FINOPS MEETUP: AUSTIN
From AI Cost Sprawl to Control.
Register Now
Your CFO asks what you got for $2M in AI spend. Most teams can’t trace it to a customer, feature, or outcome.
This will be an interactive discussion on how granularity increases as you move up the stack. Infra gives you cost per workload. GenAI ties spend to team, product, and feature. Agentic is the goal, cost per run, per workflow, per resolved task.
Three playbooks you can take back to your desk and start to apply:
1
Infra: Allocate GPU cost by time + memory to get cost per workload.
2
GenAI: Attribute spend across providers with LiteLLM, AWS Inference Profiles, Azure APIM, and GCP labels, no more unassigned AI spend.
3
Agentic: Catch retry loops and tool-call explosions, and map unit cost to outcomes like cost-per-resolved-ticket to prove ROI.
Leave with best practices on tracing GPU → LLM → agent → outcome, and the unit metrics the team needs to proactively manage costs around AI initiatives.
Meet the Speaker
