Pulumi Ushers in the Agentic Infrastructure Era with Enhanced AI Capabilities

Pulumi is positioning itself at the forefront of what it terms the "agentic infrastructure era," a period characterized by the increasing involvement of AI agents in managing cloud resources. The company recently unveiled a suite of new product capabilities and strategic partnerships designed to simplify the deployment and operation of AI infrastructure for developers, platform teams, and AI agents alike. A central component of this initiative is the expansion of Pulumi's Neo agent. Previously confined to Pulumi Cloud, Neo is now directly accessible within the Pulumi CLI, GitHub issues and pull requests, and Slack workspaces. This move allows users to leverage Neo's capabilities for scaffolding, migrating, investigating, and operationalizing infrastructure tasks without leaving their preferred development environments. The CLI integration, in particular, enables local execution of Neo, allowing it to inherit existing configurations and credentials for a more fluid interactive experience. Further enhancing agent-driven operations, Pulumi introduced the `pulumi do` command. This new feature provides a direct, imperative way to perform create, read, update, delete, and list operations across all Pulumi-supported cloud providers and resource types. This command is designed to be highly usable for both human developers and AI agents, simplifying complex infrastructure tasks into single commands. Pulumi is also addressing the specific needs of AI infrastructure with new integrations. Partnerships with NVIDIA and CoreWeave aim to provide infrastructure-as-code access to specialized GPU platforms and AI cluster runtimes. These integrations are crucial for AI teams looking to manage GPU infrastructure and AI workloads efficiently, bridging the gap between AI-generated code and safe, production-scale infrastructure deployment. The company emphasizes that these advancements are a response to the rapid adoption of AI in infrastructure management, noting that large language models (LLMs) are now responsible for over 20% of infrastructure deployments, a figure projected to exceed 50% within the year. Pulumi's goal is to provide a unified platform that allows both engineering teams and AI agents to provision, govern, and operate infrastructure across various cloud and SaaS providers with built-in verifiability, policy enforcement, and audit trails.