Pundi AI, based in Singapore, has launched a feature called Data Pump that allows contributors to convert their AI training datasets into tradable BEP-20 tokens.

Dataset owners can earn royalties and trade these tokens on decentralized exchanges like PancakeSwap.

The platform has already facilitated over 4.5 trillion data tokens across 122,097 datasets from 28,152 contributors.

Data Pump uses a bonding curve pricing model to apply market-driven valuation and includes safeguards against front-running.

Contributors verify dataset ownership through NFTs on the Pundi AI platform.

Once tokenized, datasets can attract community funding and be used by AI agents, turning static data into active, monetizable assets.

Pundi AI is also part of Nvidia’s Inception program and offers tools like a data marketplace and omnichain layer.

🔗 Source: Pundi AI

🧠 Food for thought

1️⃣ Data ownership revolution addresses market imbalance in AI economy

The data monetization landscape reveals a stark imbalance that Pundi AI is targeting with its Data Pump feature.

While the AI training dataset market is projected to grow to $9.58 billion by 2029 (mentioned in the original article), only 17% of companies have established data monetization initiatives, showing significant untapped potential 1.

This gap exists despite data being increasingly recognized as a valuable business asset, with successful monetization offering benefits including new revenue sources (69%) and improved services (66%) 1.

The shift toward decentralized ownership models responds to growing concerns about centralized control, as Web 3.0 emphasizes user ownership and privacy over the centralized model of Web 2.0 2.

Traditional AI development has concentrated profits with tech giants while data contributors receive minimal compensation, creating the imbalance Pundi AI aims to address with its tokenization approach.

2️⃣ Tokenization creates new asset class with verifiable value mechanisms