Slashing the cost of AI-powered local government services

Onetera teams up with Lenovo to make its local government AI solutions available on-premises, dramatically cutting the cost of running token-intensive workloads.

Eager to enable its municipal customers to benefit from AI in their own protected environments and to drive down AI workload processing costs, Onetera worked with the Lenovo AI Center of Excellence (CoE) to validate the accelerated breakeven and favorable token economics of on-premises infrastructure.

The challenge

• •       Local governments are under constant pressure to do more with less
• •       Large token volumes and high cloud API pricing push up the cost of agentic AI
• •       Data sovereignty is another key concern

The solution

• •       Deployed LLMs on a Lenovo ThinkSystem SR675 V3 server
• •       Every stage of the pipeline was kept inside the infrastructure boundary
• •       During the 40-day pilot, only one of the server’s eight GPUs was actively used

The outcome

• •       Cost savings of up to 98% for token-heavy workloads
• •       Lower cost makes agentic AI more accessible to local governments
• •       Ensures complete control over where data is stored and where data flows