AI Infrastructure Cost Optimization

Routes AI workloads to lower-cost models and spots infrastructure cost inefficiencies.

Business impact

• Share of employee AI queries powered by open-source models — AT&T reports 40%.
• Quality of AI performance on those tasks — AT&T reports quality fell just 2% on the tasks routed to open-source models.

Data requirements

• Employee AI queries (text) — AT&T routes employee task queries among models.
• Infrastructure cost signals (tabular) — The cost agent looks for cost inefficiencies in infrastructure.

AI methods and techniques

• Model routing — Tasks are directed to different models based on complexity.
• Cost analysis — An internal agent identifies cost inefficiencies in infrastructure.

AI models and model families

Nemotron

Industries

Real-world evidence

2 documented case studies on record.

Companies using this: AT&T, Snap.