Gata Partners with Walrus to Build Open Execution Infrastructure for AI
Gata builds on Walrus to slash data costs for its open execution network, making decentralized AI economically viable at scale
Gata, a company developing decentralized large model inference, training, and data technologies, is integrating Walrus to power its open execution infrastructure for AI. Starting with Gata’s first product, DataAgent, Walrus will provide the critical data layer for Gata’s suite of applications, which are designed to bring the benefits of decentralized AI directly to users while making AI universally accessible.
Building on Walrus, Gata will overcome the prohibitive data access costs that have constrained decentralized AI development. The integration directly powers Gata’s DataAgent, a decentralized inference-powered data factory. Anyone can use DataAgent to contribute idle compute directly from their browser with a single click, collaboratively generating the synthetic datasets essential for AI training.
To generate these datasets, DataAgent requires its distributed network of contributors to frequently access and process large data files. This high access volume can lead to significant storage and retrieval fees on traditional decentralized networks.
Walrus makes this decentralized infrastructure cost-effective for Gata by offering storage costs and performance comparable to centralized clouds, and, most critically, eliminating download fees entirely. This allows Gata’s community to perform the heavy, permissionless data interactions required for AI workloads without incurring unpredictable and unsustainable costs.
“The promise of decentralized AI has been hampered by a fundamental economic challenge: data access is prohibitively expensive,” said Rebecca Simmonds, Managing Executive of the Walrus Foundation. “Teams building the future of AI shouldn't have to choose between decentralization and affordability. With Walrus, Gata can now build on a data layer that provides both. Gata’s work in creating an open execution layer for AI is exactly the kind of innovation we want to empower, and we're thrilled to provide the critical infrastructure to help them succeed.”
This partnership unlocks the next phase of Gata’s roadmap, including fully decentralized AI inference and training. Looking ahead, Gata plans to leverage Walrus’s unique onchain data programmability to pioneer an open, auditable pricing model. By storing verifiable compute proofs on Walrus, Gata can use Move-based smart contracts to transparently validate compute work and settle payments providers onchain. This creates a fair pricing system, an improvement on the opaque models used by today's centralized AI providers.
“Today, hyperscalers like AWS are extracting 39% operating margins from the $80 billion AI cloud market,” said Oliver, Co-founder at Gata. “This centralized control not only drives up costs but throttles access—stalling AI’s massive economic potential. Our mission is to make AI universally accessible, starting with an open execution infrastructure – which depends on an equally open and cost-effective data layer. Walrus provides decentralized infrastructure that eliminates our data access cost friction. This partnership is foundational for us, not only for our DataAgent application today, but for our future roadmap of decentralized inference and training for large AI models.”
The partnership is a step forward in building a complete, economically sound technology stack for decentralized AI. Gata’s DataAgent application is now live, leveraging Walrus to power a new generation of community-driven data creation. To learn more about Gata’s mission to decentralize AI, please visit: https://www.gata.xyz.
