Veritone Data Refinery Hits 22.2T Tokens with 3.5x H1 Volume Surge
Data Refinery processed 22.2 trillion tokens in H2 2025, a 3.5x jump from H1, while aiWARE applied policy controls to unstructured audio and video. Deployments include Air Force Office of Special Investigations and media clients, tapping a dataset market set to grow from $7.48B to $52B over next decade.
1. Platform Expansion and Data Refinery Capabilities
Veritone expanded its aiWARE platform on February 11 to include enhanced Data Refinery capabilities that process unstructured audio and video, enforce governance with policy-based controls and audit logs, and generate AI-ready assets for model training.
2. Record Token Processing Growth
On January 29, the company reported that Data Refinery processed 22.2 trillion tokens in the second half of 2025, representing a 3.5x increase from the first half as customers indexed and formatted large datasets for AI model refinement.
3. Government and Media Deployments
The platform has secured deployments in regulated environments, including the Air Force Office of Special Investigations, and serves media, entertainment and talent acquisition clients seeking proprietary, licensed data for AI applications.
4. Dataset Market Opportunity
Veritone highlighted that the AI training dataset market is projected to expand from $7.48 billion in 2026 to $52 billion within the next decade, underscoring potential growth in data monetization and demand for ethical, high-quality datasets.