3 - 6 Hours 4 Modules
Main Topic: FinOps for Generative AI
Workload: 3 - 6 Hours
Level: Advanced
Target Audience: Cloud Architects, Developers, CTOs, and Financial Managers
Deep understanding of how AI billing works. The student will learn to measure and predict token consumption in real-time, transforming abstract spending into clear business metrics to prevent the organization from exceeding the planned budget.
Architecture of variable costs. How to create a "Router" that analyzes the complexity of the user's question: it sends simple requests to economical models and reserves elite models only for complex tasks.
Implementation of intelligent memory layers. The student will learn to configure Semantic Caching: if a similar question has already been answered previously, the system delivers the saved response instead of processing and paying for a new API call, drastically reducing latency and cost.
Setting up financial barriers. Techniques to prevent programming errors or infinite loops from draining the company's credits. Implementation of automatic alerts and strict limits by department or project, ensuring profitability and budget predictability.
Flexera’s “State of the Cloud Report 2025” shows that cost remains a priority. With cloud spending expected to grow 28% in 2026, budget forecasting is still a challenge and organizations are exceeding their budgets by 17%.
At the same time, organizations keep maturing their governance over cloud investments, looking for ways to optimize the cost of software licences in the cloud.
AI ADOPTION IS EXPLODING
Unsurprisingly, adoption of public cloud services related to artificial intelligence (AI) is growing fast. 79% of organizations are already using or experimenting with AI and machine learning (ML) PaaS services. Use of data warehouse services, often used to feed AI models, has grown as well.
72% of organizations already use generative AI (GenAI) - heavily or moderately - and another 26% are currently experimenting with it. Needless to say, GenAI is here to stay and is on its way to becoming mainstream, at least to some degree, in the near future.
| Challenge | % |
| Managing cloud spend | 84% |
| Security | 77% |
| Managing software licences | 75% |
| Governance | 75% |
| Lack of resources/expertise | 75% |
