Concurrence Processes 1.2 Trillion Tokens Annually Using Unity Gateway

Concurrence processes an annualized run rate of 1.2 trillion input tokens, reflecting a significant scaling of its clinical AI capabilities. The company utilizes Unity Gateway for centralized AI access and security, which facilitated the generation of 35.85 billion input tokens in July 2026 alone. This marks a 5x growth in monthly token volume since late 2025. Monthly, Concurrence handles 100.8 billion input tokens and makes 11.2 million LLM calls, with a peak of 90,000 world-model events recorded daily. The architecture is built on Databricks' shared data foundation and employs Apache Spark™ to derive its world model. Concurrence's data governance is managed through Unity Catalog, while Zerobus Ingest streams events into governed Delta tables. Compliance is a priority, with HITRUST and ISO 27001/42001 certifications currently in progress, alongside existing HIPAA, GDPR, and SOC 2 compliance. The integration of intelligent routing is expected to optimize model choice for healthcare workloads, as stated by Concurrence. Currently, 14 individual users are generating tokens through Unity Gateway, with 14 models serving production inference and a total of 46 models in the governed catalog. This follows previous reports on DeepSeek and Moonshot AI's challenges with hidden costs and routing through Claude.