Celebrating One Year of Walrus: What’s Been Built and What’s Next for Verifiable Data

Main Takeaways

• Walrus crossed 467TB of unencoded data in its first year on Mainnet, making it the second-largest decentralized storage protocol by total storage.
• Three major platform releases (Seal, Quilt, and Upload Relay) expanded Walrus from a storage layer into a programmable, developer-friendly data platform.
• Partnerships with Team Liquid, Allium, and Alkimi, among others, showed Walrus handling enterprise-scale data across industries where onchain finance and AI are moving fast.

Overview

Walrus Mainnet launched in March 2025 with a straightforward pitch: that a data layer could be fast enough, reliable enough, and programmable enough to power the future of critical applications in AI, DeFi and more.

What the first year has revealed is what becomes possible when data is verifiable, programmable, and controlled by whoever creates it rather than whoever hosts it. The features released in year one reflect that, and so do the organizations that have chosen to build on it.

Over 467TB of data is now stored on Walrus, making it the second-largest decentralized storage protocol by total data stored. That number is a measure of adoption and a measure of trust, with data from organizations like Team Liquid, Allium, and Myriad, committing to decentralized infrastructure and verifiability at scale.

How the platform has evolved

Walrus’s first significant expansion came with Seal. Decentralized infrastructure is transparent by default, which suits some use cases and disqualifies others.

For example, healthcare data, financial records, proprietary AI datasets, and more couldn’t live on a public layer without controls around who could read them. Seal made Walrus the first decentralized data platform with native access control, letting developers encrypt data and define exactly who could access it, enforced onchain.

Next came Quilt, which made working with large volumes of small files more efficient. Encoding overhead compounds quickly for NFT collections, AI agent logs, or messaging applications. Quilt groups up to 660 small files into a single storage unit through a native API, cutting that overhead significantly and saving Walrus partners over 3 million WAL.

The TypeScript SDK upgrade and Upload Relay made developers’ lives easier by handling the distribution of data across Walrus’s storage nodes. This made uploads faster and more reliable, even in low-connectivity environments, and gave developers a lightweight, self-hostable alternative to running a full Publisher.

Walrus in the wild 

With 200+ projects now building on the protocol, Walrus is proving that the data problems defining industries like adtech, institutional finance, and prediction markets are solvable at scale.

Team Liquid

Team Liquid migrated 250TB of match footage and brand content to Walrus, the largest single dataset entrusted to the protocol. The decision was driven by the practical value of turning a static archive into onchain-compatible assets that could support new fan experiences and content monetization without requiring another migration.

Alkimi

Alkimi built a new approach to digital advertising on Walrus, storing all campaign data onchain with Seal managing confidentiality, giving advertisers and publishers a verifiable audit of every impression and dollar spent. The platform now processes over 25 million ad impressions per day and 3.5 billion transactions total.

Allium

Allium, whose blockchain data infrastructure is used by Visa, Stripe, and Coinbase, brought 65TB of indexed institutional data to Walrus in Q1 2026. Datasets are encrypted natively and unlockable upon purchase by both humans and AI agents. 

Myriad

Myriad built transparent prediction markets on Walrus, processing over $5 million in transactions with all underlying data stored verifiably on the protocol. For prediction markets, where trust in the underlying data is everything, verifiable infrastructure is central to the product.

What comes next

The first year answered the question of whether Walrus could handle serious data at scale. Enterprise content, institutional finance, and programmatic advertising all found their way to Walrus. The more interesting question is what comes next.

Among the use cases taking shape, AI agents stand out. As agents take on more complex roles across financial systems and autonomous workflows, the data they rely on needs to be persistent and verifiable. That's a gap Walrus is built to fill.

Explore the protocol, the ecosystem, and what's coming next at walrus.xyz.