Pulumi today announced a series of new product capabilities and partnerships aimed at simplifying AI infrastructure deployment and operations for developers, platform teams, and AI agents. The updates include new integrations with NVIDIA and CoreWeave, alongside expanded capabilities for Neo, Pulumi’s platform-engineering AI agent.
The company said the new features are designed to help engineering teams deploy and manage cloud infrastructure more efficiently while maintaining governance, operational visibility, and production-scale control in AI-driven development environments.
As agentic coding platforms such as Claude Code, OpenAI Codex, Cursor, and GitHub Copilot accelerate software development workflows, organizations continue facing operational challenges when deploying and managing production infrastructure.
Pulumi stated that while AI coding tools can rapidly generate code, safely deploying and operating infrastructure requires broader operational context, governance controls, policy enforcement, infrastructure state awareness, and monitoring visibility.
“The gap between generating code and getting it safely into production is the defining bottleneck of the AI era,” said Joe Duffy, CEO of Pulumi. “We're building the infrastructure platform that closes that gap for human engineers, for platform teams, and for the AI agents that are increasingly part of every development workflow.”
Pulumi announced significant updates to Neo, its AI-powered infrastructure agent first introduced for Pulumi Cloud in late 2025.
Neo is now integrated directly into the Pulumi CLI, allowing developers to scaffold, deploy, and operate infrastructure from the command line without leaving their existing workflows.
The company also introduced GitHub integration capabilities that enable teams to mention @neo within pull requests to investigate infrastructure issues, review changes, and recommend fixes.
In Slack, Neo can participate in operational conversations, monitor incident discussions, and provide infrastructure context directly within team collaboration channels.
Pulumi stated that these integrations are designed to give Neo broader contextual awareness across development and operational workflows.
Pulumi also introduced a new Pulumi Integration Catalog that allows organizations to extend Neo using remote MCP servers and operational tooling integrations.
The catalog supports integrations with platforms such as Datadog, GitHub, Linear, and Sentry, enabling Neo to reason across operational, development, and infrastructure datasets.
Additionally, CLI integrations beginning with kubectl allow Neo to interact directly with runtime infrastructure systems.
According to Pulumi, these enhancements enable Neo to operate beyond Pulumi state data and reason across a broader operational environment.
A new recurring task capability allows organizations to schedule ongoing infrastructure operations and governance activities using Neo.
Teams can automate processes such as:
Pulumi noted that scheduled tasks continue to follow the organization’s RBAC policies, audit controls, and governance frameworks while delivering outputs through pull requests and standard review processes.
“I started looking at Pulumi and I recognized the problems that we were trying to solve around the reliable deployment of infrastructure from code, and also managing fairly complex configuration across a number of accounts… this is a company who are making a bet and making an investment on agentic solutions,” said Ewan Dawson, CTO, Compostable AI.
Pulumi also announced several new partnerships designed to simplify AI training and inference infrastructure management.
Pulumi introduced the first infrastructure-as-code provider for NVIDIA AI Cluster Runtime (AICR), enabling organizations to standardize GPU infrastructure configuration across Kubernetes environments.
The integration allows teams to create reusable infrastructure snapshots that preserve known-good combinations of GPU drivers, operating systems, and Kubernetes versions to reduce configuration drift in GPU-intensive AI workloads.
Pulumi also confirmed its participation in NVIDIA’s Inception program.
The company also partnered with CoreWeave to deliver infrastructure-as-code integrations for CoreWeave’s AI cloud platform and Weights & Biases.
According to Pulumi, the integrations connect compute provisioning, model training, optimization, and deployment workflows into a unified infrastructure management experience for AI teams.
The company stated that the new integrations aim to help organizations move AI workloads from experimentation into production more efficiently.
Pulumi additionally announced that all Pulumi Cloud capabilities are now available directly within the Pulumi CLI. The CLI is designed for both human users and AI agents, providing command-line access to all platform operations along with agent-friendly formatting and automation support.
The company said the newly announced capabilities are rolling out immediately and throughout the week.
Pulumi is the platform the agentic infrastructure era runs on. By modeling cloud infrastructure in real programming languages — Python, TypeScript, Go, C#, Java, and more — Pulumi gives both engineering teams and AI agents a single substrate for provisioning, governing, and operating infrastructure across every cloud and SaaS provider. Built-in verifiability, policy enforcement, and audit trails make agentic infrastructure work safe at production scale. Trusted by thousands of organizations worldwide, including leading AI labs, hyperscalers, and Fortune 500 enterprises, Pulumi is how the next decade of infrastructure gets built. For more information, visit pulumi.com.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.