Agentic AI Redefines Cloud: Omdia's 2026 Analysis

Cloud-native architectures are seeing significant changes, with traditional layered systems beginning to break down. This is happening as different layers move towards joint scheduling in specific scenarios. The rise of large language models, or LLMs, since 2023 is a major factor. A single model API call can now cover functions across several layers of the cloud-native architecture. This integrates previously independent layers into the model itself. Also, enterprise procurement is shifting from paying "per-layer" to paying "per-token." Tokens are becoming the core billing unit, reducing the independent commercial value of the original layered structure. The relationship between humans and machines is also changing. Models and AI are increasingly acting as "labor." Examples include Salesforce's Agentforce and Microsoft's Copilot. Users of AI models are falling into two groups. The "API-first" camp prioritizes speed and cost, using off-the-shelf models for general tasks. The "Agent-centric" group focuses on complex problems requiring orchestration, where an agent acts as a decision-maker. These shifts impact how businesses will interact with and pay for cloud services in the future.