We live under data serfdom, where we create valuable data and see no economic upside in the value we've helped create.
OpenAI, Meta, and Google all train models on publicly scraped data from the searchable internet, and are starting to buy up private data as they need access to more training data. Reddit for example, earns $200M from selling user-generated content as AI training data. As AI starts to play a larger and larger economic role in society, the economic impact of this data will grow.
We risk heading towards a future where AI models trained on our data displace us, with the economic gains flowing to a small set of shareholders. At the same time, we hold all the power through our data, but it requires collective action.
We believe that you should own a piece of the AI models that your data helps create.
Our north star is to empower users to own their data and the value it creates. We believe data will power the AI economic shift over the next decade. Giving users true ownership of their data opens up walled gardens and pushes AI progress forward through data abundance.
We apply the sovereign, decentralized technology that powers bitcoin and ethereum to personal data, shifting power from monopolistic big tech and distributing it back into the hands of the users who created the data. Vana provides the infrastructure to generate user-owned datasets that can replicate and supersede the datasets that big tech companies are today selling for hundreds of millions of dollars.
Vana is a decentralized data liquidity network designed to establish the first trustless open data economy.
Vana turns data into currency to push the frontiers of decentralized AI. It is a layer one blockchain designed for private, user-owned data. It allows users to collectively own, govern, and earn from the AI models trained on their data. For more context on why we built Vana, see this blog post.
At its core, Vana is a data liquidity network. It makes data liquid by solving the double spend problem for data, ensuring that data can be used like a financial asset in flexible, modular ways. This is achieved through two mechanisms:
Proof-of-contribution, which verifies the value of private data in a privacy-preserving manner
Non-custodial data, which ensures that the data is only used for approved operations
These mechanisms create a trustless environment where data can be securely tokenized, traded, and utilized for AI training without compromising user privacy or control. This paradigm shift not only democratizes AI development but also introduces a new economic model where data contributors become active stakeholders in the AI value chain.
Vana aligns incentives between data owners, developers, and data consumers. It creates a data-powered economy owned by its participants rather than centralized entities.
Learn more about the core concepts of the Vana Network by exploring these sections:
To participate in the network, you can:
How to establish ownership of the AI models created using our data
Empower users to own their data and the value it creates by decentralized technology
Build a User-Owned Data Treasury and a User-Owned Foundation Model
Understand the core building blocks of the Vana ecosystem
Explore the different participants and their role in the Vana network
Understand how data is transformed and validated and incitives will work
Build your own DLP based on provided templates and deploy to the Vana Network
Start the validation of data for specific DLPs on your own hardware
Submit your data to a DLP and observe your contribution onchain
With Vana, users and developers can incentivize global data contribution and accelerate the development of user-owned data applications, AI models, and data liquidity pools. These use cases have guided our architecture:
Incentivize 100 million people to export their Google, Facebook, Instagram, and Reddit data to create the first user-owned data treasury.
Vana enables non-custodial data storage, attributes voting rights based on data contributions, and verifies the legitimacy of data to ensure quality.
Each user adds their data to their personal server and grants access to a trusted verifier. Users then contribute their data to a collective server by encrypting it with the server's public key. The collective server operates according to rules set by the data contributors.
Build a model owned and governed by 100 million data-contributing users.
Vana stores model weights in a non-custodial way, secures distributed training on private data, allows users to earn through model usage, and enables collective governance of the model.
Users train a piece of the model on their personal servers and grant access to the foundation model DAO to merge all individual pieces. The foundation model DAO evaluates the value each person's data contributes and rewards them with a model-specific token. Developers interact with the model API by burning this token.