Take control of runaway AI costs with AI Gateway | THR902
Mike Budzynski presents how platform teams can apply FinOps controls directly in the AI runtime, managing consumption as AI usage grows across applications and agents.
Managing AI consumption at runtime
The session focuses on using AI Gateway to make AI spending more predictable and support adoption at scale. It covers five capabilities for understanding and controlling consumption:
- Centralized visibility into AI usage across applications and agents.
- Token quotas to control consumption.
- Rate limits to constrain usage rates.
- Semantic caching to improve efficiency and reduce waste.
- Intelligent routing as part of runtime cost management.
Together, these controls bring cost management into the runtime rather than treating it only as a reporting concern.
Session details and resources
- Event: Microsoft Ignite 2026
- Session: THR902 — Cloud & AI
- Format: Theater
- Level: Advanced (300)
- Speaker: Mike Budzynski
Explore the Cloud & AI next-step resources for further learning.