In this episode of The Beacon Podcast, we explored Anna’s path from working at the Federal Reserve to YC-backed AI startups, her thoughts on data as an asset, and how AI founders can navigate this new frontier.
A High Schooler with a Picture of Janet Yellen: Anna’s Early Fascination with Currency
Unlike most teenagers, Anna’s high school bedroom featured a portrait of Janet Yellen, then Chair of the Federal Reserve.
“I was fascinated by central banks and how controlling currency means controlling the economy,” she shared. This obsession with finance led her to intern at the Federal Reserve in high school, where she saw firsthand how financial systems operate.
But when she got to MIT, everything changed. She discovered the MIT Bitcoin Club—a niche, ideological group discussing decentralized money.
“It was a bunch of anarchists wanting to take back control of currency,” she recalled.
This radical shift in thinking led her to mine Ethereum in her dorm, using discarded GPUs from MIT’s labs and free dorm electricity to power her rigs. It was her first hands-on experience with decentralization—a theme that would define her career.
AI and Decentralization: The Idea Behind Vana
Anna’s AI journey began with Iambiq, a YC-backed machine-learning startup. She quickly realized that better AI models required better data, but the real problem wasn’t just access, it was ownership.
“Whoever controls the data, controls the AI. And today, big tech owns it all,” she explained.
She and her co-founder Art started asking:
- What if users owned their data?
- What if AI models were built on data contributed by individuals, not corporations?
- How can blockchain help decentralize data ownership?
Early on, they experimented with data labeling, getting people to tag data from their phones manually. But they soon realized the real bottleneck wasn’t labeling—it was liberating data trapped in walled gardens.
“We saw that big data companies kept failing contracts because they couldn’t actually access user data,” Anna said. That’s when we knew decentralization wasn’t just a philosophy, it was a compliance advantage.
The Reddit Data DAO: A Surprising Success
One of Vana’s biggest breakthroughs was the Reddit Data DAO, a community-driven initiative where 140,000 real users pooled their Reddit data.
“We honestly didn’t expect it to blow up the way it did,” Anna admitted.
The DAO allowed users to sell access to their own Reddit data, bypassing Reddit’s corporate data deals. Reddit responded aggressively, trying to shut it down.
“That was a huge validation of why decentralization is necessary,” Anna said.
Since each user in a Data DAO controls their own data, it aligns with GDPR and other data ownership laws, which grant individuals rights over their personal data. This means that while platforms like Reddit may try to restrict access, users leveraging Data DAOs can legally assert their right to share and monetize their own data collectively. This forces platforms like Reddit to comply, creating a real-world use case for decentralized data ownership.
Now, Data DAOs are spreading. LinkedIn, Twitter, and even 23andMe data pools are emerging, letting users control how their data is used in AI training.
Why AI Needs a Decentralized Data Revolution
The AI industry is hitting a wall. The largest models today (like LLaMA 3) have already been trained on the entire public internet—about 15 trillion words.
“We’ve basically compressed the entire internet into an AI model,” Anna said.
So what’s next? Private data—the type that lives in social media, emails, private communities, and chat apps—is becoming the new frontier. AI models need more than just raw text; they require high-context, user-driven data to improve personalization, reduce bias, and enable more adaptive learning. This makes decentralized data-sharing solutions even more critical for the future of AI.
- Companies like Reddit are already selling their user data for hundreds of millions of dollars.
- AI researchers are desperate for new, high-quality data sources.
- But big tech owns this data, leaving users with zero control or compensation.
Anna believes AI is going to create a massive labor disruption—but the next wave of wealth won’t be from trading tokens, it will come from users owning and monetizing their data in structured ways, such as selling access through smart contracts or participating in revenue-sharing models within Data DAOs. This shift allows individuals to benefit directly from AI development rather than just being passive contributors.
“People don’t use products for ideological reasons. They use them because they unlock something new,” she emphasized. That’s why Vana’s Data DAOs work—they provide real economic incentives, not just ideals.
Building in AI: Anna’s Take on Founding a Startup in This Space
She emphasizes that founders must balance long-term vision with real-world adoption today. While patenting non-custodial data was important, their success with the Reddit Data DAO proved that practical use cases are key to traction.
She also underscores the importance of strong co-founder relationships. “The number one thing I wouldn’t change? My co-founder. Having someone you trust is the foundation of everything.”
Lastly, Anna believes that AI and blockchain are deeply interconnected, and the future of Web3 will be shaped by their convergence. Rather than seeing AI and crypto as separate industries, she sees decentralization as a key enabler for AI-driven economies.
Final Thoughts
Anna Kazlauskas’ story is one of bold vision, technical depth, and resilience. From hacking together Ethereum mining rigs in college to pioneering decentralized AI infrastructure, her journey is a case study in identifying trends early and executing relentlessly.
For Web3 AI founders, her message is clear:
- Build for real use cases, not ideology.
- Data ownership is the next economic frontier.
- Decentralization is a competitive advantage, not just a philosophy.
Want to dive deeper? Listen to the full episode of The Beacon Podcast here!
