI. His argument is that political control over machine intelligence would pile enormous power into very few hands.

His proposed fix is the one he has championed for years. Keep the code open, keep power spread out, and let transparency do the work that regulators promise to do.

What set Voorhees off

The spark was a clip from the All-In podcast. In it, co-host David Sacks predicted that Democrats would push for a new federal department with the power to approve frontier AI models before they are released.

Sacks made that call on November 7, 2026. Voorhees weighed in five days later.

Frontier models are the most advanced AI systems being built. A pre-release approval office would work something like a drug regulator for software: nothing ships until the government signs off.

That is precisely the arrangement Voorhees objects to. He argued that a veto over AI, packaged as a safety measure, could hand the state what he called a “coercive monopoly” over machine intelligence.

His case flips the usual safety logic. Supporters of central oversight tend to present it as the cautious choice. Voorhees presents it as the dangerous one.

In his view, open-source transparency and decentralized power are the better guardrails. Code that anyone can inspect is harder to quietly bend than code reviewed by a single agency behind closed doors.

It is worth stressing what Sacks offered: a prediction. No such department exists, and the source material does not confirm that Democrats have formally proposed one.

A live legislative debate

The prediction did not land in a vacuum. US lawmakers are already arguing over how tightly to regulate advanced AI.

One proposal in that conversation is the Sanders-Casar bill, which seeks to impose regulatory approval requirements on advanced AI models. That is the general shape of what Voorhees is warning about: a permission slip before launch.

Who is Venice.ai

Venice.ai pitches itself as a private, permissionless AI service. Users can access models without asking anyone’s permission, and the company says it does not store user conversations.

The business has grown quickly. On July 1, 2026, Venice closed a $65 million Series A round at a post-money valuation of $1 billion.

The platform serves over three million users. It has also been profitable since the first quarter of 2026.

Venice processes 1.3 trillion tokens per month. In AI terms, a token is a small chunk of text a model reads or writes, so that figure is a measure of usage rather than anything on a blockchain.

Venice does, however, have a crypto token of its own: VVV. That gives the company a foot in both the AI and digital-asset worlds.

Voorhees’s position here tracks with his longer record in crypto. He has consistently pushed back against authoritative governance structures, particularly ones he sees as threats to privacy and free expression. Applying that lens to conversations between humans and AI is a natural extension.

What this means

For AI developers, the stakes are practical. A pre-release approval regime would add a new gate between building a model and shipping it.

Gates tend to favor whoever can afford to stand in line. Large labs with deep legal teams could absorb a lengthy approval process more easily than smaller or open-source projects, which is part of why critics like Voorhees see centralization risk baked into the design.

For Venice specifically, the tension is direct. Its whole pitch is private, permissionless access with no stored conversations. A federal sign-off requirement for frontier models could sit awkwardly alongside a product built around not asking permission.

Readers should also keep the incentives in view. Voorhees runs a $1 billion AI company whose model depends on light-touch access, so his warning is both a policy argument and a defense of his business.

The things to watch are fairly clear. First, whether Sacks’s prediction materializes into an actual proposal for a federal AI approval department. Second, how far the Sanders-Casar bill and similar efforts advance in Congress.