Pulumi Survey Reveals Maturing Infrastructure Practices and the Rise of AI Agents in Cloud Management

Pulumi has released the "State of Agentic Infrastructure 2026" report, surveying 510 platform, DevOps, and product engineers to understand current trends in cloud infrastructure management. Key findings indicate that 82% of respondents deploy to production weekly or more often, highlighting a mature and fast-paced deployment culture. Multi-cloud adoption is also prevalent, with 64% using AWS, 55% Google Cloud, and 45% Azure. Furthermore, 63% already operate an internal developer platform or golden paths in production. This report is critical for practitioners because it validates the ongoing shift towards highly automated, multi-cloud environments. The rapid deployment cadence means that manual processes are no longer sustainable, driving the need for sophisticated IaC tools like Pulumi that can manage diverse cloud resources programmatically. The increasing presence of AI agents in infrastructure workflows, as highlighted by the report, signifies a future where AI will assist in everything from code generation to compliance enforcement. This necessitates that engineers not only understand IaC but also how to effectively integrate and govern AI-driven automation. These trends fit squarely within the broader movement towards platform engineering and the "AI-first" paradigm in DevOps. The widespread adoption of internal developer platforms (IDPs) reflects a desire to abstract away complexity for developers, providing self-service capabilities while maintaining centralized control and governance. The integration of AI agents, such as Pulumi Neo, into this ecosystem extends the capabilities of IDPs by enabling intelligent automation and proactive issue resolution. This evolution is a natural progression from traditional IaC, where human-written code defines infrastructure, to a more dynamic model where AI can interpret intent and generate or modify infrastructure configurations autonomously, always under human oversight and approval. Practitioners should pay close attention to the evolving role of AI in infrastructure. This means not only familiarizing themselves with AI-powered tools but also understanding the principles of "agentic infrastructure" – how AI agents can be safely and effectively deployed to manage cloud resources. Key areas to focus on include policy-as-code for AI governance, ensuring auditability of AI-driven changes, and developing strategies for human-in-the-loop approval processes. The report suggests that while AI offers immense potential for acceleration, robust guardrails and clear oversight will be paramount to prevent unintended consequences and maintain security and compliance in increasingly complex cloud landscapes.