AT&T and H2O.ai Celebrate Commercial Milestone for AI Feature Store, Deepen Work in Vertical AI Super Agents

MOUNTAIN VIEW, Calif.--(BUSINESS WIRE)--H2O.ai today announced a significant milestone in their long-standing agentic AI co-creation collaboration with AT&T. H2O.ai will make a royalty payment to AT&T consistent with the commercial and intellectual property framework established at the outset of the collaboration.

The milestone reflects sustained enterprise demand for scalable, governed AI infrastructure designed to accelerate innovation while maintaining performance, reliability, and responsible AI standards.

From Co-Creation to Agentic AI Commercialization

Originally built to support AT&T’s massive data ecosystem and spanning hundreds of petabytes of network, operational, and customer data, the H2O AI Feature Store enables teams to reuse, govern, and operationalize machine learning features across the enterprise.

Feature engineering remains one of the most time-intensive components of AI development, often consuming up to 80% of a data scientist’s time. Without centralized infrastructure, teams repeatedly rebuild the same features, increasing cost and slowing deployment.

The H2O AI Feature Store addresses these challenges by providing:

• Offline and online feature serving for both batch training and real-time inference
• Sub-millisecond latency for production transactions
• Automatic Feature Recommendations to improve model performance
• Automatic feature drift detection and bias identification
• FeatureRank™ scoring to measure feature value across use cases
• Deep integrations with enterprise data ecosystems

“Our work with H2O.ai has always been about co-creating durable AI infrastructure that scales across AT&T’s network, operations, and customer platforms. What began as a shared innovation effort has become a foundational layer for our agentic AI systems — enabling reuse, governance, and real-time intelligence across the enterprise. This milestone reflects both the strength of our teamwork and the long-term value of building AI the right way.”

Andy Markus

Chief Data and AI Officer at AT&T

Building the Foundation for Next-Generation AI

The H2O AI Feature Store serves as a core component of AT&T’s broader AI architecture, supporting predictive models, generative systems, and emerging agentic workflows.

As enterprises move from experimentation to fully operational AI systems, foundational infrastructure such as feature stores ensures structured data, governance policies, and predictive intelligence can be seamlessly integrated into advanced AI platforms.

The royalty milestone reflects a shared intellectual property framework established at launch, aligning incentives for long-term innovation and commercialization while reinforcing the durability of enterprise-grade AI architecture.

Enabling Enterprise AI at Scale

AT&T continues to use the Feature Store to power high-impact use cases including:

• Network optimization
• Fraud prevention
• Predictive maintenance
• Financial forecasting and tax calculations

“Feature Stores are the digital water-coolers for Agents sharing context and memories. AT&T helped co-create one of the first truly enterprise-grade Feature Stores for Predictions. Our banking customers world-wide insist on using this co-innovation over homegrown feast derivatives. This royalty milestone is a model for how large enterprises can co-create AI infrastructure that benefits the broader market.”

Sri Ambati

Founder and CEO of H2O.ai

A Model for Strategic AI Collaboration

The royalty payment reflects a shared IP and co-created framework established at launch — aligning incentives for long-term innovation and commercial success.

As enterprises move from experimentation to scaled, operational AI systems, infrastructure components like feature stores are increasingly recognized as foundational to ROI, governance, and performance.