Venice AI Just Hit $70 Million—Without Siloing a Single Prompt
Erik Voorhees has been here before. Back in 2013, he launched Satoshi Dice—the first crypto gambling site—and watched a generation of bitcoin users discover permissionless money through dice rolls. Fast-forward to 2026: he’s back, this time armed with Venice AI and a $1 billion valuation.
Here’s the thing no one’s saying outright: most AI tools today are surveillance machines in disguise. You get a crisp response, sure, but your prompt becomes someone else’s training data, your query history gets logged, and if you ask something mildly controversial? Good luck getting past the content filter.
Venice AI flips that script. It’s not just another ChatGPT clone with a shinier interface. It’s built around the same philosophy that made bitcoin possible: neutrality, privacy, and permissionless access. And according to Voorhees—who sat down with TechCrunch this week—the platform is already profitable, running at over $70 million in annualized revenue.
That’s not a typo. Less than two years after launch, the company has hit unicorn status with $65 million in fresh funding led by Dragonfly Capital. But more importantly, it’s proof that people still want tools that respect them instead of extracting value.
Let me be clear: this isn’t the story of some VC-backed hype cycle. It’s about a founder who’s been walking this path for a decade, finally hitting product-market fit in the AI era—not by chasing the crowd, but by betting against it.
How It Works: Privacy-by-Default, Not Privacy-by-Accident
Most AI startups talk about privacy like it’s an afterthought—some checkbox you can tick if you pay extra. Venice AI treats it like the core protocol.
Here’s how they pull it off. Every prompt you type never touches Venice’s servers in plaintext. Instead, it gets encrypted right on your device, then routed through an external proxy before landing at the AI model—whether it’s open-source Llama 3 or closed-source ChatGPT. The response comes back encrypted and gets decrypted only on your side.
The result? Even if someone inside Venice wanted to snoop, the data’s unreadable. They don’t store conversation logs on their servers, and for users who want even tighter controls, the platform offers end-to-end encryption via its Pro tier.
That alone would make Venice notable. But they went further: they made uncensored AI the default.
See, most platforms implement "Safe" modes out of the box—filtering responses to keep things polite, compliant, and corporate-friendly. Venice AI flips that. The default experience lets users ask whatever they want; if you’d prefer a softer touch, there’s an optional "Safe Venice" toggle you can disable.
It sounds small. But in practice, it’s huge. It treats users like adults instead of managing them like children. Voorhees put it best when he compared Venice to Bitcoin: “Bitcoin, as a neutral protocol, works the same way for all people,” he told TechCrunch. “I think it’s actually quite dangerous from a safety perspective for the world to enter this next phase and have everyone be constantly watched.”
The Numbers Behind the Unicorn Status
Let’s talk scale because the growth numbers are wild—especially for a company launching in 2024:
- 3 million+ active users
- 850,000 unique monthly website visitors (NavTools.ai records 13.2 million monthly visits)
- 1.7 million API calls per day
- Annualized run-rate revenue exceeding $70 million
- $1 billion valuation on a $65 million Series A round, led by Dragonfly Capital
- 8% of users pay with crypto (VVV or DIEM), the rest use fiat
Most startups spend years chasing unit economics. Venice AI hit breakeven in under two years, partly because the product is the distribution. People aren’t just signing up for the tech—they’re staying because it aligns with their values.
The Pro tier ($18/month) drives most of the revenue, offering unlimited prompts, custom characters, advanced models, and high-resolution upscaling. The free tier holds them at 25 text prompts and 15 image generations per day—a generous enough offer to keep casual users onboard while nudging power users toward paying.
The Token Experiment: VVV and DIEM, One Day at a Time
Now for the crypto twist.
Venice AI launched two tokens last year: VVV and DIEM. VVV dropped in January 2026; DIEM followed in August with a simple promise—stake your VVV and you get $1 worth of AI credits per day, forever. Think of it as a recurring dividend for using the platform.
Voorhees admits the token rollout hasn’t gone perfectly. Only 8% of users currently pay with crypto, while the majority stick to credit cards and bank transfers. That’s fine by him: he’s building for the long term, not next quarter.
“The strength of the tokens lies in their utility,” he told TechCrunch. “Not speculation, not fluff—real utility that gets baked into the product.”
In practice, that means developers can access Venice’s API with VVV to run AI agents without per-request fees. Power users get daily AI credits via DIEM. The more you interact with the platform, the more the tokens pay you back.
If it sounds like Venmo meets Web3 in a calm, sensible way—that’s the point. Voorhees isn’t trying to disrupt banking or payments. He’s just adding a layer of transparency and ownership on top of a great product.
The Uncensored Edge: Why Users Don’t Just Stick to ChatGPT
Let’s be honest—ChatGPT is impressive. Its text generation, code snippets, and even image capabilities are refined. So why switch to Venice AI?
Here’s what people tell us after using both:
- “I can ask anything”: Whether it’s a controversial political opinion, a medical edge-case scenario, or graphic creative writing, Venice AI doesn’t censor its responses by default.
- “My prompts feel mine”: You walk away with full access to your chat history, no log-in required for basic use. Privacy-focused users appreciate not having their data resold.
- “More models, less friction”: Venice hosts over 200 open-source models—including Llama 3 and Stable Diffusion—and routes to closed-source ones like GPT-4. You pick the model, decide how much censorship you want, and go.
- “It’s just faster to get things done”: According to NavTools.ai data, the site averages 6.47 pages per visit and nearly nine minutes of engagement—suggesting users aren’t just clicking around. They’re getting things done.
That last point is critical. Many assume privacy-first means lower quality or slower speed. Venice proves otherwise.
“We were very far away from what ChatGPT could do when we launched,” Voorhees confessed. “But people would use us because it was private. Today, we’re very close to what ChatGPT can do… so as we’ve closed that gap, it’s become an increasingly compelling alternative.”
In other words: privacy brought people in. Quality kept them.
What’s Next for Venice AI?
The new $65 million round isn’t just about marketing spend or hiring engineers. Voorhees wants to finally own the stack.
Right now, Venice AI leases GPUs from cloud providers. It’s functional but expensive—and leaves margins vulnerable to fluctuating hardware costs. The plan is to use the Series A proceeds to buy its own GPUs and build dedicated data centers.
“The goal is to increase our gross margins,” he said, “by taking control of the infrastructure.”
That’s a classic crypto founder move—build on someone else’s stack until you can replace the middleman. Once Venice owns its own hardware, scaling becomes a cost optimization problem instead of a funding dependency.
Short-term, the focus is on better character customization for creators and more robust image generation tools. The Pro tier’s custom characters already let users design AI companions for storytelling, therapy, or just fun—but the team is pushing toward photorealistic avatars and longer-form video interactions.
Long-term? Venice could become the foundation for a truly decentralized AI layer on top of the internet. No single company decides what gets filtered, censored, or promoted. Just code, cryptography, and a protocol that works the same way for everyone.
Sound idealistic? Maybe. But理想 rarely built billion-dollar companies without real revenue to back them up.
Venice AI has both. And that’s why this isn’t just another AI startup story—it’s the first one you should be paying attention to.
