Most conversations about enterprise AI focus on models. Veritone is making a different argument, and a multi-year deal with Oracle Cloud Infrastructure is the latest expression of it.
The California-based AI company signed an agreement this week to migrate key workloads onto OCI, designating it as a preferred platform for several core products including aiWARE, its platform for processing large volumes of audio and video, alongside Veritone Data Refinery and Veritone Data Marketplace. CEO Ryan Steelberg described the move as a step toward supporting customers with the performance, security, and scalability that growing AI demand requires.
The deal reflects something Veritone has been building toward for several years. The company spent considerable time re-architecting its AI stack, containerizing its core infrastructure in a way that allows workloads to move across cloud environments without being locked to any single provider. That flexibility matters now because the company is expanding globally and fielding increasing demand from governments and enterprises that want AI systems hosted within specific geographic and legal boundaries. Steelberg acknowledged that sovereignty has become a central concern for many clients, noting that some government customers are particularly deliberate about where their data and AI systems actually reside.
The data side of this partnership carries as much weight as the infrastructure side. Veritone’s Data Refinery converts raw audio and video into structured, usable AI training data, while its Data Marketplace distributes that processed content to developers.
A significant portion of the source material still exists in legacy formats, including tape archives that require digitization before any AI processing can begin. That conversion process has become faster and cheaper over time but remains a meaningful operational challenge for organizations sitting on large media libraries.
Steelberg’s view on where long-term value concentrates in the AI stack is direct. Models will continue to evolve and commoditize. The data pipelines feeding those models, and the organizations controlling high-quality training data at scale, hold the more durable advantage. Veritone’s positioning around processing and monetizing unstructured data rather than building foundation models reflects that belief in practical terms.
The partnership covers Veritone’s work across media and public sector clients, including broadcasters managing large content libraries and government agencies operating in environments with strict federal compliance requirements.
Veritone continues working alongside other major cloud providers rather than consolidating onto a single platform.
