a16z Backs Yupp: $33M Bet on AI Meets Crypto Rewards

3 mins mins

Key Points:

Yupp rewards users for voting on AI outputs, generating transparent feedback for model improvement.
Blockchain enables on-chain incentives, ensuring auditable and fair data contributions across the Yupp platform.
a16z pushes for new governance models, calling Ethereum and Solana’s foundations limiting for growth.

Yupp, a new platform for AI model evaluation, has raised $33 million in seed funding led by venture firm a16z. The platform uses blockchain incentives to crowdsource human preferences for improving AI systems. The funding round includes participation from investors familiar with both artificial intelligence and crypto.

Yupp allows users to enter prompts, review AI-generated responses side-by-side, and vote on the best outputs. The choices made by users create signed preference data that can be used to fine-tune AI models. Participants are rewarded for their input, which gives users a direct role in shaping model development.

Blockchain Brings Transparency to Model Training

The team behind Yupp argues that current AI model evaluation systems are limited by secrecy and a lack of user insight. Companies often keep training data and evaluation processes private, making it difficult to assess model performance or improve quality based on real-world input.

In response, Yupp combines AI and crypto to offer a transparent, reward-driven system. The platform generates digitally signed feedback packets that model developers can audit. a16z stated in its announcement,

“Users get incentives, builders get trustworthy data, and everyone can audit either side of the open market.”

This structure also ensures that all participants follow the same rules. The reward system is visible on-chain, and no single entity can alter or hide evaluation results.

Experienced Team and Technical Strategy

Yupp’s founding team includes developers with backgrounds at Google, Coinbase, and Twitter. Pankaj Gupta previously led engineering at Google Pay and Coinbase, while Gilad Mishne worked on machine learning projects at GoogleX. The team has built consumer-facing AI products and plans to scale Yupp as a core evaluation layer for future AI systems.

Yupp introduces a system where preference data naturally refreshes. Older feedback is replaced with newer choices from users, creating a loop that helps improve model quality and relevance. This flow helps maintain a live and adaptive evaluation environment that grows as user activity increases.

Crypto Governance in Focus

Alongside the funding news, Miles Jennings, head of crypto policy at a16z, shared his view on current blockchain governance models. On June 2, he said the foundation structures behind chains like Ethereum and Solana are “limiting industry growth.” He called for new models that improve responsiveness and accountability.

DISCLAIMER: The information on this website is provided as general market commentary and does not constitute investment advice. We encourage you to do your own research before investing.